Srini's Point of View

Disqus comments not loading

2019-05-13T08:30:00-04:00

I am a little upset that my disqus comments widget is not loading on my previous post from last week. It is loading fine on all the older posts but not on the last one. Posting a random rant just to see if it is an issue with that post or everthing going forward. I hope it is a onetime glitch. I have a feeling it maybe got to do with me changing the slug on that article. Will find out shortly.

Update - after a bunch of trials, found the issue with how my SITEURL on my pelican config was causing this. If I have RELATIVE_URLS set to true, SITEURL becomes relative too. So places where I need the SITEURL be the absolute URL, like for instance in my disqus generated section for each article.. it was breaking that articles disqus widget.

This article gave me a solution on how to fix this. I defined a new variable called SITEURL_ABS = SITEURL in publishconf.py and changed the template where disqus widget was getting generated and use the SITEURL_ABS variable instead of SITEURL. This needs to be done as SITEURL gets changed if RELATIVE_URLS is set to True.

How to setup kubernetes cluster on EC2 to run the Dask NYC Taxi Demo

2019-05-08T11:00:00-04:00

I heard about Dask at my local Python Meetup last month. After reading about it, I came across Matthew Rocklin’s Pycon 2017 talk on Dask and was blown away by his demo of crunching through around 20 gigs of csv data (60 gigs in memory) in matter of a minute. You have to see the demo to really appreciate the power of Dask. He also wrote an article on his blog, which goes over the same example.

I thought the best way to get started with Dask would be to re-create the demo. Reading on the blog, I saw that Matthew was using an eight node cluster on EC2 of m4.2xlarges (eight cores, 30GB RAM each). So I started to look into how to setup that eight node cluster. The main documentation now recommends deploying Dask with Kubernetes and Helm.

This is where my in-experience working with kubernetes became an obstacle. Luckily the documentation and watching a few youtube videos helped me get started. I am also doing all this from Windows 10 on WSL.

Steps to get this demo working!

1. Make sure your aws cli is setup. Read up install instructions here and configuring instructions here. In my case aws cli was already setup as I use it for other work.

2. Install Docker. As of May of 2019, docker does not work natively on WSL. The way around it is to have Docker be installed on windows and just install the docker client on WSL. Nick Janetakis has a great write up on his blog and I got docker working from WSL in no time.

3. Install kubectl. The official docs covers a lot of ways but none that covers wsl. I found a writeup on DevKimchi where I found out that the windows Docker settings has an option to enable Kubernetes. Once enabled just copy the configuration from windows into WSL and also install the kubernetes cli in WSL.

$ mkdir ~/.kube
$ cp /c/Users/[USERNAME]/.kube/config ~/.kube/
$ curl -LO https://storage.googleapis.com/kubernetes-release/release/$(curl -s https://storage.googleapis.com/kubernetes-release/release/stable.txt)/bin/linux/amd64/kubectl
$ chmod +x ./kubectl
$ sudo mv ./kubectl /usr/local/bin/kubectl

4. Install Kops. Kops is a tool used to create, destroy, upgrade and maintain production-grade, highly available, Kubernetes clusters from the command line. Installing on WSL was pretty straight forward.

$ curl -LO https://github.com/kubernetes/kops/releases/download/$(curl -s https://api.github.com/repos/kubernetes/kops/releases/latest | grep tag_name | cut -d '"' -f 4)/kops-linux-amd64
$ chmod +x kops-linux-amd64
$ sudo mv kops-linux-amd64 /usr/local/bin/kops

5. Setup AWS IAM group and user. I am going to create a group called kops and user called kops.

$ aws iam create-group --group-name kops
$ aws iam attach-group-policy --policy-arn arn:aws:iam::aws:policy/AmazonEC2FullAccess --group-name kops
$ aws iam attach-group-policy --policy-arn arn:aws:iam::aws:policy/AmazonRoute53FullAccess --group-name kops
$ aws iam attach-group-policy --policy-arn arn:aws:iam::aws:policy/AmazonS3FullAccess --group-name kops
$ aws iam attach-group-policy --policy-arn arn:aws:iam::aws:policy/IAMFullAccess --group-name kops
$ aws iam attach-group-policy --policy-arn arn:aws:iam::aws:policy/AmazonVPCFullAccess --group-name kops
$ aws iam create-user --user-name kops
$ aws iam add-user-to-group --user-name kops --group-name kops

6. Set up an S3 bucket to save your cluster state. Just choose your own unique name.

$ aws s3api create-bucket --bucket some-kops-state-bucket --region us-east-1

7. Setup environment variables to help create our clusters. These variable names are defaults which are used by kops. Just make sure sure the cluster name you choose ends with “k8s.local”. You can also add this to your .bashrc file.

export KOPS_CLUSTER_NAME=somekops.k8s.local
export KOPS_STATE_STORE=s3://some-kops-state-bucket

8. Create your cluster. I am using 1 t3.xlarge (4cpu/16GB RAM) as my master and 8 m4.2xlarge (8 cpu/32GB RAM) as my worker nodes.

$ kops create cluster --name=${KOPS_CLUSTER_NAME} --node-count=8 --node-size=m4.2xlarge --master-size=c4.4xlarge --zones=us-east-2c
$ kops update cluster --name=${KOPS_CLUSTER_NAME} --yes

Validate your cluster with the command below -

$ kops validate cluster

You should proceed further only after you see that your cluster is ready. This was also one of the pain points for me as I wasn’t sure why my clusters were not getting created properly. It turned out that it was due to AWS availability in the “us-east-1c” zone. I switched to “us-east-2c” and all was good.

9. Install Helm. Helm is your package manager for Kubernetes. We will install Dask on the kubernetes cluster we created using helm. Installing helm is pretty straight forward. Zero-to-jupyterhub docs do a great job.

$ curl https://raw.githubusercontent.com/kubernetes/helm/master/scripts/get | bash

Now setup tiller on your kubernetes cluster

$ kubectl --namespace kube-system create sa tiller
$ kubectl create clusterrolebinding tiller --clusterrole cluster-admin --serviceaccount=kube-system:tiller
$ helm init --service-account tiller

Check if helm is installed correctly

$ helm version
Client: &version.Version{SemVer:"v2.13.1", GitCommit:"618447cbf203d147601b4b9bd7f8c37a5d39fbb4", GitTreeState:"clean"}
Server: &version.Version{SemVer:"v2.13.1", GitCommit:"618447cbf203d147601b4b9bd7f8c37a5d39fbb4", GitTreeState:"clean"}

10. Install Dask on your cluster. The official docs have a good explanation on how to install a basic cluster. Here is the command

$ helm install stable/dask

The install gives a name to your deployment and you need to use this name if you want to upgrade or later delete your helm install. To validate if your install was successful and your pods are running, use the following command

$ kubectl get pods

Here is a screenshot of the stages the pods go through

You can see here that the deployment gave my cluster the name of billowing-labradoodle and the status is running.

The thing which tripped me the most was the docs saying “By default, the Helm deployment launches three workers using two cores each and a standard conda environment.” So it took me a while to get to a Dask cluster which had 8 workers with 64 cores. The basic install does create a cluster with 1 jupyter server, 1 scheduler and 3 worker nodes, but the size of the worker nodes is based on your original node size you provisioned in step 8. After a few permutations, I realized I just want to say I want 8 replicas of the workers with my default kubernetes config for cpu and memory the way I set it up (8 cpu/32gig ram) for each worker. Another issue I faced was that the install has basic anaconda installed and thats it. So if you want to use s3 or Google cloud storage you need to have either s3fs or gcsfs installed on all your nodes of your cluster, including your scheduler, jupyter server and worker nodes. These changes can be made easily by overriding the settings used by helm chart by writing your own config.yaml file. Here is my config.yaml file

# config.yaml
worker:
  replicas: 8
  env:
    - name: EXTRA_CONDA_PACKAGES
      value: numba xarray -c conda-forge
    - name: EXTRA_PIP_PACKAGES
      value: gcsfs s3fs dask-ml --upgrade
# We want to keep the same packages on the worker and jupyter environments
jupyter:
  enabled: true
  env:
    - name: EXTRA_CONDA_PACKAGES
      value: numba xarray matplotlib -c conda-forge
    - name: EXTRA_PIP_PACKAGES
      value: gcsfs s3fs dask-ml --upgrade

Now update your helm install with this command

$ helm upgrade billowing-labradoodle stable/dask -f config.yaml

Check your pods and make sure the status says they are all Running with the command “kubectl get pods”.

Now run the following command to get the external IP for your jupyter server and scheduler server. I have shortened the external-ip names so you can see what the result looks like.

$ kubectl get services
NAME                                   TYPE           CLUSTER-IP       EXTERNAL-IP            PORT(S)                       AGE
billowing-labradoodle-dask-jupyter     LoadBalancer   100.xx.xxx.93    a251--.amazonaws.com   80:32431/TCP                  1h
billowing-labradoodle-dask-scheduler   LoadBalancer   100.xx.xxx.136   a026--.amazonaws.com   8786:32624/TCP,80:31475/TCP   1h
kubernetes                             ClusterIP      100.xx.x.1       <none>                 443/TCP                       1h

Now copy paste the external IP into two browser windows. You should see the Jupyter hub login page and Scheduler page show up. If it doesn’t show up, just be patient and wait a few minutes. Typically the Scheduler page comes up pretty fast and Jupyter hub page takes a while. Eventually you will be greeted by the login page.

11. Finally Running the New York Taxi Example. Type in ‘dask’ as password for jupyterhub and login. Go into the examples folder and open the 05-nyc-taxi.ipynb notebook. If you were to run this now, it will run.. but it will be really slow and no where close to how fast the demo ran in the pycon video. It eventually occured to me that the given example is using csv files from google storage, whereas all our infrastructure we setup is on AWS. Of course there will be a huge delay! These files need to go over the network between two providers and to top it off, these files are huge.. about close to 2 gigs each. After I realized my mistake, I changed the code to use files from s3 instead and finally the notebook ran as fast as the demo!! Here is the code you need to change to use s3 instead of google cloud (also on MattMatthew Rocklin’s article). The thing to notice in the screenshot below is the number of Workers, the number of Cores and the total Memory across all the workers in step 3 -

from s3fs import S3FileSystem
s3 = S3FileSystem(anon=True)
s3.ls('dask-data/nyc-taxi/2015/')
import pandas as pd
with s3.open('dask-data/nyc-taxi/2015/yellow_tripdata_2015-01.csv') as f:
    df = pd.read_csv(f, nrows=5, parse_dates=['tpep_pickup_datetime', 'tpep_dropoff_datetime'])
df
import dask
from dask.distributed import Client, progress
client = Client()
client
import dask.dataframe as dd
df = dd.read_csv('s3://dask-data/nyc-taxi/2015/*.csv',
                parse_dates=['tpep_pickup_datetime', 'tpep_dropoff_datetime'],
                storage_options={'anon': True})
df = df.persist()
progress(df)

The rest is the same. Hope this helps somebody!

Send Email In Python Bugfix

2018-02-22T11:20:00-05:00

I have been using this python function for a while to send html emails from within my python scripts -

def sendhtmlemail(emailbody, subject):
    """ Send html email with given subject and body """
    fromaddr = readconfig(EMAIL_FROM)
    toaddr = readconfig(EMAIL_TO)

    msg = MIMEMultipart('alternative')
    msg['From'] = fromaddr
    msg['To'] = toaddr
    msg['Subject'] = subject

    body = MIMEText(emailbody, 'html')

    msg.attach(body)


    s = smtplib.SMTP('localhost')
    s.sendmail(fromaddr, toaddr, msg.as_string())
    s.quit()

The emailbody is my html body I want in my email which may also include css in it. the variable EMAIL_FROM and EMAIL_TO are global variables which stored the file path where I could have the from email address and to email address. If I want to send a plain text email, I use the following function -

def sendmail(subject, body):
    """ Send email with given subject and body """
    fromaddr = readconfig(EMAIL_FROM)
    toaddr = readconfig(EMAIL_TO)

    msg = MIMEMultipart()
    msg['From'] = fromaddr
    msg['To'] = toaddr
    msg['Subject'] = subject
    msg.attach(MIMEText(body, 'plain'))

    s = smtplib.SMTP('localhost')
    s.sendmail(fromaddr, toaddr, msg.as_string())
    s.quit()

I was recently using the sendhtmlemail function to send an email to multiple addresses. All I did was store the multiple emails as a comma separated string in the EMAIL_TO file and found out that it was not being send to all the recipients. It was only sent to the first recipient and the person who got it could see all the other people the email was sent to.. Strange!

After poking around and reading the official docs, I found out that the second parameter in the smtplib.SMTP’s sendmail function takes a list of RFC 822 to-address strings (a bare string will be treated as a list with 1 address). DOH!

I fixed this by splitting the toaddr variable on the comma and storing it in a new list and feed the list to the sendmail function! Here is my final plain text sendmail version -

def sendmail(subject, body):
    """ Send email with given subject and body """
    fromaddr = readconfig(EMAIL_FROM)
    toaddr = readconfig(EMAIL_TO)

    msg = MIMEMultipart()
    msg['From'] = fromaddr
    msg['To'] = toaddr
    msg['Subject'] = subject
    msg.attach(MIMEText(body, 'plain'))
    toaddr_list = [x.strip() for x in toaddr.split(',')]

    s = smtplib.SMTP('localhost')
    s.sendmail(fromaddr, toaddr_list, msg.as_string())
    s.quit()

Lost SSH During Ubuntu Update

2018-01-24T22:00:00-05:00

I was trying to install security updates on my digitalocean vps and lost my ssh connection. The one time I was not using tmux! GAH!! It was also stuck at a stage where it was waiting for me to put in what config changes to use for some software package (use my previous or use the default)

When I logged in again and tried to run the apt-get update again, I saw the following error -

E: Could not get lock /var/lib/dpkg/lock - open (11: Resource temporarily unavailable)
E: Unable to lock the administration directory (/var/lib/dpkg/), is another process using it?

So had to find my running process and kill it. Even after killing it, I was seeing the same error. Turns out I can resume my upgrade from where my ssh connection died, using the following command -

% sudo dpkg --configure -a

How to install MongoDB on a Mac

2016-02-02T00:00:00-05:00

I am starting to teach myself the MEAN stack. The MEAN stack makes use of MongoDB, Express.js, Angular.js and Node.js. The best part about this stack is that all parts of the stack can be written in Javascript. Read this wikipedia article to know more.

The best way to install mongodb is using homebrew. First make sure homebrew is upto date and there are no issues

$ brew update
$ brew doctor

Then install mongodb -

$ brew install mongodb

Once mongodb is installed, create a data directory where mongodb will create all its database files -

$ sudo mkdir -p /data/db

Now change ownership to your account so you dont have to use sudo to run mongodb-

$ sudo chown `id -u` /data/db

Thats it! Now run the ‘mongod’ command to run your instance of mongodb without using sudo.

How to install npm packages on osx without sudo

2016-02-02T00:00:00-05:00

One of the best features of python is how you can have project specific python environments using virtualenv. That way you are not putting any python packages in the global area. I knew I wanted to do something similar for managing and installing npm packages. Basically I did not want to install the packages in the global area which could mean I would have to use sudo to install.

Reading up on the internet, I found that npm and node can cause a lot of headache if not setup correctly.

I followed the following links to setup my environment http://www.johnpapa.net/how-to-use-npm-global-without-sudo-on-osx/ https://gist.github.com/DanHerbert/9520689 https://github.com/sindresorhus/guides/blob/master/npm-global-without-sudo.md

Pindi Chole

2016-01-24T10:00:00-05:00

This is my aunt Leela Mangu’s special pindi chole recipe.

Boil 2 cups Chana with 2 tsp ground tea leaves, 2-3 Laung, 2 tsp amla pdr, salt.

Fry in oil 1 inch ginger, some garlic till brown, add 2 small tomatoes chopped to small pieces. Add Haldi, little jeera pdr n dhaniya pdr (1/2 tsp), chole masala- 1.5 tsp. When oil starts to leave the masala, add a little of the boiled channa (mashed somewhat.) Add 1 tsp chai leaves. Add rest of the Chana, let it cook. Then Add anardana 2 tsp, mix well and garnish.

sambar

2016-01-24T00:00:00-05:00

This is a recipe I got from my mother-in-law. The cool part about this recipe is that it is made without using any sambar powder.

Ingredients

Imli - nimbu size
Vegetables -any of the veggies drumstick, bindi, shimla mirch, red pumpkin, white peta, onion
coconut - Half coconut- grated
Masala - chana dal, sabud dhaniya, pepper, methi, hing, sabud mirch
Toor daal
Tadka - rai, methi, kadi patta and sabud mirch

Method

Soak imli and microwave and cool
Sqeeze out imli water
Boil any two or three vegetables in imli water
Add haldi and namak while boiling
Pressure cook toor dal with haldi
Mash the toor dal and add it to the cooked vegetable.
Roast and grind the masala with coconut and add to the sambhar
after it has boiled for some time add tadka

sambar

2016-01-24T00:00:00-05:00

This is a recipe I got from my mother-in-law. The cool part about this recipe is that it is made without using any sambar powder.

Ingredients

Imli - nimbu size
Vegetables -any of the veggies drumstick, bindi, shimla mirch, red pumpkin, white peta, onion
coconut - Half coconut- grated
Masala - chana dal, sabud dhaniya, pepper, methi, hing, sabud mirch
Toor daal
Tadka - rai, methi, kadi patta and sabud mirch

Method

Soak imli and microwave and cool
Sqeeze out imli water
Boil any two or three vegetables in imli water
Add haldi and namak while boiling
Pressure cook toor dal with haldi
Mash the toor dal and add it to the cooked vegetable.
Roast and grind the masala with coconut and add to the sambhar
after it has boiled for some time add tadka

Bellam Menthikai

2015-12-09T10:00:00-05:00

This is my aunt Suguna Rao’s recipe. This is a quick way of making this recipe. This does not involve drying the mango in the sun.

Ingredients

One raw green mango (remove skin and cut into chips) - 2 cups
Gud - 1/2 cup
Salt - 2 tsp
Mirchi Powder - 1 tsp
Menthi podi - 2 pinch

Method 1

Mix the cut mango with gud, salt and mirchi powder. Cover and keep aside
Next day add 2 pinches of menti podi and mix really well.

The above recipe does not last long and you need to finish it up quickly. If you want to make a bigger batch and want it to last longer you need to put it out in the sun.

Method 2

Mix the cut mango with gud and salt
keep it out in the sun everyday till you see it oozing a syrup
Add the mirchi powder and menti podi and mix well

Coconut Chutney

2015-12-09T09:30:00-05:00

Coconut chutney goes great with Idli’s, Dosa’s and vadda’s. One recent thread on my family’s recipes group on facebook had a couple of variations on this traditional chutney. I will try to post all the different styles on making this dish.

M.S. Leela Method

Ingredients - Take half a coconut two or three green chillies less than half teaspoonful of lal mirch powder, 5-6 curry leaves A pinch of hing Tamarind or lime juice per taste

Method - Add all above and make a smooth paste in a grinder/Mixie. Add salt to taste and finally add a tadka of mustard seeds and curry leaves

Variation to above method from my aunt Suguna Rao - Add one table spoon of yogurt while making paste in the grinder.

Sita Rao Method

Take equal quantity of grated coconut and fried chutney dal. Add green chillies, salt and grind to smooth paste. Add a tempering of mustard seeds, jeera, udad dal and dried red chillies.

variation - add raw onion and ginger while grinding the smooth paste.

Mukkala Pulusu

2015-12-09T08:45:00-05:00

My relatives started a new group on facebook to share family recipes. I thought I would start writing them down on my blog so they are searchable. The first recipe is a Andhra dish called Mukkala Pulusu. Some other names it is known as are ‘Theepi Pulusu’, ‘Dappalam’ and ‘Meetha Pulusu’. This recipe is from M.S. Leela. She is one of my Dad’s elder sisters. Here is the recipe -

Ingredients

Pumpkin - 250gms
Eggplant - 250gms
Okra/Ladies finger - 250gms
Optional vegetables - Gheeya/Sweet Potato/Carrots
Gud - one big ladoo (1 fist)
Besan
Green Koriander/Hara Dhaniya
Green Chillies
Salt
Haldi
Imli Water - 1/2 cup
Tadka/Tempering - jeera, Mustard seeds, Urad Dal, Meethi seeds, Dried Red Chillies, Kaddi Patta, Hing

Recipe

Boil all the vegetables in 3 glasses of water along with dhaniya, hari mirch, haldi, salt and gud in a pot.
Boil imli and add the pulp to the boiled vegetables.
Roast besan; cool it and stir in half a cup of water.
While Stirring the pulusu add the besan mix.
Temper all the ingredients meant from tempering from above list of ingredients (item 12) and add to the pot.
In the end, add more hara dhaniya and serve it hot.

Adding new Content Using Pelican and Github

2015-10-05T14:00:00-04:00

I am super excited to finally get this end to end workflow to work. I am writing this post from Github, which in turn causes travis-ci to do a pelican build and finally push the build to shireenrao.github.io and also pushes the build to my personal blog.

I will follow this up with a detailed how-to soon.

Kadai Gosht

2015-05-10T22:34:00-04:00

This dish is a very popular dish from North India. I have had this at restaurants and Dhaba’s in and around New Delhi. So how is this different than all the different types of mutton I have made in the past. Well.. I did some research and found this recipe and tried it today. It came out really well!! I am really satisifed with the results and all I can say it’s all in the execution. Here is the recipe -

Ingredients

Goat meat - 2 1/2 pounds in 1 to 2 inch chunks
Onions - 2 big red onions sliced
Tomatoes - 3 Vine tomotoes sliced
Ginger and Garlic paste - 3 tbls
Chili Powder - 1 tsp (2-3 if you want spicy)
turmeric powder 1/2 tsp
Cumin POwder - 2 tsp
Coriander Powder - 3 tsp
Salt - to taste
Crushed Dried Fenugreek leaf - 1 tbls
Green Chilies - 4-5 (may add 3-4 more if you want it spicy)
Cilantro - 1/2 cup
Yogurt - 1/2 cup
Oil - 1 cup

Recipe

Heat oil in a dutch oven
Add the ginger garlic paste and cook for a minute
Add the goat meat and brown it for maybe 5 minutes.
Add the sliced onions and tomatoes
Add chili powder, cumin powder, coriander powder, turmeric powder, salt and crushed fenugreek leave
Mix all the spices well into the meat, onions and tomatoes.
Now reduce heat to medium and put a lid on the dutch oven
You don’t need to add any water. You will see a lot of liquid starts forming in the dish because of all the ingredients.
After 5 minutes add the yogurt and mix it into the dish.
Keep mixing every 5-10 minutes and make sure nothing sticks to the bottom. Cook for an hour or till the meat is tender.
Once you see the oil oozing at the top, add the green chilies and cilantro and mix it in the curry.
Now you can remove the lid and cook till the gravy reduces to required consistency.
Serve with hot Naan or tandoori rotis

Where is the sqlite file for my Flask app

2015-04-24T13:34:00-04:00

I am really enjoying working with the Python Flask Web Framework. I won a O’Reilly book last year on Flask Web Development by Miguel Grinberg and am finally getting around to reading it and trying out what the book teaches. So far it’s been excellent!

From the start of the book we develop a simple app and add new features as we learn new concepts. I got to the Databases chapter and hit my first roadblock! The book teaches you to use SQLAlchemy for database management which is good. That way you interact with one library and behind the scenes you can have any of the supported databases. For the examples in the book we used sqlite.

Here is the snippet I typed in to create my database instance. This is not a complete working snippet. It just has parts which are relevant to start using SQLAlchemy.

from flask_sqlalchemy import SQLAlchemy

basedir = os.path.abspath(os.path.dirname(__file__))

app = Flask(__name__)
app.config['SQLALCHEMY_DATA_BASE_URI'] =\
    'sqlite:////' + os.path.join(basedir, 'data.sqlite')
app.config['SQLALCHEMY_COMMIT_ON_TEARDOWN'] = True

db = SQLAlchemy(app)

I then define my database models. The book shows us examples on how to create User table and Role table and how they are related. Then to exercise different database operations it shows how you can open a Flask shell and try out the different commands on the command prompt. The flask shell is possible because of the Flask-Script extension. All you have to do is to first install flask-script in your virtualenv

$ pip install flask-script

then in your flask app use the Manager provided by flask_script -

from flask import Flask
from flask_script import Manager

app = Flask(__name__)
manager = Manager(app)

# ...

if __name__ = '__main__':
    manager.run()

This allows you to interact with your flask app on the command line. This also lets you open a python shell while the app is running. Works great for testing purposes.

So with the database instance defined, I opened the python shell to test out the different database operations. The first command you learn is the create_all command. This creates your tables based on the model you defined. Here is my interactive session -

(venv) $ python hello.py shell
>>> from hello import db
>>> db.create_all()

At this point the application directory should have a new file called data.sqlite. This is the name we gave to our database in app.config above. I did not see the file in the application directory. I proceeded with the exercise of creating roles and users and commiting my work. I could query the tables too. I checked the application directory again and again didnt see the file. This was getting wierd!! where is the file???

I turned to google and found a solution on stackoverflow. It turns out that I had a typo in my configuration for the database, because of which my database was created in memory. That was sneaky! I would have preffered an error instead. In the config secriotn, instead of ‘SQLALCHEMY_DATA_BASE_URI’ it should have been ‘SQLALCHEMY_DATABASE_URI’.

Here is the corrected snippet

from flask_sqlalchemy import SQLAlchemy

basedir = os.path.abspath(os.path.dirname(__file__))

app = Flask(__name__)
app.config['SQLALCHEMY_DATABASE_URI'] =\
    'sqlite:////' + os.path.join(basedir, 'data.sqlite')
app.config['SQLALCHEMY_COMMIT_ON_TEARDOWN'] = True

db = SQLAlchemy(app)

Installing homebrew MacVim with YouCompleteMe on Yosemite

2014-12-08T10:53:00-05:00

I was pulling my hair to get my vim setup with YouCompleteMe on my macbook after I upgraded to Yosemite. On Mavericks it was pretty straight forward. The following used to work -

# install xcode and command line tools  
$ xcode-select --install  
# install homebrew  
$ ruby -e "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/master/install)"  
# install python  
$ brew install python  
# install macvim  
$ brew install macvim

Now following instructions on the Vundle Github page to setup Vundle.
Add the reference to YouCompleteMe plugin in your .vimrc file along with other plugins you use. Here is what mine looks like -

set nocompatible " be iMproved, required  
filetype off " required

" set the runtime path to include Vundle and initialize  
set rtp+=~/.vim/bundle/Vundle.vim  
call vundle#begin()  
" alternatively, pass a path where Vundle should install plugins  
"call vundle#begin('~/some/path/here')

" let Vundle manage Vundle, required  
Plugin 'gmarik/Vundle.vim'

Plugin 'altercation/vim-colors-solarized'  
Plugin 'Lokaltog/powerline', {'rtp': 'powerline/bindings/vim/'}  
Plugin 'kien/ctrlp.vim'  
Plugin 'scrooloose/syntastic'  
Plugin 'Valloric/YouCompleteMe'

" All of your Plugins must be added before the following line  
call vundle#end() " required  
filetype plugin indent on " required

Now Install all plugins including YouCompleteMe using Vundle by launching vim and executing :PluginInstall

Complete YCM install by running the following from within the YCM vim bundle -

$ cd ~/.vim/bundle/YouCompleteMe  
$ ./install.sh --clang-completer

Now after doing the above on Yosemite, everytime I opened vi, I would get a python crashed error message pop up.

After looking around for a fix, this post gave me the right direction.
I first have to make sure I have the file -

/Library/Developer/CommandLineTools/usr/lib/libclang.dylib

If the above file is missing, that means command line tools has not been installed.
Run “xcode-select —install” and follow the prompts to install command line tools.

I exported the variable EXTRA_CMAKE_ARGS on terminal -

EXTRA_CMAKE_ARGS="-DEXTERNAL_LIBCLANG_PATH=/Library/Developer/CommandLineTools/usr/lib/libclang.dylib"

Then in the same terminal session running the following -

$ cd ~/.vim/bundle/YouCompleteMe  
$ ./install.sh --clang-completer --system-libclang

That did the trick. MacVim with YouCompleteMe should now be working. CodeCompletion on vim could not get any better :)

How to pip install python packages offline

2014-06-04T17:18:00-04:00

You may wonder why would you even need to do that.. Well in my case, I was using virtualenv for a project I was working on and when it came time to install the needed python packages, I realized I was on a train with no internet connection. I had those packages installed on my system, but not in my virtualenv. That put the kibosh on my work that day. So when I found that you can actually create a kind of a local repository on your computer so you could then pip install from that location.. I was excited!!

Start by deciding where would that location be. I put it as a dot directory in my home folder on my mac. The dot folder makes it a hidden folder so it doesn’t clutter your finder window. I called my location .mypypi. So for instance if you wanted to make sure your virtualenv should have ipython, ipdb and django whenever you create it, you would first need to download it into your pypi repository you created. I would do the following

$ mkdir .mypypi  
$ pip install --download $HOME/.mypypi ipython ipdb django  
Downloading/unpacking ipython  
Downloading ipython-2.1.0-py2-none-any.whl (2.8MB): 2.8MB downloaded  
Saved ./.mypypi/ipython-2.1.0-py2-none-any.whl  
Downloading/unpacking ipdb  
Downloading ipdb-0.8.zip  
Saved ./.mypypi/ipdb-0.8.zip  
Running setup.py
(path:/private/var/folders/qk/qk0fnpvd7jg3z8cp48ph2pvw0000gn/T/pip_build_shireenrao/ipdb/setup.py)
egg_info for package ipdb

Downloading/unpacking django  
Downloading Django-1.6.5-py2.py3-none-any.whl (6.7MB): 6.7MB
downloaded  
Saved ./.mypypi/Django-1.6.5-py2.py3-none-any.whl  
Downloading/unpacking gnureadline (from ipython)  
Downloading gnureadline-6.3.3-cp27-none-macosx_10_9_x86_64.whl
(132kB): 132kB downloaded  
Saved ./.mypypi/gnureadline-6.3.3-cp27-none-macosx_10_9_x86_64.whl  
Successfully downloaded ipython ipdb django gnureadline  
Cleaning up...

The above command will also make sure all dependencies are also downloaded. The command also will take a requirements file. Just put “-r requirements.txt” instead of the package name in the above pip install command.

Now when you are in your virtualenv and are ready to install your packages, this is all you have to do -

$ pip install --no-index --find-links=file:/$HOME/.mypypi ipython
ipdb django  
Ignoring indexes: https://pypi.python.org/simple/  
Downloading/unpacking ipython  
Downloading/unpacking ipdb  
Running setup.py
(path:/Users/shireenrao/mydev/python/temp/build/ipdb/setup.py) egg_info
for package ipdb

Downloading/unpacking django  
Downloading/unpacking gnureadline (from ipython)  
Installing collected packages: ipython, ipdb, django, gnureadline  
Running setup.py install for ipdb

Installing ipdb script to /Users/shireenrao/mydev/python/temp/bin  
Successfully installed ipython ipdb django gnureadline  
Cleaning up...

Boom!! It’s done!!

Again, the above command also works with “-r requirements.txt” instead of the python package name.

How to enable upload of files over 2MB in Wordpress running on nginx

2014-06-03T16:04:00-04:00

I was writing a post on a chicken dish I cooked yesterday. As I was trying to upload some pictures, I saw that the upload failed because the files were over 2MB. This was the first time I was posting pictures on my wordpress since I move to Digital Ocean as my hosting company. I knew this was a limit on the php configuration on the server, so it was an easy fix.

I logged in to my server and edited my php.ini located at /etc/php5/fpm/php.ini. All I had to do was to change -

upload_max_filesize = 2M

upload_max_filesize = 8M

Now restart the following services -

sudo service php5-fpm reload  
sudo service nginx reload

I went back to trying to upload the image. This time, the upload started to happen but I eventually saw another failure. Turns out, this time it was from Nginx. I found that nginx throws a 413 error if a file upload exceeds its default limit.

All you have to do to fix that is to over-ride the default in your sites config file. In my case I add the following in my default configuration at /etc/nginx/sites-available/default in the server block right under server_name -

client_max_body_size 8M;

Restart the services again-

sudo service php5-fpm reload  
sudo service nginx reload

The picture uploads worked this time :)

It’s been great with Digital Ocean so far.. and am really loving them!! Highly recommend using them. These guys have great documentation and their IRC channel on freenode are really helpful too!

Yummy Chicken Adraki

2014-06-03T15:10:00-04:00

This dish came out really good!! Very tasty if I can say so myself. It is not very different from my previous effort which you can read here. There were slight changes, so I am writing it down so I remember what I did and of course to share it with everybody.

Ingredients -

Ghee - 2 tbsp
Green Chillies - 3-5 depending on how spicy you want
cumin - 1 tsp
Chopped Onions - 2
Ginger Garlic Paste - 3 tsp
Dhaniya powder - 2 tsp
Garam Masala - 1 tsp
Turmeric Powder - 1/2 tsp
Kashmiri Mirch - 1/4 to 1/2 tsp depending on how spicy you want
Julienned Ginger - from a 1 and half inch of ginger
Cilantro - 1/2 cup chopped
Tomatoes - 2 chopped
Kasoori Methi - 2 tsp crushed
Salt - to taste

Marinate 2 pounds of Chicken thighs(cut bite sized) for at least 30 minutes -

2 tbl spoons ginger paste
2 tsp Black pepper powder
1/2 tsp Turmeric
1 tsp Salt
Juice from 1 lemon

Method -

heat two table spoons of ghee in a thick bottomed pan. Once the ghee is hot, add 3-4 slit green chillies. Once it is slightly cooked, take them out with a slotted spoon and save for later.
Add 1 tsp cumin seeds to the hot ghee.
Once cumin starts sputtering, add the chopped onions and sauté. Add a tsp of salt to help with the caramelizing.
Once the onions are slightly browned, add the 2 tsp of fresh ginger garlic paste and mix it well into the onions and sauté for another 2 minutes
Now add all the powdered spices.. 2 tsp of Dhania powder, 1 tsp of garam masala, 1/2 tsp Kashmiri mirch and 1/2 tsp turmeric powder. Mix all the spices into the onions.
Add the sautéed green chillies and chopped Chicken and mix it well into the onions. Saute the chicken so all the pink goes away. It may take 3-5 minutes.
Add the julienned ginger and chopped cilantro and mix it well into the onions and chicken and sauté some more till you see some oil releasing around the edge of the vessel
Add the 2 chopped tomatoes into the vessel and mix it well into the chicken, scraping the bottom making sure nothing is getting stuck.
Lower the heat and cover the vessel. Check every two minutes and mix the chicken and make sure nothing is sticking at the bottom.
Once tomatoes loose their structure and you see a nice consistent gravy, crush 2 tsp of Kasoori Methi between your palms and add to the dish. Mix it well
Taste to make sure salt is right
Check to see if chicken is cooked
Serve with hot chapati’s or lachedar parathas

Here is a nice close up -

How to Reset Django’s admin password

2014-06-01T22:29:00-04:00

I was working on a django application and because of getting busy at work and home, the project took a backseat. After getting back to the project, I realized I don’t remember what was the username/password I used to create the admin account.

Here is what I did. I started a django shell session (% is my prompt) -

% python manage.py shell  
Python 2.7.6 (default, Apr 9 2014, 11:48:52)  
[GCC 4.2.1 Compatible Apple LLVM 5.1 (clang-503.0.38)] on darwin  
Type "help", "copyright", "credits" or "license" for more information.  
(InteractiveConsole)  
>>> from django.contrib.auth.models import User  
>>> users = User.objects.all()  
>>> users  
[<User: admin>]

I see I had created a user called admin. Now to change the password is easy..

>>> u = users[0]  
>>> u  
<User: admin>  
>>> u.set_password('somepassword')  
>>> u.save()

As I just had one user, this was easy enough.. if you have a lot of users, you could get the individual users object like this -

u = User.objects.get(username__exact='someusername')  
u.set_password('somepassword')  
u.save()

How to install numpy scipy pandas matplotlib and scikit-learn on Mavericks

2014-05-19T23:08:00-04:00

These are all excellent data analysis tools. I cam across some python tutorials on data sciences and wanted to follow the examples from the video.. so went ahead and installed these python packages. This post is here for future reference on the hurdles I had and how I solved them and maybe help somebody else.

1. Make sure you have Xcode command lines tools. You can install that by -

xcode-select install

2. Install Homebrew if you don’t already have it. As simple as just running this on your command line -

ruby -e "$(curl -fsSL https://raw.github.com/Homebrew/homebrew/go/install)"

3. Run the following to make sure homebrew is installed properly. You will be suggested to change your path so /usr/local/bin has priority

4. Install homebrew python -

brew install python  
which python

you should see ‘/usr/local/bin/python’ when you execute which python

4. Now begin installing packages one at a time

brew install gfortran  
pip install numpy  
pip install scipy  
pip install pandas

5. The next package I tried was matplotlib, but that failed with an error about a file now found.. here is a part of what I saw

pip install matplotlib  
# lots of install details here...  
/usr/X11/include/ft2build.h:56:10: fatal error:
'freetype/config/ftheader.h' file not found  
#include <freetype/config/ftheader.h>  
^  
1 error generated.  
error: command 'clang' failed with exit status 1  
----------------------------------------

so I installed freetype from homebrew and retried installing matplotlib -

brew install freetype  
pip install matplotlib

Unfortunately, I got the same error. I saw a solution on stack overflow to create a symlink to the freetype directory in the expected location.

sudo ln -s /usr/X11/include/freetype2/freetype/
/usr/X11/include/freetype  
pip install matplotlib

It worked this time. It’s a hack to do it this way.. just remember you did this. If the package gets updated later, just remember to remove this link and try installing the normal way.

6. Install scikit-learn

pip install scikit-learn

7. In case you want to play with iPython notebook, you would need the following -

pip install ipython  
pip install tornado  
pip install pyzmq  
pip install jinja2

Thats it.. you now have all scientific packages installed!!

You could also take the easier route and install one of the following python distributions -

Anaconda from Continuum Analytics
Enthought Canopy

Happy Data hacking!

YouCompleteMe install fails due to internal compiler error

2014-05-18T15:44:00-04:00

While I was trying to set up my development environment on a vagrant vm, I saw a compiler error while installing the excellent YouCompleteMe vim plugin. Well let me back up a bit..
You install the plugin using Vundle. After the plugin is installed you need to run a script to compile YCM with semantic support for C-family languages:

cd ~/.vim/bundle/YouCompleteMe  
./install.sh --clang-completer

My install was failing with this wierd error -

[ 75%] Building CXX object
BoostParts/CMakeFiles/BoostParts.dir/libs/thread/src/pthread/thread.cpp.o  
Linking CXX static library libBoostParts.a  
[ 75%] Built target BoostParts  
Scanning dependencies of target ycm_core  
[ 77%] Building CXX object
ycm/CMakeFiles/ycm_core.dir/CustomAssert.cpp.o  
[ 77%] Building CXX object ycm/CMakeFiles/ycm_core.dir/Utils.cpp.o  
[ 78%] Building CXX object
ycm/CMakeFiles/ycm_core.dir/ClangCompleter/ClangCompleter.cpp.o  
c++: internal compiler error: Killed (program cc1plus)  
Please submit a full bug report,  
with preprocessed source if appropriate.  
See <file:///usr/share/doc/gcc-4.8/README.Bugs> for instructions.  
make[3]: ***
[ycm/CMakeFiles/ycm_core.dir/ClangCompleter/ClangCompleter.cpp.o] Error
4  
make[2]: *** [ycm/CMakeFiles/ycm_core.dir/all] Error 2  
make[1]: *** [ycm/CMakeFiles/ycm_support_libs.dir/rule] Error 2  
make: *** [ycm_support_libs] Error 2

Looking for a solution to the problem, I came across the same issue on github. Basically I was running out of memory on my VM while compiling the plugin. Somebody hinted on adding swap space would help. So I used this script to add swap file -

#!/bin/sh

# size of swapfile in megabytes  
swapsize=512

# does the swap file already exist?  
grep -q "swapfile" /etc/fstab

# if not then create it  
if [ $? -ne 0 ]; then  
    echo 'swapfile not found. Adding swapfile.'  
    fallocate -l ${swapsize}M /swapfile  
    chmod 600 /swapfile  
    mkswap /swapfile  
    swapon /swapfile  
    echo '/swapfile none swap defaults 0 0' >> /etc/fstab  
else  
    echo 'swapfile found. No changes made.'  
fi

# output results to terminal  
cat /proc/swaps  
cat /proc/meminfo | grep Swap

Now go ahead and compile YouCompleteMe to Success :)

You can delete the swap file if you don’t need it. Here use this script -

#!/bin/sh

# does the swap file exist?  
grep -q "swapfile" /etc/fstab

# if it does then remove it  
if [ $? -eq 0 ]; then  
    echo 'swapfile found. Removing swapfile.'  
    sed -i '/swapfile/d' /etc/fstab  
    echo "3" > /proc/sys/vm/drop_caches  
    swapoff -a  
    rm -f /swapfile  
else  
    echo 'No swapfile found. No changes made.'  
fi

# output results to terminal  
cat /proc/swaps  
cat /proc/meminfo | grep Swap

I got the scripts from here.

How to capture shell script output to a logfile and standard output

2014-04-28T15:44:00-04:00

I wanted the ability to review the output from a shell script. I could always re-direct output to a log file, but then I don’t see where in the script I am currently at. I can always open a second terminal and then tail the lot file I used in the first place.

Well there is another way.. in short..

$ command 2>&1 | tee -a logfile

This works on a bash shell. I tried to get this to work on csh or tcsh.. but get an error.

How to make a bootable install of Mavericks on USB Flash Drive

2014-04-15T15:12:00-04:00

Your USB drive should be called Untitled and formatted as Mac OS Extended (Journaled). The installer should be called Install OS X Mavericks.app and should be in your Applications folder.

Run the following command to prepare your USB to be a bootable installer, Don’t worry.. it takes 20-25 minutes to finish..

$ sudo /Applications/Install\ OS\ X\ Mavericks.app/Contents/Resources/createinstallmedia --volume /Volumes/Untitled --applicationpath /Applications/Install\ OS\ X\ Mavericks.app --nointeraction

You should see something like this:

Erasing Disk: 0%... 10%... 20%... 100%...  
Copying installer files to disk...  
Copy complete.  
Making disk bootable...  
Copying boot files...  
Copy complete.  
Done.

You can then boot up from the USB by holding down the option key, then install the Mavericks from the USB.

*Source

Desi Murgh Curry

2014-04-13T23:35:00-04:00

It has been ages since I have posted any recipes and I had to post this dish as it is something special. It’s not very complicated and is very tasty. This is a North Indian Chicken curry. It is better to cook this with chicken on the bone. I do it with chicken drum sticks and wings. Turns out fabulous!!

Here are the ingredients -

Large onions, sliced - 2
Ginger and Garlic paste - 2 Table Spoons
Tomatoes, peeled and chopped
Red chilli powder 1 tsp
Turmeric - 1/2 tsp
Cumin Seeds - 1 tsp
Coriander Powder - 3 tsp
Black Cardamoms - 3
Cloves - 5-6
Bay leaf - 1
Salt to taste
Yogurt - 1/2 cup
Chopped cilantro

Method -

1. Heat 1/4 cup of cooking oil in a thick bottom vessel. Add the sliced onions and cook till brown. Remove the onions and make a fine paste of the fried onions.
2. Add a little more oil to the vessel and heat. Add ginger garlic paste and stir for a minute or so.
3. Add the chopped tomatoes and stir it well. a couple of minutes later, add the yogurt and cook for 4-5 minutes and keep mixing.
4. Add the browned onions paste and mix it well into the vessel.
5. Add all the spices along with salt. Cook and stir the dish until oil separates out.
6. Add the chicken pieces and mix the masala into the chicken.
7. Add 2-3 cups of water and cover up the vessel and reduce the heat to low heat.
8. Stir occasionaly and cook for 45-50 minutes till chicken is cooked completely and tender.
9. Remove the lid and cook for another 10 minutes.. you should see the oil in the sauce separate out. Add the chopped cilantro and mix that into the curry
10. Serve with hot naan or rice!!

Here are some more pics I took with my 50mm lens :)

Change wordpress wp_mail using a plugin

2014-04-13T21:53:00-04:00

Hosting wordpress on bluehost has been nice except for one problem. All notifications from my wordpress come out weird. The domain name of the sender does not show up as my domain name. It shows up as username@box###.BlueHost.com as explained here.

I had figured out a way to fix this by editing some core wordpress scripts and had shared my method in a previous post. It was working out fine, except for one small problem.. Every time wordpress got updated, I would loose my fix and I would have to make those changes again.

I then found out, I could use filters to over-ride two variable wp_mail_from and wp_mail_from_name which would do exactly what I wanted. I then did some digging to understand how wordpress plugins work to write my own plugin. I also wanted the ability to have my own plugin settings page, just so I can set the email address and name and have it saved in the internal wordpress database instead of having it hard coded in the plugin. These two excellent tutorials on plugins helped me a lot - here and here.

Thus was born the plugin email-over-rider. Just download the zip and upload the plugin and activate it. Then go to the Settings area in your wordpress admin section and you should see the Email Over-rider sub menu. Just set the required email address and Name and you are set. No need to update the wordpress code everytime wordpress gets updated.

If plugins are not your cup of tea, you could also just add the filters to your themes functions.php file. The entire method is explained very nicely here.

How to import VMWareVM file from a backup in VMWare Fusion

2013-09-03T23:05:00-04:00

I recently re-installed Mountain Lion on my mac book pro. There were a couple of reasons why I did that. Mainly because I wanted get rid of macports and try out homebrew. The reason I wanted to do that was because I am trying to setup a build and package process for my PyQt applications, and macports based pyqt was not working out. Doing a setup from scratch is a lot of work but it’s fun for me.

In my previous setup I had two VM’s, one was a mountain lion VM and the other was a Snow Leopard VM for VMWARE. I had backed these two vmwarevm files on my time machine. After re-installing vmware fusion I restored the vmwarevm files in its default location.

I was expecting to see the two vm’s in my library, but that wasn’t the case. Then I thought maybe I need to import these in. Clicking on File->Import… did not work either. I couldn’t even select the vmwarevm files as shown in the image below -

I thought I would loose the vm’s and would have to re-create these from scratch :(

I tried to google for this, but couldn’t find any solution for it. I don’t know what prompted me, but I remember I saw that the icon on the two vm’s was that for fusion. I then double clicked on one of the VM’s and viola!! It popped in my vmware library and started up!! Hurrah!!

That was easy!!

How to update individual packages in the enthought Python Distribution

2013-08-20T16:28:00-04:00

I am recently came across a youtube video about data analysis using Python by Wes McKinney. So to learn a new library I thought I would follow the tutorial. As I already had Enthought Python distribution installed on my macbook, I thought I had all the prerequisites. Sadly that was not the case. I had to update some of the packages. Luckily epd comes with its own package manager called enpkg. This is what I did to update python, numpy and pandas -

$ sudo enpkg enstaller  
Password:  
prefix: /Library/Frameworks/Python.framework/Versions/7.3  
enstaller-4.6.2-1.egg [fetching]  
267 KB
[.................................................................]  
enstaller-4.5.3-1.egg [removing]  
607 KB
[.................................................................]  
enstaller-4.6.2-1.egg [installing]  
765 KB
[.................................................................]  
$ sudo enpkg ipython  
prefix: /Library/Frameworks/Python.framework/Versions/7.3  
ipython-0.13.1-2.egg [fetching]  
2.55 MB
[.................................................................]  
ipython-0.12.1-3.egg [removing]  
7.31 MB
[.................................................................]  
ipython-0.13.1-2.egg [installing]  
7.42 MB
[.................................................................]  
$ sudo enpkg pandas  
prefix: /Library/Frameworks/Python.framework/Versions/7.3  
MKL-10.3-1.egg [fetching]  
34.74 MB
[.................................................................]  
numpy-1.7.1-2.egg [fetching]  
3.31 MB
[.................................................................]  
pandas-0.12.0-1.egg [fetching]  
3.97 MB
[.................................................................]  
numpy-1.6.1-0.egg [removing]  
10.97 MB
[.................................................................]  
MKL-10.3-1.egg [installing]  
106.08 MB
[.................................................................]  
numpy-1.7.1-2.egg [installing]  
11.58 MB
[.................................................................]  
pandas-0.12.0-1.egg [installing]  
14.50 MB
[.................................................................]

Run a macro in all buffers in vim

2013-08-12T17:50:00-04:00

I knew vim was powerful, but realized recently that with combined command line tools it is really really powerful!!

Our Oracle production database was having issues at work. As a way to find out the cause, our group was asked to switch our jboss instance to point to a different node on the rack server. The switch, in theory is quick and painless. One who has permissions can directly go change the jboss oracle datasources, but in this day and age with all this change control crap, it’s really gotten tough doing such things. All changes at my workplace need to be done by a deployment group.

As the guy who maintains the jboss instances was out, it was upto me to to make this change. All the datasource config files for all environments are stored in one place. It would have been painful to open all the prod ds files and check which ones were pointing to the oracle server. Thats where the grep command came in handy. I can just grep the name of the server (say RACK1) and get a list of files.

$ grep -l "RACK1" *

(The reason I used the -l flag is so I just get the list of files and not get the file contents where the line matched)

You can combine the above command with vim to open all the files in vim

$ vim $(grep -l "RACK1" *)

All the files are now open in different buffers. Now I could just use a macro to make my change on the first file and then replay the macro on all the buffers to make my change on all files at once :)

In command mode on the first buffer -

qa/RACK1^McwRACK2^[q

Breakdown -
qa - Start recording macro
/RACK1^M - Search file for reference for RACK1
cw - Change word to new server name
^[ - back to command mode
q - stop recording

Now to apply this macro on all buffers (all files), run this in command mode -

:bufdo execute "normal @a" | write

The above command is from here
Please note that I have not included how to checkout the files here.

All these changes were done within a minute but took me 2 hours to have the changes applied in PROD. That is a different story :)

replay a vim macro until end of buffer

2013-08-12T16:06:00-04:00

This is an awesome trick I learnt recently. It’s a great trick if you don’t know how many times you need to run your macro. In short it is called a recursive macro :)

The gist is, you first make sure the register you are going to record your macro in is empty. Now while recording your macro call the register you are recording in as the last command.

For example if you are recording your macro in register a -

In command mode, first make sure your register is empty -

qaq

:let @a=""

You can confirm your register is empty by printing the contents of register a by using -

"ap

Now suppose I am working on a python script and need to comment out the script from current line to end of buffer, I would run this in command mode -

qaI#^[j@aq

Here is the breakdown
qa - start recording in register a
I# - Insert # in the beginning of current line
^[ - Thats what gets types when you hit escape to come out of insert mode
j - go down to next line
@a - nothing happens as register a is empty
q - quit recording

When you execute @a again as register a now is set it calls itself and repeats till end of buffer. A very cool trick indeed!!

Another way using the visual mode is by using the following command

VG:normal @a

The problem with the visual mode is that it only works when your macro is only editing on a single line. If you macro spans across lines, the visual mode trick will not work. I have to confess, I still need to fully understand how to use normal in vim commands.

Compile Vim 7.3 with +clientserver feature on Mac OS X

2013-08-08T22:56:00-04:00

I have been using vim for the longest time but I know I am not the most efficient and fastest vim user. My interest in vim recently started after I watched a youtube video on how to use vim as a Python IDE. Here is a link to the video. Here is the link to Martin’s github repo.

Seeing this video, I saw that you can build vim from source and have it built with python features. Below I show how I built it and installed this under my home directory in $HOME/opt/vim. I then added a sym link to my bin directory under my home directory. As the bin directory is in my path, the built version of vim gets invoked.

$ wget wget ftp://ftp.vim.org/pub/vim/unix/vim-7.3.tar.bz2  
$ tar xvjf vim-7.3.tar.bz2  
$ cd vim73  
$ ./configure --enable-pythoninterp --with-features=huge --prefix=$HOME/opt/vim  
$ make  
$ make install  
$ cd ~/bin  
$ ln -s $HOME/opt/vim/bin/vim vim  
$ ln -s $HOME/opt/vim/bin/vi vi

This led me to further learn new features/tips and tricks in using vim. I then found Derek Wyatt’s excellent and entertaining vim tutorials. It has ignited the desire to learn the editor properly. One of the features I found out about was the ability to edit multiple files. These files can be based on some filter on your file system by using the grep command (My next post will talk about this feature and how I saved the day at work by editing files based on grep). Unfortunately the feature Derek talks about needs vim to have the +clientserver feature built in.

Unfortunately the above build did not add that feature, even thought I have —with-features=huge set. After some googling around, I found that all I had to do was configure the build slight differently. I just have to add the flag —enable-gui=gtk2 to the configure command.

$ cd vim73  
$ make distclean  
$ ./configure --enable-pythoninterp --with-features=huge --enable-gui=gtk2 --prefix=$HOME/opt/vim
$ make  
$ make install

Now you have vim with clientserver features. I will have more updates coming about vim soon. Keep posted!

Kerala Chicken

2013-07-07T20:24:00-04:00

This recipe is a variation on the Malabar mutton curry. I am using chicken thighs instead of mutton or lamb. You can replace the chicken with lamb or goat. It’s quick and easy to do.

Ingredients
Chicken thighs - 2 pounds
Onions - 1 chopped
Coriander powder - 1 tbsp
Turmeric powder - 1/2 tsp
Chili powder - 1/2 tsp
Salt to taste

Spice Paste
Grated coconut - 1/2 cup
Fresh ginger - 1 inch
Garlic - 3 cloves
Cinnamon stick - 1 inch
Cloves - 3
Bay leaves - 2
Curry leaves - 10
Black peppercorns - 5

For tempering
Mustard seeds - 1/2 tsp
Curry leaves - 10
Green chilies - 2 slit

Method
Roast the ingredients for spice paste till coconut gets browned. Once it cools down, grind to a fine paste in a food processor by adding water gradually. Be careful and not add a lot of water and make it watery.

Now in a heavy bottom pan heat 3 tbsp of oil. Add your chopped onions and fry for 5 minutes.
Add the spice paste, ground coriander powder, turmeric powder, chili powder and salt. stir the mixture for a minute and then add 1 and 1/2 cup of water.. Bring this to a boil.

Once a boil comes in, add your chicken. Reduce the heat and cook covered for 30 minutes or till the chicken is cooked.

Prepare the tempering mixture in a small frying pan. Heat 1 tbps of oil and add mustard seeds.
Once they start popping, add curry leaves and slit green chilies. Stir-fry for a minute. Add the tempering mixture to your meat curry and continue cooking until curry is very dry and thick.

This should go really well with either rice or parathas!

Here are some more pics…

While roasting the spices for the paste -

Spice Paste -

Cooking to dry out the sauce and get it thick -

The finished dish -

Checkout this quick video of the recipe:

Shell script to download Vimeo Videos

2013-06-27T21:24:00-04:00

I know you can easily download vimeo videos, but I wanted to download all of Derek Wyatt’s excellent vim tutorials for offline viewing in one shot. Ideally downloading all videos directly from the rss feed would be the best option. I could not find any application which did that but found a shell script from http://ossguy.com/?p=841. Unfortunately the script no longer worked :(

I then found another script on github here. Reading that script I saw that it was based on the first script I had found and John Teslade had fixed it up after Vimeo had made changes to their site.

I saved the script in my bin directory which is on my path. I then tested it with a sample video and it worked beautifully. The best part was that the saved video file was named with the title of the video. Perfect! Well almost. This script does not dowmload videos based on a playlist or rss feed.

That’s where I come in :). I opened the Derek Wyatt’s rss feed on vimeo. saved the xml file locally on my computer. I called it vimeo.xml. Stupid thing all came in one line. Never mind… opened the file in vim and used the following commands to beautify it -

:%s/></>\r</g  
:0 
=:$

Now just grep the file and look for the string ” ” and pipe it to a file. Lets call that file playlist.txt

% grep "<link>" vimeo.xml > playlist.txt

The contents of the file looked like this -

<link>http://vimeo.com/user1690209/videos</link>  
<link>http://vimeo.com/user1690209/videos</link>  
<link>http://vimeo.com/15443936</link>  
<link>http://vimeo.com/7133419</link>  
<link>http://vimeo.com/7129986</link>  
<link>http://vimeo.com/7096383</link>  
<link>http://vimeo.com/7035132</link>  
<link>http://vimeo.com/6999927</link>  
<link>http://vimeo.com/6342264</link>  
<link>http://vimeo.com/6332848</link>  
<link>http://vimeo.com/6329762</link>  
<link>http://vimeo.com/6307101</link>  
<link>http://vimeo.com/6306508</link>  
<link>http://vimeo.com/6216655</link>  
<link>http://vimeo.com/6185584</link>  
<link>http://vimeo.com/6170479</link>  
<link>http://vimeo.com/6154082</link>  
<link>http://vimeo.com/6110008</link>  
<link>http://vimeo.com/6093081</link>  
<link>http://vimeo.com/6025010</link>  
<link>http://vimeo.com/6024774</link>  
<link>http://vimeo.com/4457161</link>  
<link>http://vimeo.com/4456458</link>  
<link>http://vimeo.com/4454614</link>  
<link>http://vimeo.com/4449258</link>  
<link>http://vimeo.com/4448635</link>  
<link>http://vimeo.com/4448265</link>

Clearly you can tell the first two are not video links. So just delete those two lines. Then do a global search and replace to remove the strings ” ” and “”. The following vim recipe from http://rayninfo.co.uk/vimtips.html worked for me -

:%s#<[^>]\+>##g

Save and close the file. Now just run the following command to download all the videos -

$ for line in `cat playlist.txt`; do vimeo_downloader.sh $line; done

Maybe some day I will enhance this script to do all this for me.

Kasoori Meethi Chicken finally

2013-06-24T22:25:00-04:00

I made kasoori methi chicken long time ago and posted a picture saying I would post a recipe soon.. The day is finally here. Better late than never right?? :)

This recipe is a slight twist from a different interesting recipe I came across. The slight twist is the addition of Kasoori Methi. In english it is called Fenugreek leaves. I will be using dried fenugreek leaves. I think it is better than using fresh leaves as the dried leaves give a very distinct flavor. You can get this from any Indian store. This dish can be made without the addition of Kasoori Methi and it is still a very tasty dish.

Another reason I thought this dish was interesting was because of the use of fried onions paste. Checkout the ingredients and method to make this dish!

Ingredients

Chicken - a whole chicken around 2 pounds(1.2kg) skinned and cut into 8 pieces
Onions - 2 large finely sliced
Garlic Paste - 2 tsp
Ginger Paste - 2 tsp
Tomatoes - 4 chopped
Yogurt - 1/2 cup
Red Chilli Powder - 1 tsp
Turmeric Powder - 1/2 tsp
Coriander Powder - 2 tsp
Black Cardamom - 3 pods
CLoves - 6
Bay Leaf - 1
Dried Fenugreek leaves - 1 tbsp
Cilantro - 3 tbsp chopped

Method

A point to note about the chicken is that if you don’t like bone in chicken you can make this dish with boneless thighs. Just use 2 pounds of chicken thighs cut into chunks.

Heat about 1/4 cup of cooking oil in a heavy bottom pan. Add the onions and cook till golden brown. Remove the pan from the heat and remove the onions with a slotted spoon and let it cool down. Once it has cooled down use a food processor to make a fine paste.

Return the pan to heat and add garlic and ginger paste. Stir for a minute and then add the chopped tomatoes. Cook the tomatoes for a couple of minutes and then mix in the yogurt. Now cook for 5-6 minute, mixing every couple of minutes.

Mix in the onions paste you made earlier, along with all the spices including the bay leaf, black cardamom and salt to taste. Cook, stirring until the oil separates out.

Now add the chicken pieces and spoon the sauce over the chicken. Add 2 cups of water and mix the water in the sauce. Put the lid on the pan and cook on low heat for 30-35 minutes, stirring occasionally so that the bottom does not stick till the chicken is cooked. Add more water if needed.

Remove the lid. Crush the fenugreek leaves in your hands and add to the pan. Mix it into the sauce. Now cook with the lid open, mixing occasionally till the oil separates out.

Add half of the cilantro to the pan and mix it into the sauce. Garnish with the remaining cilantro and serve this with hot chapati’s or Naan.

Hope you enjoy this Yummy Recipe! Here are some more pics of the dish..

Installing Snow Leopard in VMWare on Mountain Lion

2013-06-24T13:29:00-04:00

I am learning how to build GUI applications in Python using PyQt/PySide. I still have ways to go before I can start building some real applications. My end goal is to be able to learn how to build a mac application and then package it as a disk image (.dmg) like how most mac software is available. I would also like to learn how to create those installer packages (.pkg/.mpkg) eventually.

One of the PySide apps I learnt was a currency converter. I thought, it would be nice if I could create a dmg file as a proof of concept to the entire build process. After facing a lot issues to build this on my Mountain Lion, I finally succeeded. I will be documenting all the issues I faced during this process in my blog.

Once I was able to do a build which worked on my own laptop, I sent the dmg file to my brother to test out. He is still running Lion on his macbook pro, so I thought it would be a good test for the build. The app crashed on launch :(

After trying a couple of things we still could not get this to work. That’s when I decided to run an older version of MAC OSX as a VM on VMWare. That way I can look at console logs and see what failures are coming up and figure out a fix myself. As I have Snow Leopard media, I thought it would be easy to do. Unfortunately NOT!!

Finally got my Snow Leopard VM working after reading this thread on MacRumors. I am capturing all that info here just to keep a copy. All credit goes to the original source. I will document my Pyside Build ordeal in a separate blog post.

Step One - Preparing the Snow Leopard Installation DVD

1. Insert a fresh Snow Leopard Installation DVD into your drive and launch Disk Utility (/Applications/Utilities/). Since the release of Mountain Lion, you can obtain this DVD for $19.99 from the Apple Store or call their telesales agents [1-800-MY-APPLE (1-800-692-7753)] or Customer Service and Sales Support at 1-800-676-2775 or 1-800-275-2273 (ask for a Sales Representative). Do not use a grey, machine-specific Install DVD: see the attached photos.

2. Select the device for your optical drive (note that this is NOT the Mac OS X Installation disk, but rather the device that has mounted it; in this example the device is named “SuperDrive”). BE VERY CAREFUL HERE!

3. From Disk Utility’s File menu, choose “New > Disk Image from [device]” from the File menu, set the image format to “DVD/CD master” and Encryption to none and create the disk image on your Desktop. When the creation of the disk image file is complete, quit Disk Utility and eject the Snow Leopard Install DVD from your drive.

4. Launch Terminal (/Applications/Utilities/); cut and paste or carefully type the following 4 commands into Terminal and, after each, hit Return/Enter:

5. Change to the directory where the image is located.

% cd Desktop

6. Mount the disk image in a way that allows any changes to be kept in a ‘shadow’ file.

% hdiutil attach SuperDrive.cdr -shadow

7. Add the important file to the mounted image with this (including the quotation marks):

% touch '/Volumes/Mac OS X Install DVD/System/Library/CoreServices/ServerVersion.plist'

8. Unmount the disk image by ‘ejecting’ the ‘Mac OS X Install DVD’ in Finder. At this point you have two files - SuperDrive.cdr and SuperDrive.cdr.shadow on the Desktop.

9. Create a new DVD/CD master disk image ’SuperDrive-Server.cdr’ which is the same as SuperDrive.cdr, but with the change included.

% hdiutil convert -format UDTO -o SuperDrive-Server.cdr
SuperDrive.cdr -shadow

10. Quit Terminal, and continue using SuperDrive-Server.cdr in place of SuperDrive.cdr. If you wish you can now delete SuperDrive.cdr and SuperDrive.cdr.shadow - they are no longer needed.

(steps 4-10 are from here)

Step Two - Installing Snow Leopard into VMWare on Mountain Lion

1. Launch VMWare

2. From the File menu go to New… Then click on the “Choose Continue without disc” button

3. Drag and Drop the SuperDrive-Server.cdr image you created onto “Use operating system installation disc or image” radio button and click Continue.

4. Change Operating System to “Apple Mac OS X” and Version to “Mac OS X Server 10.6 64-bit” and click continue.

5. You can click on Customize Settings to make changes to the default. I took defaults and clicked on Finish.

6. OS X Snow Leopard will now start to be installed into the VMWare file called “Mac OS X Server 10.6 64-bit” located in “~Documents/Virtual Machines”. As with all Snow Leopard installations, the following selections need to be made, as they come up:

a. Use English for the main language (click on the right arrow).

b. Install Mac OS X (click on Continue).

c. Agree to the terms of the software licensing agreement by clicking Agree.

d. Select the disk where you want to install Snow Leopard by clicking on the Macintosh HD, but DO NOT YET CLICK ON INSTALL.

7. In the Utilities Menu to the right of the Mac OS X Installer Menu (at the top of the Snow Leopard installation screen), Launch Terminal;

8. Type the following command (including the quotation marks) in Terminal and then click RETURN/ENTER:

% mkdir -p '/Volumes/Macintosh HD/System/Library/CoreServices/'

9. Now type this command (including quotation marks) in Terminal and hit RETURN/ENTER:

% touch '/Volumes/Macintosh HD/System/Library/CoreServices/ServerVersion.plist'

10. Quit Terminal and go back to the installation of Snow Leopard; BUT FIRST: Click Customize and select Rosetta and optionally Quicktime 7 and all printer drivers. Click on OK.

11. Click on Install and the process will take upwards to 30 minutes or more as Snow Leopard is installed into your VM.

12. As it initiates Snow Leopard be sure to remember the Account Name and Password you establish.

13. Upon completion the installation process will reboot and you will now have a Snow Leopard VMWare Window. Cancel ANY attempt by Software Update at this time. I recommend the following optional settings in the Snow Leopard environment:

a. Go to Preferences in the Finder menu and make sure that Hard Disks and Connected Servers are selected.

b. Launch System Preferences and go to Date/Time and then Clock and select show seconds.

14. Shut down Snow Leopard. Make a snapshot of your install for backup.

Step Three - Updating the Snow Leopard VMWare environment

1. Installing VMWare Tools - Start your Snow Leopard VMWare session. Go to Virtual Machine menu option and click on Install VMWare Tools.

2. Preparation for updating Snow Leopard

a. a. Open TextEdit in Snow Leopard. NOTE: The first time you open TextEdit in the Snow Leopard VMWare environment, go to the TextEdit Menu, down to Preferences and under Format, instead of the default Rich Text, click Plain Text; then close the Preferences window.

b. b. Copy the following xml code into a blank open document in TextEdit (NOTE: if you open Safari in Snow Leopard and type in the following URL — http://forums.macrumors.com/showthread.php?p=14799174 — you can then easily copy and paste the following xml text into TextEdit):

<?xml version="1.0" encoding="UTF-8"?>  
<!DOCTYPE plist PUBLIC "-//Apple Computer//DTD PLIST 1.0//EN"  
"http://www.apple.com/DTDs/PropertyList-1.0.dtd">  
<plist version="1.0">  
<dict>  
<key>Label</key>  
<string>com.rectalogic.vmware</string>  
<key>ProgramArguments</key>  
<array>  
<string>/bin/bash</string>  
<string>-c</string>  
<string>/bin/rm -f /System/Library/CoreServices/ServerVersion.plist;
trap "/usr/bin/touch /System/Library/CoreServices/ServerVersion.plist;
exit" SIGINT SIGTERM SIGHUP; sleep 999999 &amp; wait $!</string>  
</array>  
<key>KeepAlive</key>  
<true/>  
<key>RunAtLoad</key>  
<true/>  
</dict>  
</plist>

NOTE: the 12th line in the middle of the text above to be entered into Text Edit, runs off the screen here on MacRumors. BE SURE to get the complete line; either copy and paste the whole text or be sure you are properly entering this COMPLETE line:

/bin/rm -f /System/Library/CoreServices/ServerVersion.plist; trap “/usr/bin/touch /System/Library/CoreServices/ServerVersion.plist; exit” SIGINT SIGTERM SIGHUP; sleep 999999 & wait $!

c. Go to Save As in the File Menu of TextEdit.

(1) Enter the file name: “com.rectalogic.vmware.plist” (without the quotation marks) in the Save As box (delete: “.txt”). You will see the correct default - Plain Text Encoding: Unicode (UTF-8)

[If you do NOT see “the correct default - Plain Text Encoding: Unicode (UTF-8)” when you try to Save As, it is probably because, when you pasted in the code, it pasted as Rich Text. You can fix this easily by going to Format > Make Plain Text, after which the UTF-8 option (and .txt option) will be available.];

(2) Uncheck the box “If no extension is provided, use ‘.txt’.” and save it to the Desktop. If it asks which extension, choose .plist.

(3) Quit TextEdit.

(4) Go to the Desktop and move the file to /Macintosh HD/Library/LaunchDaemons/ and when it asks - click Authenticate and then enter your Snow Leopards account password.

d. Open Terminal in Snow Leopard from /Macintosh HD/Applications/Utilities/. Enter the following text in Terminal and then hit RETURN/ENTER:

% sudo chown root:wheel /Library/LaunchDaemons/com.rectalogic.vmware.plist

e. It will ask you for your Snow Leopard account password; enter it (nothing will appear on screen as you type the password) and hit RETURN/ENTER.

f. Enter the following text in Terminal and then hit RETURN/ENTER:

% sudo launchctl load /Library/LaunchDaemons/com.rectalogic.vmware.plist

g. Quit Terminal

3. Updating Snow Leopard and other software updates
a. In the Apple Menu in Snow Leopard go to Software Updates…

b. After it Checks for New Software, it will tell you updates are available:

(1) Click the Show Details button.

(2) My advice is that at this time (May, 2012) all updates should be accepted except for the Apple Software Installation Update 1.0 (which has had some problems and probably only fixes installation problems of software that you would preferably be installing into Lion anyway).

(3) You will most likely have to restart after the Software Installation and there may be additional software updates that will appear after a restart.

(4) Repeat this process until all recommended updates are complete.

c. In System Preferences, go to Sharing and turn on File Sharing. NOTE: you may have to turn on printer sharing in your Lion environment to access your printers in Snow Leopard.

4. Congratulations: you now have a completely functional Snow Leopard environment in VMWare

Step Four - Backup your Snow Leopard VMWare Files
1. Shut Down Snow Leopard.

2. Make a backup of the VMWare file “~Documents/Virtual Machines/Mac OS X Server 10.6 64-bit”.

END OF INSTALLATION INSTRUCTIONS

Salmon Papillote

2013-06-24T00:16:00-04:00

This is the easiest dish to make and it is full proof and quick to make.

Ingredients

Salt - 1 tsp sprinkle all over the fish
Pepper - sprinkle all over the fish
Lemon - quarter wedge
Garlic powder - sprinkle all over the fish
Extra virgin olive oil

Apply salt, pepper and garlic powder to the fish. Then drizzle with the olive oil and squeeze a quarter of lemon. Now seal the salmon in aluminum foil.

Preheat the oven at 375 degrees. Put the packaged salmon on a baking sheet and let it bake for 15 minutes.

Serve it hot with some rice pilaf, mashed potatoes and grilled asparagus! Do not throw away the sauce you see in the above picture.. Drizzle that on top of fish before serving! It’s super tasty!!

The picture below shows how the packaged foil looks like.

Chicken Saag

2013-06-19T22:50:00-04:00

This dish came out soo tasty! It goes really well with both cooked white rice and chapatti’s..

Here are the ingredients for this this -

Chicken thighs cut into small chunks - 2 pounds
Spinach - 2 bunches
Ginger - one inch
Garlic - 4 cloves
Green Chillies - 2
Bay Leaves - 2-3 leaves
Pepper Corns - 10
Onions - 1 chopped
Tomatoes - 3-4 chopped
Chilli Powder - 1/2 tsp
Coriander Powder - 2 tsp
Garam Masala - 1 and 1/2 tsp
Turmeric - 1/2 tsp
Yogurt - 1/2 cup
Salt to taste

Method

Cook the spinach in pan. Toss the spinach till it reduces for maybe 5 minutes. Put the cooked spinach with ginger, garlic, green chillies and water if required in a blender and process to a thick puree.

Heat the oil in a large heavy pan. Add bay leaves and black peppercorns and sauté for a minute. Stir in the chopped onions and cook till golden brown. Add the chopped tomatoes and cook for 6-8 minutes.

Now add the Chilli powder, Coriander Powder, Garam Masala, Turmeric and salt and mix it into the tomato onion mixture. Cook for 2-3 minutes till the oil starts to separate.

Stir in the spinach puree you made earlier. Add about 3/4 cup of water and simmer for 5 minutes. Add your beaten yogurt and mix it well into the spinach mixture and simmer for another 5 minutes. Now add your chicken and stir the spinach to coat the chicken. Cover and cook for 25-30 minutes on medium to low heat till the chicken is tender.

Garnish with a dollop of yogurt and either server with warm Naan or hot white rice

This dish can also be made with Paneer (indian cottage cheese) instead of chicken. Just cut paneer into cubes and add at the point where you add the chicken. Paneer does not need to be cooked for 25-30 minutes. It will be ready in just 5-10 minutes

Nihari

2013-06-18T09:51:00-04:00

This is one of those dishes which brings really fond memories of my childhood going to these shops in Old Delhi(India) serving Nihari early in the morning. I think this is my favorite dish of all time!! This dish is usually eaten early in the morning. Read this wikipedia entry for more details on this dish.

Here is the recipe -

Ingredients for Nihari Masala

3 tbsp saunf/fennel
2-3 sticks dalchini/cinamom
6 choti elaichi/green cardamom
2 badi elaichi/black cardamom
12-15 kali mirch/black pepper seeds
1 tbsp sonth/dried ginger powder
6 pieces - javatiri/mace
nutmeg 1/2 inch piece
1.5 tbsp spoon lal mirch/red chili powder
1 tbsp dhaniya/coriander powder
1/4 tsp haldi/turmeric

Ingredients for the recipe

1 KG Lamb meat cut into big chunks. You can replace this with any red meat.
1 onion - chopped
2 tbsp adrak lahsun/ginger garlic paste
1 tbsp salt
1 tbsp lal mirch.red chili powder
4 tbsp of Nihari Masala
1/2 cup yogurt

Take all the ingredients for Nihari Masala and use a dry grinder or coffee grinder to make a powder.
Now take a heavy bottom vessel as you will be cooking the meat for atleast 3-4 hours. I used an enameled dutch oven to cook this.

Add 1/2 cup of ghee to your vessel and heat on a medium flame.
Add 1 chopped onion and cook till light golden
Add 2 tbsp adrak lahsun/Ginger Garlic paste and cook for a minute
Add your meat - 1 KG/2.2 Pounds and let it brown on all sides (maybe 2-3 minutes on each side)
Add 1 tbsp salt
Add 1 tbsp lal mirch/red chili powder
Add 4 tbsp nihari masala pwder from above and mix it well.
Now Add 1/2 cup dahi/yogurt and mix it well into all the meat.

At this point this is what you should be seeing -

Lower the heat and add about 10 cups of water. Cover up the vessel and forget about this dish for the next 3-4 hours. Well almost. Just check every 30 minutes to make sure the masala does not stick to the bottom of the vessel. If the liquid evaporates and your meat hasn’t become tender, just add more water.

In my case the meat got tender in about 4 hours. I had to replenish the water at about the 3 hour mark but your mileage might vary.

Once the meat is tender, make a very thin paste by mixing 1/2 cup of all purpose flour and 4 tbsp of wheat flour with water. Add this mixture to your dish. This is used as a thickening agent. Now keep a close eye on your dish and make sure the bottom is not sticking. Don’t overly mix the dish as you don’t want to break down the meat. Taste and add more salt if required. Once you see a boil after adding the thin paste the dish should be ready.

Add a garnish of julliened ginger, chopped cilantro and crispy fried onions. Serve it with Naan or chappati.

How to reset QmailAdmin password

2013-05-19T07:07:00-04:00

Actually you can use this command to reset any users qmail account password.

Simply run the following command to change the password for user@domain.com either by using sudo or as root.

# /home/vpopmail/bin/vpasswd user@domain.com

The above command will ask you for the new password twice. That was easy!

Roast Pork Banh Mi at 163 Vietnamese Sandwiches and Bubble Tea

2013-05-16T15:09:00-04:00

My favorite Banh Mi (Vietnamese Sandwich) place in Chinatown Boston introduced a bunch of new sandwiches on their menu. The place is called 163 Vietnamese Sandwiches and Bubble Tea. It’s in 66 Harrison Avenue. The name is a mouthful for sure :)

The roast pork looked like thickly cut ham with the usual fixings of vietnamese mayo, drizzle of some kind of fish sauce, thinly sliced cucumbers carrots and daikon, chopped scallions and cilantro on a crispy baguette. I usually ask for my sandwich to be spicy so they add some chopped up Thai Chillies. For $3.75 it’s a steal.

Here are some pics

The sandwich is good, but I think their pork chop sandwich, BBQ pork Sandwich and cold cuts sandwich are much much better than this.

Cold Noodle Salad at Fugu Truck Boston

2013-05-14T14:27:00-04:00

Is it wrong if I said “I Fugu You”?? (Read Fugu=Love). I think the food in Fugu is made with such care and love that every dish comes out tasty and you just want to come back for more..
And that’s what I did today. I went back for more.. Today I tried their Cold Noodle salad with sesame sauce with Pork.

I think it is their best dish yet! It looks as pleasing as it is tasty.. the combination of cilantro, scallions and thinly sliced cucumbers and carrots give a very colorful and fresh look! I should add that the noodles were also cooked well.

Here are some pictures -

Hmmmm.. Where is my pork?? looks like somebody forgot to put pork in my order.. but the Guys on Fugu Truck are really nice about it and hook me up with some extra :) Thanks Guys!! Look at that pile!

This is after mixing it all up with the extra hot chilli oil I asked for (It’s soo tasty… You must!!)

and don’t worry, I did not forget the customary pork buns this time too..

Thanks @fugutruck for another wonderful lunch!

Soy Ramen at Ginza

2013-05-14T14:26:00-04:00

Finally got to go back to Ginza last week to try their Soy Ramen. So tasty and refreshing.. I liked it! I still feel the Pork Broth Ramen is the best tasting among all the options at Ginza.. But second place is a tough choice. It is a tie between the miso and soy flavored Ramen!

Here are some pics

Roasted lamb sandwich from Flour Bakery

2013-05-01T13:33:00-04:00

Today I decided to walk to Flour Bakery + Cafe to try one of their sandwich’s and indulge in their famous sticky buns!

It was another beautiful day in Boston.. Bright, sunny and pleasant.

When I got their I decided on their Roast Lamb sandwich with tomato chutney and goat cheese.. Sounds yum! For desert I went with their sticky buns. Actually the bun was going to be a snack for later in the day ;)

Decided to do take out and eat my lunch over looking the water.. Found a beautiful wooden bench.. it was more like nicely varnished wooden pillars joined together to look like a bench. Flour does a nice job with their packaging for take out.

Here is a view of their famous Sticky Buns… droool!!!

This is another view of the buns taken indoors

Finally the sandwich itself..

and here is another angle..

Sandwich does look tasty. Loved the bread. I would never have thought of ever trying the combination of roast lamb with tomato chutney and goat cheese. They sure go well! The goat cheese adds a nice tangy/tarty flavor to the sandwich. I would have liked a wee bit more salt but maybe it’s just me.

The bakery has a good vibe and decor is hip. They have a lot of different sandwiches and baked goods to offer so will have to come back soon.

Bi Bim Bap at Fugu Truck Boston

2013-04-30T15:09:00-04:00

I had a nice twitter exchange with @fugutruck last week where I was recommended to try their Bi Bim Bap. I promised I would come to their Downtown Boston location this week and true to my promise, I was there!

I took @Erickinboston and @steveobc with me as I wanted them to join me this week. It was a beautiful day out and we decided to eat at the Post Office square. Great decision!!

I knew what I wanted today… The Bi Bim Bap which is a Korean style rice bowl mixed with meat, vegetables and a sweet pepper sauce, and off course their sooo tasty pork buns!!

Here is what their Bi Bim Bap looks like

The week long wait for the Bi Bim Bap was totally worth it!! I loved the flavors of the rice bowl. The best way to eat this is to mix everything together. I didn’t add the pepper sauce at first as I wanted to taste this dish on its own. Then mixed the sauce to a little rice to see what the flavors were like. The sauce definitely gives this dish more meaning and taste!! All I can say is I am totally at peace after having another wonderful meal at this fine establishment!!

@Fugutruck I bow to you for bringing such high standards to the Boston Food Truck scene!!

See you guys soon..

Pork Belly at Fugu Truck Boston

2013-04-23T13:15:00-04:00

I got a chance to go out and try something new today at Fugu Truck. But of course I still had to get their pork belly sandwich.. Just love me some Fugu truck Pork Belly Sandwich!! This time I decided to go with the grilled Pork Belly Sandwich instead of the steamed. The difference lies in the texture of the bun. The grilled bun adds a certain crispness to the dish. Very fresh and tasty!

The new dish I tried today was the Kimchee Bulgogi Panini..

The kimchi gives it a nice flavor. Both sandwiches were very tasty and completely different.

These guys are making it really difficult for me to narrow down one favorite dish. I still have 3 more dishes to try before I have eaten their entire menu.. Keep posted to see the other dishes :)

Xcode cache causing problems

2013-04-10T13:35:00-04:00

The keyboard sequence Command-Option-Shift-K is used to clean out the build folder. Even better, quit Xcode and clean out ~/Library/Developer/Xcode/DerivedData manually. Remove all its contents because there’s a bug where Xcode will run an old version of your project that’s in there somewhere. (Xcode 4.2 will show you the Derived Data folder: choose Window > Organizer and switch to the Projects tab. Click the right-arrow to the right of the Derived Data folder name.)

How to preserve hyperlinks when you turn a web page to pdf

2013-04-09T13:00:00-04:00

This is sooo easy to do on a mac. You can create a PDF from anywhere you can print on a mac. The problem comes to preserving hyperlinks when printing a webpage to PDF. The trick is to use Safari. If you print a webpage on Safari to PDF, all hyperlinks are preserved :)

Miso Ramen at Ginza

2013-04-09T12:48:00-04:00

After trying out the [clear Pork Broth Ramen at Ginza last week, I got a chance to try their Miso Ramen today. Me so like it ;)

Here are a few pics I got

Here is another one from the top..

Honestly.. it was good! The broth has a nice miso flavor and very satisfying! The only difference between the Miso based Ramen and Pork based Ramen other than the wonderful broth was a slice of butter on my Miso Ramen. Also today the pork was tastier and more tender than the one I had last time with my Pork Broth Ramen!

Last time I was there, the place was super busy and waited 30 minutes for my order to come out. Today I was at Ginza by 11:40AM and I believe was one of their first customers of the day. The service was quick in comparison.

My friend from work who got the pork broth ramen was kind enough to let me taste his pork broth so I could taste and compare the two broths.. it’s a little tough to decide, but we both agreed that the pork broth is the the winner!!

I now have the Soy based broth left to taste.. Will definitely come back for it!!

How to Join AVCHD M2TS files and convert to MKV

2013-04-05T15:30:00-04:00

There is still no cheap(free) and easy way to work with the M2TS files created by my Canon H100 video camera. When I am shooting with that camera, for any event, I end up with a bunch of .m2ts video files. VLC now plays this format, but still I have to either create a playlist of all my files or play them one at a time.

This is where the freeware tsMuxer comes to the rescue. SmartLabs doesn’t offer this app anymore, but you can download this from videohelp.com. You can download the app for windows, linux and Mac. If you are running OSX, use the VoxMac Unofficial Build. Even this build does not work correctly because of some korean fonts (I would love to know why). Here is a snippet from macrumors forums on how to fix that problem.

tsMuxeR is not working due to fonts interpretations problem.
We got a ‘bus error’ when trying to execute the command in command-line mode.
The problem is due to 2 particulars korean fonts supplied in Snow Leopard.
To solve the problem, if you don’t use those korean fonts, delete
the two following fonts : AppleMyungjo.ttf and Gungseouche.ttf located in /Library/Fonts
When done, the tsMuxeR command Line should Work

Instead of deleting the fonts, I just renamed them using the command line. tsMuxerGUI now launches with out any problems.

Joining .M2TS file
I use tsMuxer to combine all the .M2TS files into one .M2TS file. It’s very easy and very quick to join these videos. These steps should work on Windows and Linux too. Here are the step -

1. Launch the app

Drag and Drop the first file in the sequence in the Input files window. The Add button doesnt work for M2TS files.

Now Select the rest of the files you want to join together, drag and drop on the Join button at the top.

Make sure the Output is set to M2TS muxing and click on the Browse button to choose an appropriate place to save the output. Now click Start Muxing to join all the .M2TS files.

You will get a status window showing progress. Click Ok when it finishes to get back to tsMuxerGui

You should now have your joined .M2TS file.

But I still can’t play This .M2TS on XBMC. The solution is to convert this to an MKV file :)
The trick is to demux this joined .M2TS file using tsMuxer and then use mkvmerge to create an MKV file. Problem Solved! You can download mkvmerge from the MKVToolNix site. They have binaries for windows, multiple linux platforms and Mac.

Demuxing .M2TS file

Drop your joined .M2TS file in the input files section of the tsMuxerGUI

Select Demux in the output section at the bottom and click on Start Demuxing. You will see the progress window again.
The Demux process creates two files, a .264 (video) file and a .ac3 (audio) file.
Click on the add button on the input files section of mkvmerge and add your .264 and .ac3 files.

Click on Start muxing at the bottom to create your MKV file. A progress window comes up during the muxing process.

You now have an MKV file which works on XBMC. As tsMuxer and mkvmerge are tools also available on linux and windows, this process should work on all platforms.

Unix Aliases with variables in bash

2013-04-05T13:23:00-04:00

A lot of times you are typing the same command again and again on the unix command line. Unix Aliases help to save time from typing out these long commands. Back when I was using C Shell, I could use parameters in aliases. For example I have this in my .cshrc file to find certain running processes-

alias chk "ps auxwww | grep \!:1"

running this would show me all commands which have java in them

$ chk java

Unfortunately this alias does not work in Bash. The work around is to create a function for that, rather than an alias, and then exported it, like this:

function chk { ps auxwww | grep "$1"; }  
export -f chk

Note -f switch to export: it tells it that you are exporting a function. Put this in your .bashrc and you are good to go

Ramen at Ginza

2013-04-04T13:48:00-04:00

My brother gave me some really good news 2 weeks ago.. Ginza in Boston Chinatown now offers Ramen as part of their menu!! Soo exciting!! The best part is that Ginza is a stones throw away from my workplace. If the ramen is any good, it’s going to be my goto place whenever I get a craving!! Bro told me that they offer it in three different broths.. Clear Pork broth, Miso broth and finally Soy based broth. I have got to try them all!!

Today I finally got a chance to go try it out.. I got the Clear Pork broth Ramen and here is how it looks..

The broth tastes good and is very fulfilling. The Ramen noodles itself had a good texture. I did wish I had more noodles… but it could also just be the big bowl :). The pork slices could have been more tender but over all I really enjoyed my lunch!

I had a quick chat with the staff and they promised me that the Ramen was not a short time promotion and it was going to be part of their menu going forward!! Yay!!!

I am now looking forward to going back to try the Soy and Miso based broths.

Get Table creation date on SQL Server

2013-04-02T10:49:00-04:00

I was trying to find some changes I had made to some tables and was having a hard time finding the table I modified.. This helped me find the ones I was last working on ;)

For SQL Server 2005 and up, you can use

SELECT  
[name]  
,create_date  
,modify_date  
FROM  
sys.tables

How to build Growl 2.0.1 from source on Mountain Lion 10.8.2

2013-03-10T23:16:00-04:00

Growl is the ultimate notification system for OS X. A lot of apps support and use growl to notify the user about any status updates. You can either buy the app from the app store for $3.99 or build and install from source. This page explains how to build Growl from source.

(Click here to skip my ordeal and see the instructions which worked for me)

Unfortunately the instructions were not enough. I went through a lot of issues trying to build this. This post is my ordeal in trying to build and install Growl on my laptop.

Here is a list of requirements as mentioned in the official build notes -

OS X Mountain Lion - 10.8.2
Xcode - 4.6
Mercurial Client - I used the one from here
Git Client

We will now pull down the source code using mercurial to an appropriate location on our mac. I have a directory called projects where I do all my development. You may do this, where ever you want under your home.

Open a terminal window and execute the following

$ export LC_ALL="en_US.UTF-8"  
$ hg clone https://code.google.com/p/growl/  
$ cd growl  
$ hg tags | sort

The “hg tags | sort” command shows you all the versions of growl you can build. Now update the source to what ever version you want. We are building the latest version as of today, so run the following command.

$ hg update "Growl.app 2.0.1"

Now we are ready to do a build. The official notes misses out on telling us how to do this build.
Reading instructions from other sources I found that running the following command inside the growl directory should do the trick..

$ xcodebuild -project Growl.xcodeproj -target Growl.app -configuration Release

Here I got my first failure

$ xcodebuild -project Growl.xcodeproj -target Growl.app -configuration Release  
=== BUILD AGGREGATE TARGET VCS Data OF PROJECT Growl WITH CONFIGURATION Release ===  
Check dependencies

=== BUILD NATIVE TARGET GrowlLauncher OF PROJECT Growl WITH CONFIGURATION Release ===  
Check dependencies  
Code Sign error: The identity '3rd Party Mac Developer Application: The
Growl Project, LLC' doesn't match any valid, non-expired
certificate/private key pair in your keychains

** BUILD FAILED **

The following build commands failed:  
Check dependencies  
(1 failure)

This is one of the failures the official document warned us about. A Quick fix is to generate a self signed certificate with the prefix ‘3rd Party Mac Developer Application: The Growl Project, LLC’.

Create Self Signed Certificate

Open the Keychain Access.app, go to Keychain Access menu then go to ‘Certificate Assistant’ and finally click on ‘Create a Certificate…’. Fill out all details as shown in the screenshot below

Here are the details -
Name: 3rd Party Mac Developer Application: The Growl Project, LLC
Identity Type: Self Signed Root
Certificate Type: Code Signing

Now click on create and click on continue button when it shows you the “You are about to create a self-signed certificate” and finally click on Done when you see the below screen.

Close the Keychain Access app and re-run the build command on the terminal window.
This time you should see a popup window like the one shown below saying “codesign wants to sign using key ‘3rd Party Mac Developer Application: The Growl Project, LLC’ in your keychain”. Looks promising… Click Always Allow.

but this time the build failed with two new errors

/Users/shireenrao/projects/growl/Plugins/Actions/Speech/GrowlSpeechDisplay.m:15:9:
fatal error: 'ShortcutRecorder/ShortcutRecorder.h' file not found  
#import <ShortcutRecorder/ShortcutRecorder.h>  
^  
1 error generated.

/Users/shireenrao/projects/growl/Plugins/Actions/Speech/GrowlSpeechPrefs.m:12:9:
fatal error: 'ShortcutRecorder/ShortcutRecorder.h' file not found  
#import <ShortcutRecorder/ShortcutRecorder.h>  
^  
1 error generated.

2013-03-10 15:19:13.982 xcodebuild[45006:7b03] DVTAssertions: Warning in
/SourceCache/IDEXcode3ProjectSupport/IDEXcode3ProjectSupport-2108/Xcode3Sources/XcodeIDE/Frameworks/DevToolsBase/pbxcore/SpecificationTypes/XCGccMakefileDependencies.m:87  
Details: Failed to load dependencies output contents from
``/Users/shireenrao/projects/growl/build/Growl.build/Release/Speech.build/Objects-normal/x86_64/GrowlSpeechPrefs.d''.
Error: Error Domain=NSCocoaErrorDomain Code=260 "The file
“GrowlSpeechPrefs.d” couldn’t be opened because there is no such file."
UserInfo=0x402b75740
{NSFilePath=/Users/shireenrao/projects/growl/build/Growl.build/Release/Speech.build/Objects-normal/x86_64/GrowlSpeechPrefs.d,
NSUnderlyingError=0x402e22420 "The operation couldn’t be completed. No
such file or directory"}. User info: {  
NSFilePath =
"/Users/shireenrao/projects/growl/build/Growl.build/Release/Speech.build/Objects-normal/x86_64/GrowlSpeechPrefs.d";  
NSUnderlyingError = "Error Domain=NSPOSIXErrorDomain Code=2 \"The
operation couldn\U2019t be completed. No such file or directory\"";  
}.  
Function: void
XCGccMakefileDependenciesParsePathsFromRuleFile(NSString *, void
(^)(NSString *))  
Thread: {name = (null), num = 11}  
Please file a bug at http://bugreport.apple.com with this warning
message and any useful information you can provide.

** BUILD FAILED **

The following build commands failed:  
CompileC
build/Growl.build/Release/Speech.build/Objects-normal/x86_64/GrowlSpeechDisplay.o
Plugins/Actions/Speech/GrowlSpeechDisplay.m normal x86_64 objective-c
com.apple.compilers.llvm.clang.1_0.compiler  
CompileC
build/Growl.build/Release/Speech.build/Objects-normal/x86_64/GrowlSpeechPrefs.o
Plugins/Actions/Speech/GrowlSpeechPrefs.m normal x86_64 objective-c
com.apple.compilers.llvm.clang.1_0.compiler  
(2 failures)

So at this point I started looking around the growl google group. I found a couple of things. Firstly Growl 2.0.1 was built on Xcode 4.5.x. I am trying to do this on 4.6. Secondly, the rake build tool is now being used to build growl. Armed with that info this is what I tried next

$ cd Release  
$ VERSION=2.0.1 rake setup  
xcrun: error: unable to find utility "packagemaker", not a developer tool or in PATH

Another error.. This time about a tool called packagemaker. After some google foo, found out that packagemaker is no longer part of xcode. It is now available in a separate package called Auxiliary tools for Xcode Developer Preview. Get it from here.

Just copy PackageMaker to your Applications directory so it is available on the command line. Once installed, try the build again.

$ VERSION=2.0.1 rake setup  
(in /Users/shireenrao/projects/growl/Release)  
$ VERSION=2.0.1 rake build:growl

Two new errors come up… -

/Users/shireenrao/projects/growl/Plugins/Displays/GrowlDisplayWindowController.m:400:42:
error: cast from function call of type 'void *' to non-matching type
'NSInteger' (aka 'long') [-Werror,-Wbad-function-cast]  
NSInteger startPercentage = (NSInteger) NSMapGet(startTimes,
transition);  
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

/Users/shireenrao/projects/growl/Plugins/Displays/GrowlDisplayWindowController.m:401:42:
error: cast from function call of type 'void *' to non-matching type
'NSInteger' (aka 'long') [-Werror,-Wbad-function-cast]  
NSInteger endPercentage = (NSInteger) NSMapGet(endTimes, transition);  
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~  
2 errors generated.

First Code Change

This is a casting error and I think mainly because I am trying to build this using xcode 4.6. I opened the file “growl/Plugins/Displays/GrowlDisplayWindowController.m” and removed the cast “(NSInteger)” from the two lines right before NSMapGet. Tried running the command “VERSION=2.0.1 rake build:growl” again and got a new error

/Users/shireenrao/projects/growl/Plugins/Actions/SMS/GrowlSMSDisplay.m:141:97:
error: format specifies type 'unsigned int' but the argument has type
'CFIndex' (aka 'long') [-Werror,-Wformat]  
CFStringRef contentLength =
CFStringCreateWithFormat(kCFAllocatorDefault, NULL, CFSTR("%u"),
CFDataGetLength(postData));  
~~ ^~~~~~~~~~~~~~~~~~~~~~~~~  
%lu  
1 error generated.

Second Code Change

This was an easy fix as the fix is in the error shown above. I just change %u to %lu in the file growl/Plugins/Actions/SMS/GrowlSMSDisplay.m on the error line.

After retrying the build command once again, I finally saw this

** BUILD SUCCEEDED **

Yippeee!! The Growl.app is now available under growl/Release/distribution/build/growl/Release/Growl.app

Just copy that to your applications folder and you have now built Growl from Source :)

Ok, in all honesty, I am not sure if I am doing the right thing by taking out the cast from the file “growl/Plugins/Displays/GrowlDisplayWindowController.m”. If I see any problems with the app, I will revisit this problem.

I quickly wanted to mention another problem I encountered. A while back, I had done a build using xcode. I don’t remember what version of xcode I had used and I was definitely using an older version of growl. Actually I don’t know which version I was using because I had not updated the mercurial tag to say which version I should be using. As I said, I used xcode to do the build. The build succeeded. I copied the build to my applications and it was working for a while. I then started seeing the message - “This Beta Has Expired / Please download a new version to keep using Growl”. I didn’t know what happened for a while. Turns out, I did not copy the app from the release configuration. I had copied Growl.app from the debug configuration. Only the release configuration doesn’t have the beta flag in the code.

Now I know that to create a release build in xcode, you have to edit your current scheme. You do that by hitting (⌘<) inside xcode and highlight “Run [name of application]. On the right, select “Build Configuration” and choose “Release”. Click Ok to close the window. Here is a screenshot of what you will change -

Build by hitting ⌘B in Xcode. If you have your self signed certificate and have edited the two files as mentioned above, the build should succeed. But, there is one slight problem… The version number in the app is incorrect. It shows up as 2.0b6 instead of 2.0.1.

That’s why the rake build method is recommended.

Here is a quick synopsis of the entire build process without my hiccups

Requirements

OS X Mountain Lion - 10.8.2
Xcode - 4.6
Mercurial Client - I used the one from here
Git Client
Packagemaker - look for Auxiliary tools here
Create self signed certificate as shown here

Steps

$ hg clone https://code.google.com/p/growl/  
$ cd growl  
$ hg update "Growl.app 2.0.1"

Edit files as shown here and here

$ cd Release  
$ export LC_ALL="en_US.UTF-8"  
$ VERSION=2.0.1 rake setup  
$ VERSION=2.0.1 rake build:growl

Growl.app is available at growl/Release/distribution/build/growl/Release/Growl.app

XBMC has No Sound after upgrading to Frodo on Mountain Lion

2013-03-06T22:03:00-05:00

I wanted to test out Frodo on my laptop before I upgraded xbmc on my mac mini which is my main home theatre PC. I am still running Version 11.0 (Eden) everywhere. Good thing I tested this on my laptop first because the upgrade was a fail. I could not get any sound to play after the upgrade. After poking around the settings, I found out that the Audio Output Device was set to Airplay as shown in xbmc’s Audo Output section of the System Settings screen shot below -

I just changed it to “Built-in Output” and that fixed the issue. I was also hearing only from one side of my headphones. Seeing the release notes for this new release I saw that it was a known bug and can be fixed by changing the Speaker Configuration from 2.1 to 2.0. Here is a screenshot after my changes.

Bread Pakora

2013-03-02T14:25:00-05:00

This is how I remember eating Bread Pakora’s in Amar Colony, New Delhi. So the other day when I suggested to Priya that we should try making Bread Pakoras with Paneer, Priya was inspired!

This is Priya’s recipe -

Ingredients
Paneer
Dhaniya chutney
Chat masala
White Bread
Besan (Chick Pea Flour) - 1 Cup
Red Chili powder according to taste
Salt to taste
Baking Soda - 1 pinch
Rice Powder - 1 table spoon

Preparation

The only paneer you can buy here is the frozen kind, so to give it that fresh look just take it out of the package at put it in hot water. Once it is soft, cut it in thick chunks so it will fit in between your sandwich.

Prepare a smooth batter by mixing Chick pea flour, rice powder, salt, red chili powder, baking soda and water. The rice powder is used here to make crisp pakoras.

Spread Dhaniya Chutney on the bread, put your thick slices of paneer, sprinkle some chat masala and close it with another slice of bread.

Heat some cooking oil in a vessel.

The last step is to dip your sandwich in the batter so it is covered with the batter on all sides. A good way of doing this is to hold the sandwich and start dipping into the batter, all the edges first and then finish off by dipping on the two sides. Now put this in the hot oil and deep fry till golden brown like you see it in the picture. Take it out and put it on a wire rack to drain any oil out. Or just use some kitchen napkins to soak up the excess oil out.

Cut the pakora into half’s and sprinkle some chat masala before serving.

Syntax Checking your python script without running it

2013-03-01T11:20:00-05:00

This is a neat little trick to test your python code on the command line without running the script.

$ python -m py_compile yourpythonscript.py

All that command does is compile your python script and create a .pyc file. If there are any errors, you will see the errors be spit out.

Read Video files metadata in python using hachoir-metadata

2013-02-28T09:17:00-05:00

The only module out there for reading video files metadata I could find is hachoir-metadata. Actually you can extract metadata from multimedia files like images, audio files and video files. it also works on different types of archives. I tested it against all kinds of files and it does a great job in most of the popular ones. Their bit bucket site lists all the formats hachoir-metadata supports.

Unfortunately installing it on my mac using python pip does not fulfill dependencies. I was getting the following error trying to run their sample program..

$ hachoir-metadata  
Traceback (most recent call last):  
File
&amp;amp;amp;quot;/Users/shireenrao/.virtualenvs/default/bin/hachoir-metadata&amp;amp;amp;quot;,
line 5, in &amp;amp;amp;lt;module&amp;amp;amp;gt;  
from hachoir_core.error import error, HachoirError  
ImportError: No module named hachoir_core.error

I was surprised when doing a google search saw that nobody else had seen this issue. After reading their website, found that the module had dependencies. I Fixed this by installing the following dependencies…

$ pip install hachoir-parser  
$ pip install hachoir-core

My macs list of “Open with” applications start listing the apps twice

2013-02-27T08:17:00-05:00

When I right click on a file on my mac, I started seeing multiple apps being listed in the “Open with…” menu.

Typically this list of apps are the apps which can open the file. Seeing the same application listed multiple times in the “Open With…” list doesn’t harm anything but is definitely annoying!

So I googled the problem, and saw others were having the same issue. Basically the launchServices database gets corrupted which leads to this problem.

To fix on Mountain Lion:

1. Open Terminal.app in the /Applications/Utilities/ folder.
2. run

$ /System/Library/Frameworks/CoreServices.framework/Versions/A/Frameworks/LaunchServices.framework/Versions/A/Support/lsregister -kill -r -domain local -domain system -domain user

3. Restart

If you don’t like using the command line, the free tool Onyx also has an option under Maintenance -> Rebuild -> launchservices.

Another option is to uninstall the application which is showing up multiple time and install it back.

All the above options work.

Egg Fried Rice

2013-02-24T22:52:00-05:00

I love Egg fried rice. The best place for me to eat this is at The Best little Restaurant in Boston Chinatown. It’s plain, simple and very flavorful. It’s also very quick to make at home.

Ingredients

Rice - 1 cup, preferably made in advance or left overs are best
Eggs - 3 lightly beaten
Scallions - 1/2 cup thinly sliced
Sesame oil - 1 teaspoon
Cooking Oil - 4 tablespoons. Peanut Oil preferred.
Salt - to taste
MSG - 2 pinches

Preparation
All stir fries should be done over high heat. So heat a wok over high heat. Add 2 table spoons of cooking oil, swirling to coat the wok evenly, and heat until hot and just smoking. Season the lightly beaten egg with a pinch of salt, slightly beat the eggs again to mix in the salt. Add eggs and swirl the wok in all directions to form a thin even layer on cooking surface, and cook. As soon as the egg sets, flip the omlete. Another 30 seconds and you can now take it out of the wok. I like my omlete to be a nice bright yellow, so I try not to over cook this. Now add the remaining 2 table spoons of cooking oil and heat it. Add the one teaspoon of sesame oil and then add the rice, salt, MSG and stir-fry. Add the cooked eggs and break it up while mixing it into the rice, letting rice rest several seconds between stirs, until rice is hot, 2 to 3 minutes. Add scallion greens and stir-fry until combined well.

You can also add blanched mixed vegetables if you want to make it a vegetable fried rice. You don’t have to use Peanut oil if you don’t have it. I normally just use regular Olive Oil meant for sautéing instead of peanut oil. Enjoy!

Fix Python virtualenv errors after upgrading Lion to Mountain Lion

2013-02-21T07:00:00-05:00

I upgraded the OS of my mac mini to Mountain Lion this weekend. Unfortunately it messed up all my python settings. When I ssh’d into the mac mini I saw an ugly python error. The gist of which was the following -

“ImportError: No module named virtualenvwrapper.hook_loader”

Basically I just had to Install the Xcode Command Line tools, and that would have fixed everything. Just in case, I also installed pip, virtualenv and virtualenvwrapper again.

It’s all good now :)

“No Camera Found” when I plug in my iPhone

2013-02-19T09:58:00-05:00

I was trying to copy some videos I had taken today to my laptop. I was syncing my iPhone after a really long time.. actually since apple gave us iCloud I dont have to sync for apps and pictures anymore. as I needed to get the videos.. I had too.. Maybe I should use dropbox for this?? Ohh I digress! so back to what I was doing.. Oh yes, trying to sync my phone.. I plugged in the USB and saw an error message pop up saying “No Camera Fond”. I tried a couple of things.. rebooted my phone and then my laptop (Don’t laugh.. I am a windows user at work).. Nothing worked. Back to Google I found out what was wrong. My image capture settings had changed after I installed Canon’s imagebrowser software.

On the Mac, after the iPhone is connected to the computer, open the Image Capture application. Select the iPhone and an option is presented in the bottom of the left panel. Change the “Connecting this iPhone opens” to an option of your choice. I chose “no application”.

GNU Screen on Mac OSX

2013-02-18T23:54:00-05:00

Did you know Gnu Screen came default with your macs? I did not know that!!
I wanted to run one of my python scripts remotely on my mac mini to detect duplicates in my image library. As this would take some time, I didn’t want to leave the session running on my laptop. I remembered using screen when I was actively using linux a few years ago and thought it would be a great solution here. First thing I did was do a quick google search to see if there was a ready made package for mac OSX. To my surprise I found out that it’s already installed on my mac :) Yipppee!!

Here is a cheat sheet - http://aperiodic.net/screen/quick_reference

Python development setup on Mountain Lion - part 4

2013-02-18T09:31:00-05:00

Setup Git on mountain lion to use github to save your projects

Continuing on my series, today we will be setting up git on our newly upgraded Mountain Lion.
Go ahead and download the latest version of git from here. Go ahead and install it. Next go and open an account on GitHub if you don’t already have one.

We will be using SSH keys to establish a secure connection between our computer and GitHub. Just follow these excellent instructions as described at https://help.github.com/articles/generating-ssh-keys

Just confirm that once you have it setup you can run the following command and see the output as shown below -

$ ssh -T git@github.com  
The authenticity of host 'github.com (207.97.227.239)' can't be
established.  
RSA key fingerprint is
16:27:ac:a5:76:28:2d:36:63:1b:56:4d:eb:df:a6:48.  
Are you sure you want to continue connecting (yes/no)?

Just type yes and hit enter to add the RSA key for github.com to your list of known_hosts. You should then see the following -

Warning: Permanently added 'github.com,207.97.227.239' (RSA) to the
list of known hosts.  
Hi username! You've successfully authenticated, but GitHub does not
provide shell access.

That’s it! You can now checkin and checkout code to GitHub.

If I want to clone the mp3cleaner project from Github, I will cd to the location where I will want the project to be copied. I then run the git clone command.

$ git clone git@github.com:shireenrao/mp3cleaner.git  
Cloning into 'mp3cleaner'...  
remote: Counting objects: 11, done.  
remote: Compressing objects: 100% (9/9), done.  
remote: Total 11 (delta 2), reused 10 (delta 1)  
Receiving objects: 100% (11/11), 4.15 KiB, done.  
Resolving deltas: 100% (2/2), done.

Now you can make any changes or add any files. If you are adding new files you will have to use the command

$ git add filename

to add the file to source control. To commit the changes you made, use the command

$ git commit -a

This will open your favorite editor to make a comment on what you are committing.
Now use this command to push your changes out to GitHub

$ git push

Now if you have a new project you want to push to GitHub, you first login to your github account and create a repository for the project you are working on. You should see a button on the home screen of Github which says “New repository” -

Now fill out the repository name. The other fields are all optional.

Click on the Create repository button to create your repository. Thats all you need to do on github.
For the sake of this article, lets call this repo “Hello-World”

Now back on your mac’s terminal go to where your project lives, initialize your repo locally. Create an empty README file if it doesnt exist already. Add all files to your local repo. Commit your changes and then create a remote named “origin” pointing at your GitHub repo. The last step is to push your changes to github.

$ cd ~/projects/Hello-World  
$ git init  
$ touch README  
$ git add *  
$ git commit -m 'first commit'  
$ git remote add origin https://github.com/username/Hello-World.git  
$ git push origin master

Now if you go see your repository on github, you should see your files there.

Enjoy!

Easy Date Math in Python

2012-09-16T13:08:00-04:00

I was trying to figure out exactly when does 6 months end from a given date. Well one can just count the days and find out but I thought python would be a good way to do it quickly. Here is one way to do it…

>>> from datetime import date  
>>> from dateutil.relativedelta import relativedelta  
>>> date(2012,12,14)+relativedelta(months=+6)  
datetime.date(2013, 6, 14)

Southern Fried Chicken

2012-08-29T17:32:00-04:00

I had some craving for some fried chicken the other day and didn’t want to wait an entire day for marination. So looking around on the Internet I found Paula Deens recipe which was quick to make!

Here is my take on it.

Python development setup on Mountain Lion - part 3

2012-08-29T09:17:00-04:00

Eclipse setup to use python in a virtualenv

Today I will cover how to setup eclipse to use the python virtual environment we created in part 2 of this series.

When you first launch eclipse, create a new workspace and call it something relevant. I called it python273 and created it in a folder called workspaces under my home directory.

Now go ahead and install the pydev module. You can do this by going to the eclipse marketplace. Click on Help->Eclipse Marketplace… Type pydev in the find textbox and click on Go. Click on the Install button when you see the pydev module. Go through the installation and restart eclipse.

Once eclipse comes back, open the preferences pane and expand pydev and select the Interpreter - Python option. Click on the New button on that screen and fill out the interpreter name and select the path to python in your virtualenv (/Users/username/.virtualenvs/python273/bin/python)

Click on OK. Now you will be greeted with a dialogue box of all modules eclipse found. Select all the modules you want to add to eclipses pythonpath. I will just let the default be. just make sure you are only using paths which are in your virtualenv. Click OK. You should now see the following dialogue box

Just click on Proceed anyways. Then click on OK on the preferences page. Now lets switch to pydev mode in eclipse and test our Install.

Create a new PyDev Project (just select the defaults) and then create a new module. Just write some print statements, save and then setup some debug points and start debugging the code. You should now be able to switch to debug mode and step through code.

In the next part I will walk through setting up git on your laptop and integrate github into your workflow.

Python development setup on Mountain Lion - part 2

2012-08-28T09:06:00-04:00

Setup virtualenv and virtualenvwrapper on Mountain Lion

In the previous part I went through how you can have multiple versions of python and how you can easily switch between these versions. Here I will go through setting up virtualenv and virtualenvwrapper. This will help in keeping all the different versions of python pristine!

First step is to switch to whatever version of python you want to do this in. Use the bash switch function we created in part 1.

Next we will install the python package installer pip using the easy_install command.

$ sudo easy_install pip

Update Start (July-02-2013) **
The above step is not correct. On my mac, I assumed easy_install to be part of the python version I switched to, as described earlier. When I checked the location for easy_install, I realized that it was from the python version shipped with mountain lion. easy_install is part of setuptools. So we will need to install setuptools for the version of python you are using. As I am trying to switch to version 2.7.3, I did the following to install this -

$ curl -o setuptools-0.6c11-py2.7.egg https://pypi.python.org/packages/2.7/s/setuptools/setuptools-0.6c11-py2.7.egg  
$ sudo sh setuptools-0.6c11-py2.7.egg  
$ rm setuptools-0.6c11-py2.7.egg

You should now see easy_install in the bin directory of the python version you are trying to do this for. In my case for python 2.7.3 it was in “/Library/Frameworks/Python.framework/Versions/2.7/bin”. As this is already in my path thanks due to the python switching tools I showed you all from the first part in this series, we are good. Now run “sudo easy_install pip” as first directed and it will install pip in the right version of python.
* Update Finish (July-02-2013)*

Next install virtualenv and virtualenvwrapper using the pip command

$ sudo pip install virtualenv  
$ sudo pip install virtualenvwrapper

Now add the following variables in your .bash_profile before activating virtualenvwrapper

export WORKON_HOME=$HOME/.virtualenvs  
export PIP_VIRTUALENV_BASE=$WORKON_HOME  
export PIP_RESPECT_VIRTUALENV=true

. /usr/local/bin/virtualenvwrapper.sh

Yes, there is a dot in front of /usr/local/bin/virtualenvwrapper.sh
Next go ahead and create the directory .virtualenvs in your home directory.
Now that all the ground work is done, go ahead and source your .bash_profile file. You should see virtualenvwrapper initialise some files and folders under .virtualenvs. It will also add some more bash functions to make your life easy.

Now lets go ahead and create a virtual environment for our python 2.7.3 environment.

$ mkvirtualenv python273  
Installing setuptools...........  
Complete output from command
/Users/username/.v...python273/bin/python -c "#!python  
\"\"\"Bootstra...sys.argv[1:])

" /Library/Python/2.7/...ols-0.6c11-py2.7.egg:  
Traceback (most recent call last):  
File "<string>", line 279, in <module>  
File "<string>", line 207, in main  
File "build/bdist.linux-i686/egg/setuptools/__init__.py", line 2,
in <module>  
File "build/bdist.linux-i686/egg/setuptools/extension.py", line 1, in
<module>  
File
"/Users/username/.virtualenvs/python273/lib/python2.7/distutils/__init__.py",
line 16, in <module>  
exec(open(os.path.join(distutils_path, '__init__.py')).read())  
IOError: [Errno 2] No such file or directory:
'/System/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/distutils/__init__.py'  
----------------------------------------  
...Installing setuptools...done.  
Traceback (most recent call last):  
File "/usr/local/bin/virtualenv", line 8, in <module>  
load_entry_point('virtualenv==1.7.2', 'console_scripts',
'virtualenv')()  
File "/Library/Python/2.7/site-packages/virtualenv.py", line 942, in
main  
never_download=options.never_download)  
File "/Library/Python/2.7/site-packages/virtualenv.py", line 1052, in
create_environment  
search_dirs=search_dirs, never_download=never_download)  
File "/Library/Python/2.7/site-packages/virtualenv.py", line 598, in
install_setuptools  
search_dirs=search_dirs, never_download=never_download)  
File "/Library/Python/2.7/site-packages/virtualenv.py", line 570, in
_install_req  
cwd=cwd)  
File "/Library/Python/2.7/site-packages/virtualenv.py", line 1020, in
call_subprocess  
% (cmd_desc, proc.returncode))  
OSError: Command /Users/username/.v...python273/bin/python -c
"#!python  
\"\"\"Bootstra...sys.argv[1:])

" /Library/Python/2.7/...ols-0.6c11-py2.7.egg failed with error code 1

Turns out this error came up because Xcode which was installed from the App Store on my new Mountain Lion install did not come with the command line tools. Go to the preference pane in Xcode, and then go to the downloads section and install the command line tools. Once done, go ahead and try the mkvirtualenv command again.

$ mkvirtualenv python273  
New python executable in python273/bin/python  
Installing setuptools............done.  
Installing pip...............done.  
virtualenvwrapper.user_scripts creating
/Users/username/.virtualenvs/python273/bin/predeactivate  
virtualenvwrapper.user_scripts creating
/Users/username/.virtualenvs/python273/bin/postdeactivate  
virtualenvwrapper.user_scripts creating
/Users/username/.virtualenvs/python273/bin/preactivate  
virtualenvwrapper.user_scripts creating
/Users/username/.virtualenvs/python273/bin/postactivate  
virtualenvwrapper.user_scripts creating
/Users/username/.virtualenvs/python273/bin/get_env_details  
(python273) $ which python  
/Users/username/.virtualenvs/python273/bin/python

Notice how my prompt changed to show me which environment I am in. What mkvirtualenv did was create a copy of python under your .virtualenv directory and made it your default python.

There is a slight problem now. I checked to see the version of python in the virtualenv I just created and found that the version was 2.7.2. This was the version of python which came with the new install of Mountain Lion. After reading the docs on mkvirtualenv I found that mkvirtualenv used the python version which was used to install virtualenv.. but I thought I had switched my version of python to use python 2.7.3.. Grr.. ok.. whatever!! I deleted my newly created virtualenv and recreated a new one again, but this time define the version of python I want to use.

$ rmvirutalenv python 273  
Removing python273...  
$ mkvirtualenv python273 -p
/Library/Frameworks/Python.framework/Versions/2.7/bin/python2.7  
Running virtualenv with interpreter
/Library/Frameworks/Python.framework/Versions/2.7/bin/python2.7  
New python executable in python273/bin/python  
Installing setuptools............done.  
Installing pip...............done.

This confirms my python :)

Update Start (July-02-2013) **
If you followed my latest update from above for confirming the existance of easy_install in the version of python you want and installing if missing, the problem with creating the virutalenv instance with the wrong version of python will not happen.
** Update finish (July-02-2013)

Here are some commands you will using -

deactivate - command will disable the virtual environment.
workon - command by itself will show you a list of virtual environments. To activate a particular environment, run ‘workon env_name’
rmvirtualenv - run this command with the name of virtualenv to delete that environment

The best part is that now you don’t need the sudo command to install any new python modules as python is now under your home directory. You are now free to use the pip command to install any module you want.

In the next part I will show you how to setup eclipse to use a virtual python environment…

Tip on burning dual layer disc on your MacBook Pro

2012-08-28T08:10:00-04:00

You can solve this problem by disabling the sudden motion sensor on your laptop. When the disk is going from layer 1 to layer 2, the sudden motion sensor activates which then disrupts the burn process. So if you want to burn a dual layer disc, disable the sensor.. But there is a risk! If you drop you’re Mac book, the superdrive will be damaged. Your bes option would be to disable your sensor.. Burn your disc and then finally enable the sensor again.

to disable sudden motion sensor start Terminal and type:

$ sudo pmset -a sms 0

to reactivate type:

$ sudo pmset -a sms 1

Another thing I found was always try to use Verbatim, all the other brands didn’t work out.

Python development setup on Mountain Lion - part 1

2012-08-27T14:13:00-04:00

Install and manage multiple versions of Python on Mountain Lion

When I was doing python development on snow leopard, I was installing python modules up the wazoo on the default python which came with the laptop. I later realized that was not the right way of doing it. To keep things organized will help you when you are developing projects which have different requirements and using different modules and different versions. So when I installed mountain lion on my laptop, I thought this time around I will be more organised about this.

I didn’t want to mess with the system python, so I installed python2.7.3 and python 3.2.3 from python.org. After each install, the installation process is going to try to add that version of python in your path on the command line. I recommend commenting the lines which were added to your .bash_profile

Now one would wonder how would you switch between the different versions of python? This is where I found this cool script called update_python_switchers.py from http://sevas.github.com/2011/06/14/multiple-python-osx.html

When you execute this script, it creates a file called .python_switchers.sh in your home directory. This script detects all the different versions of python and then creates bash functions based on what were found. To make these functions available, just source this file. As recommended on the site just add the following to your .bash_profile

export OLD_PATH=$PATH

. $HOME/.python_switchers.sh

When I ran the script update_python_switchers.py, I saw the following error in the execution.

Using generic prompt  
Adding System Python 2.3  
Adding System Python 2.5  
Adding System Python 2.6  
Adding System Python 2.7  
Adding MacPython 2.7  
Adding MacPython 3.2  
Adding EPD 32 7.3  
Traceback (most recent call last):  
File "./update_python_switchers.py", line 152, in <module>  
epd64_versions = detect_epd64_installs()  
File "./update_python_switchers.py", line 48, in
detect_epd64_installs  
directories = os.listdir(os.path.join(EPD64_ROOT, "Versions"))  
OSError: [Errno 2] No such file or directory:
'/Library/Frameworks/EPD64.framework/Versions'

Looks like the script is failing because the script is trying to look for the 64 bit version of Enthought Python. As I don’t have it, I just commented out the call to detect_epd64_installs() in the script. Now running the script gives me the following output

Using generic prompt  
Adding System Python 2.3  
Adding System Python 2.5  
Adding System Python 2.6  
Adding System Python 2.7  
Adding MacPython 2.7  
Adding MacPython 3.2  
Adding EPD 32 7.3  
Saved python switcher bash functions to
/Users/username/.python_switchers.sh

where I replaced my real username with username. As you can see I have a lot of versions of python installed.
Just source the .python_switchers.sh in your home directory and now you should have all the versions of python available to you on a press of a bash function really quickly. All that function is doing is adding the version of python to the beginning of your path.
Type ‘typeset -F’ on the command line to see all bash functions available. In my case I see

declare -f select_epd_32_73  
declare -f select_macpython_27  
declare -f select_macpython_32  
declare -f select_system_python_23  
declare -f select_system_python_25  
declare -f select_system_python_26  
declare -f select_system_python_27

I can now use any of these functions to switch to any of the versions of python installed on my laptop. Typing the function select_macpython_27 will make python 2.7.3 be my default version.

One other change I did to the script update_python_switchers.py was to comment out code which was modifying my command prompt. I like my default prompt. I can easily tell what is my version of python by running the command ‘python —version’.

In part 2 I will go through setting up virtualenv and virtualenvwrapper..

My Code Editors on Mountain Lion

2012-08-26T14:12:00-04:00

I started playing with sublimetext2 sometime before I upgraded to Mountain Lion. I liked it a lot. Specially for python programming.

In order to be able to invoke it from the terminal, I created a sym link to my bin directory. Just make sure your bin folder is in your path. You can do that by adding this to your .bash_profile

export PATH=$HOME/bin:$PATH

The other editor I like to use is TextWrangler. After I installed it from the apple app store, I found out that the command line tools are not available if you get this from the app store.. something about not being upto apples standards. You can install the command line tools by going to their site.

Once installed, you now use the edit command to open textwrangler from the terminal.

The below command will open the file.

$ edit filename

Something to keep in mind is that if you already have a textwrangler session open, it will open the file in the same session. If you want a new window use the following command

$ edit --new-window filename

If you want multiple files

$ edit --separate-windows file1 file2

No default Java on Mountain Lion

2012-08-26T00:29:00-04:00

I was trying to see what version of java shipped with mountain lion. So I opened a terminal session and typed ‘java -version’. I was shocked to see that there was no java. Instead I was greeted with a popup windows with an option to install the 1.6 Java Run time.
If one needs to get a jdk, one has to get it from oracle now.

Mountain Lion on my macbook pro

2012-08-26T00:22:00-04:00

I just installed mountain lion on my macbook pro and starting to install all my development tools. This time I plan to document all the things I do for the sake of making notes…

Setup Python 2.7, Django 1.4 with virtualenv and virtualenvwrapper on Bluehost

2012-08-03T15:03:00-04:00

I am pretty excited about learning django. As part of the development process, I wanted to see what it takes to host a django app and thought it would be a good exercise to try it out on my domain. In order to do this, you need shell access. A very important point to make is that the steps below have been edited slightly. I changed the location and my username to be /home/username. You have to replace this to whatever is your location and username on bluehost.

Here are the steps I took to first install Python. The reason I am doing this is so I may have multiple versions of python.

[~]# mkdir python  
[~]# cd python  
[~/python]# mkdir src  
[~/python]# cd src  
[~/python/src]# wget http://www.python.org/ftp/python/2.7.3/Python-2.7.3.tgz  
[~/python/src]# tar xvfz Python-2.7.3.tgz  
[~/python/src]# cd Python-2.7.3  
[~/python/src/Python-2.7.3]# mkdir ~/python/python27  
[~/python/src/Python-2.7.3]# ./configure -prefix=/home/username/python/python27  
[~/python/src/Python-2.7.3]# make  
[~/python/src/Python-2.7.3]# make install  
[~/python/src/Python-2.7.3]# cd ..

Now add the new python bin to your path by adding the following line to your .bashrc file

export PATH=$HOME/python/python27/bin:$PATH

Source the .bashrc file and use the which command to confirm the location of python

[~/python/src]# source ~/.bashrc  
[~/python/src]# which python  
/home/username/python/python27/bin/python

Now install pip to make life easy. The prerequisite is to install setuptools first.

[~/python/src]# wget http://pypi.python.org/packages/source/s/setuptools/setuptools-0.6c11.tar.gz  
[~/python/src]# tar xvfz setuptools-0.6c11.tar.gz  
[~/python/src]# cd setuptools-0.6c11  
[~/python/src/setuptools-0.6c11]# python setup.py install  
[~/python/src/setuptools-0.6c11]# cd ..  
[~/python/src]# wget http://pypi.python.org/packages/source/p/pip/pip-1.1.tar.gz  
[~/python/src]# tar xvfz pip-1.1.tar.gz  
[~/python/src]# cd pip-1.1  
[~/python/src/pip-1.1]# python setup.py install  
[~/python/src/pip-1.1]# cd

Using virtualenv and virtualenvwrapper makes development on multiple projects so much better. I wont go into details here on why this is good.. but trust me on this! Here are the instructions on installing them -

[~]# pip install virtualenv  
[~]# pip install virtualenvwrapper

Create a directory to keep all your virtual environments -

[~]# mkdir .virtualenvs

Now add virtualenvwrapper.sh script to your .bashrc files to add the virtualenvwrapper functions to your environment. Add the following line to your .bashrc. Don’t forget the dot in front of the path.

. /home/username/python/python27/bin/virtualenvwrapper.sh

Source the .bashrc file to activate the functions. Then go ahead and create your virtualenv. I am gonna call mine mydjango.

[~]# source .bashrc  
[~]# mkvirtualenv mydjango

Running the which command will tell you which python you are using. Your prompt will also change to show the virtualenv you are using.

(mydjango) [~]# which python  
/home/username/.virtualenvs/mydjango/bin/python

Now go ahead install all the packages needed for Django while you are in your virtualenv.

(mydjango) [~]# pip install Django  
(mydjango) [~]# pip install MySQL-python  
(mydjango) [~]# pip install flup

Now that all requirements are done, we can start our django project. Lets start with creating a location where all your projects will live. I chose to put this under django_projects in my home directory. I then go in that directory and start a new project and call it myproject.

(mydjango) [~]# mkdir django_projects  
(mydjango) [~]# cd django_projects  
(mydjango) [~/django_projects]# django-admin.py startproject myproject

I now create a directory in my public_html folder. This is so that the project is visible using the url http://www.mydomain.com/myproject

(mydjango) [~/django_projects]# mkdir ~/public_html/myproject  
(mydjango) [~/django_projects]# cd ~/public_html/myproject

Now we will create an fcgi file in your public folder. Lets call it mysite.fcgi. This is to redirect and execute your project. Make note of the first line.. We need to point the script to the python in the virtualenv we created earlier. We will also add the project and virtualenv python instance to the python path..

#!/home/username/.virtualenvs/mydjango/bin/python  
import sys, os

# Add a custom Python path.  
sys.path.insert(0, "/home/username/.virtualenvs/mydjango")  
sys.path.insert(13, "/home/username/django_projects/myproject")  
os.environ['DJANGO_SETTINGS_MODULE'] = 'myproject.settings'  
from django.core.servers.fastcgi import runfastcgi  
runfastcgi(method="threaded", daemonize="false")

Another point which I initially missed was when I created the file mysite.fcgi. When setting the DJANGO_SETTINGS_MODULE, make sure it is set to the name of your project. In this example, I put it as the same name as my django project + .settings which is myproject.settings.

Make sure to set the permissions on the file

(mydjango) [~/public_html/myproject]# chmod 755 mysite.fcgi

Now create a .htaccess file in the same place -

AddHandler fcgid-script .fcgi  
RewriteEngine On  
RewriteCond %{REQUEST_FILENAME} !-f  
RewriteRule ^(.*)$ mysite.fcgi/$1 [QSA,L]

Now turn off your virtualenv with the deactivate command.

(mydjango) [~/public_html/myproject]# cd  
(mydjango) [~]# deactivate  
[~]#

Now if you go to http://www.yourdomain.com/myproject/ you should see your standard Django-powered page.

Here are the sources I used to get this to work. Thanks for all the help!

http://simplyargh.blogspot.com/2012/04/python-27-django-14-on-bluehost.html
http://alextreppass.co.uk/getting-django-working-on-bluehost

http://blog.ruedaminute.com/2011/01/2011-installation-instructions-for-django-on-bluehost/

My take on Butter Chicken

2012-02-09T10:50:00-05:00

This is my favourite chicken dish of all time! Not many restaurants in Boston serve this dish! The only place I know of which even comes close to the same taste one would get in delhi is Royal Bengal in Cambridge.. Here is my version..

This dish is actually made from tandoori chicken but who has a tandoor.. so this has been modified so it can easily be made at home. For those who are vegetarian can just skip the chicken and either put fried tofu or fresh Paneer. As we only get frozen paneer in the indian stores, what I do is put the frozen paneer in hot water which completely softens the paneer.

Marinate for Chicken

Cut 4 pounds of boneless skinless chicken thighs into bite size pieces
Season with 2 tsp salt, 1 tsp pepper, ¼ cup lemon juice, 2 table spoon ginger garlic paste and cover and put in ur fridge for 30 minutes
Whisk ½ tsp red pepper powder, ½ tsp cumin powder, ¼ tsp turmeric pdr, 1tsp garam masala, 1 and ½ tsp coriander powder in 1-2 cups of yogurt and mix into the chicken. Put this back in the fridge for a couple of hours.

Broil chicken
Spread the chicken in a drip tray and broil in the oven on high for 6-8 minutes on each side. Keep this aside.

The Curry

Heat a heavy bottom pan.. add 2 table spoon of butter and about ½ cup of cooking oil ( I used olive oil (not evoo)).
Once oil is hot add about 1 tsp fenugreek seeds and let it crackly a bit.. don’t burn this as this can give your dish a bitter taster
once the fenugreek seeds are a little darker than its normal color add 2 finely sliced onions. ( I only use red onions in Indian cooking)
Cook till it is soft and light brown.. We don’t want it to be completely caramelized.
Now add 2 table spoons of ginger garlic paste. It will make a world of a difference if you use fresh ginger garlic instead of the store bought.
Slit 2 Thai chili peppers and add that to your pan
Now add your dry spices at this point. I put in salt, red pepper powder, garam masala, coriander powder, turmeric.
keep mixing the onions and the spices you added and don’t let things stick to the pan. You also don’t want to over cook the spices you added. Maybe 30 seconds depending on the heat.
Add two cans of crushed tomatoes and mix the onions into the sauce. Give this a good boil and keep mixing making sure it doesn’t stick to the pan.
crush 2-3 tsp of kasoori methi (dried fenugreek leaves) in your hands and mix into the pan
Now add half a pint of light cream into the sauce and mix.
Bring it to a boil and then add the broiled chicken.
Another boil should finish the dish.

Tandoori Style Butter Chicken at Trader Joe’s

2011-10-19T00:04:00-04:00

Time Machine backup to an SMB share

2011-06-16T14:55:00-04:00

If you wanted Time Machine to work with a SMB share you will notice that it doesn’t work. To get it to work run the following command on the terminal

defaults write com.apple.systempreferences TMShowUnsupportedNetworkVolumes 1

Now mount your SMB share. The best way to do this is to click on Go->connect to server in the finder menubar. For the server name you enter smb://username:password@servername/ShareName. I have a NAS share which has a Guest account with no password so my address looks like smb://Guest:@MyServer/HomeShare/TimeMachineBackup. Notice I have no password as no password is required for the Guest account.
Once mounted, Control click on the Time machine in the Dock and select “Time Machine Preferences”. Once you select “Choose Backup Disk” you should see your SMB Share. Select your SMB share and start your Time Share.
Once Time Machine backup starts, you will see an error saying that it cannot create image error. This error message was expected as Time machine could not create a file called a sparsebundle which is required for it to work on the SMB share. For this to work correctly we will be creating the sparsebundle manually. Turn off Time machine and go on to the next step.

In your terminal find your mac address. You can do this with the following command

ifconfig en0 | grep ether

Make a note of the mac address without the colon symbol. For this example lets assume my MAC address is 000000000000. Decide on how big you want your Time Machine backup to be. Then run the hdiutil command to create your sparsebundle file. Here is the command I created for a 200 Gigs share -

hdiutil create -size 200g -fs HFS+J -volname “OSX Backup” mymac_000000000000.sparsebundle

Where 000000000000 is my mac address and mymac is my mac name.

Now copy the sparsebundle file to the Nas share you want for your backups. In my case I copied it to the TimeMachineBackup folder on my NAS Share.

Go back to Time machine preferences and turn on Time machine. You should now see that the backup starts up with no problems. Depending on what you are backing up, the first time it runs it might take some time.

An important point to note is that for this to work your NAS Share needs to be mounted. You can set Time machine to be manual so that way you can make sure your SMB share is mounted before you click on the backup button.

Super Secret features and shortcuts for your MAC

2011-05-18T15:28:00-04:00

http://mac.appstorm.net/how-to/os-x/30-super-secret-os-x-features-and-shortcuts/

terminal tips and tricks for Mac OS X

2011-05-12T15:26:00-04:00

Some interesting tips and tricks for Mac OS X

http://apple.stackexchange.com/questions/5435/terminal-tips-and-tricks-for-mac-os-x

Mac Book Pro terminal prompt keeps changing

2011-05-03T11:00:00-04:00

It got really annoying for me to see how my terminal prompt kept changing when I connected my Macbook from work. I know I can easily fix this by setting my prompt in my .profile file, but I wanted to find out why this was happening.

It turns out the culprit is DHCP. By Default the hostname is determined dynamically on your macbook. This setting can be controlled with your /etc/hostconfig file. Here is what it looks like on my laptop -

\# This file is going away

AFPSERVER=-NO-  
AUTHSERVER=-NO-  
TIMESYNC=-NO-  
QTSSERVER=-NO-

To hard code a hostname, I just added “HOSTNAME=MyMacBook” (without the quotes) to the end of the file.
Some might consider the changing hostname a feature they might want. Not me ;)

Convert uploaded file with windows line ending to unix line ending using php

2011-04-01T11:56:00-04:00

I was working on a php project which involved in uploading a tab delimited file and then processing it with php on a unix webserver to insert data into a database. The file was produced on a windows machine and so it had windows line endings. I used this trick to strip off the carriage return and line feed characters from each line of the file

Here is my html form code

<form action="process.php" method="post" enctype="multipart/form-data" name="import_form">  
<input type="hidden" name="action" value="import">  
<input type="file" name="import_file">  
<br>  
<input type="button" value="Import">  
</form>

Here is my code for process.php

if(isset($_POST['action']) && ($_POST['action'] == 'import')) {  
    if (is_uploaded_file($_FILES['import_file']['tmp_name'])) {

        // Move the file into an accessible location.  
        $buffer =
        trim(implode("",file($_FILES['import_file']['tmp_name'])));

        // Convert all line-endings to UNIX format  
        $buffer = str_replace("\r\n", "\n", $buffer);  
        $buffer = str_replace("\r", "\n", $buffer);

        //code for processing $buffer

    else {  
        echo "<p class=\"Error\">File Upload Failed.</p>";  
    }  
}

Argument list too long

2011-03-29T20:02:00-04:00

Ever tried to copy or move a bunch of files on unix/linux from one location to another and see this error message spit out back at you??

$ cp /var/www/* /var/tmp  
-bash: /bin/cp: Argument list too long

UNIX find to the rescue…

$ find /var/www -type f -name '*' -exec cp {} /var/tmp/. \;

problem solved!

MAMP - mysqlcheck: Got error: 1045

2011-03-28T13:18:00-04:00

I started seeing the following error every time I started MAMP. This began after I changed the root password for my MAMP installation on my MBP.

/Applications/MAMP/Library/bin/mysqlcheck: Got error: 1045: Access
denied for user 'root'@'localhost' (using password: YES) when trying to
connect

It turns out that the default password is hard coded in different scripts under ‘/Applications/MAMP/bin’. In my case the script quickCheckMysqlUpgrade.sh. Just edit the script to have the correct password and you wont see the annoying message again.

Join multiple mpg files into one file

2011-03-23T22:19:00-04:00

I did not realize how easy it is to join mpeg files with just the windows copy command or unix cat command.

On windows the trick is to copy the files while preserving the binary format. Here is how you would do it -

copy files1.mpg /b + files2.mpg /b joinedfile.mpg

You could also do the following if the files are named consecutively

copy file*.mpg /b joinedfile.mpg

On UNIX it couldn’t be any simpler..

cat file1.mpg file2.mpg > joinedfile.mpg

cat file*.mpg > joinedfile.mpg

SSH aliases

2011-03-23T13:07:00-04:00

I constantly connect to linux or unix machines using ssh and it can get a little bothersome typing out the IP address repeatedly. Here is the typical syntax to connect to a linux box (random IP address)

ssh username@111.11.11.1

Now imagine typing this again and again. Thats where the ssh config file comes to the rescue. You can create aliases to make your connections easy. The template for the config file \~/.ssh/config is

Host domain  
HostName domain.com  
User username

The HostName can be either an IP address or an actual hostname. So for our example above our config file \~/.ssh/config will be

Host myserver  
HostName 111.11.11.1  
User username

So now all you have to do is type the following on your command prompt to directly connect and get prompted for a password.

ssh myserver

You can create as many aliases as you want. You can also override the settings of the config file on the command line. For instance if you want to connect using a different user, you can do the following

ssh diff_user@myserver

You can also specify your SSH keys for the server in your config file as shown below so you don’t have to enter your password

Host myserver  
HostName 111.11.11.1  
User username  
IdentityFile ~/.ssh/id_dsa

Click here or here to read how to generate and use SSH keys.

Unix tar tricks

2010-08-31T10:10:00-04:00

One of the most common programs on Linux systems for packaging files is the venerable tar. tar is short for tape archive, and originally, it would archive your files to a tape device. Now, you’re more likely to use a file to make your archive. To use a tarfile, use the command-line option -f . To create a new tarfile, use the command-line option -c. To extract files from a tarfile, use the option -x. You also can compress the resulting tarfile via two methods. To use bzip2, use the -j option, or for gzip, use the -z option.

Instead of using a tarfile, you can output your tarfile to stdout or input your tarfile from stdin by using a hyphen (-). With these options, you can tar up a directory and all of its subdirectories by using:

tar cf archive.tar dir

Then, extract it in another directory with:

tar xf archive.tar

When creating a tarfile, you can assign a volume name with the option -V . You can move an entire directory structure with tar by executing:

tar cf - dir1 | (cd dir2; tar xf -)

You can go even farther and move an entire directory structure over the network by executing:

tar cf - dir1 | ssh remote\_host "( cd /path/to/dir2; tar xf - )"

GNU tar includes an option that lets you skip the cd part, -C /path/to/dest. You also can interact with tarfiles over the network by including a host part to the tarfile name. For example:

tar cvf username@remotehost:/path/to/dest/archive.tar dir1

This is done by using rsh as the communication mechanism. If you want to use something else, like ssh, use the command-line option —rsh-command CMD. Sometimes, you also may need to give the path to the rmt executable on the remote host. On some hosts, it won’t be in the default location /usr/sbin/rmt. So, all together, this would look like:

tar -c -v --rsh-command ssh --rmt-command /sbin/rmt -f username@host:/path/to/dest/archive.tar dir1

Although tar originally used to write its archive to a tape drive, it can be used to write to any device. For example, if you want to get a dump of your current filesystem to a secondary hard drive, use:

tar -cvzf /dev/hdd /

Of course, you need to run the above command as root. If you are writing your tarfile to a device that is too small, you can tell tar to do a multivolume archive with the -M option. For those of you who are old enough to remember floppy disks, you can back up your home directory to a series of floppy disks by executing:

tar -cvMf /dev/fd0 \$HOME

If you are doing backups, you may want to preserve the file permissions. You can do this with the -p option. If you have symlinked files on your filesystem, you can dereference the symlinks with the -h option. This tells tar actually to dump the file that the symlink points to, not just the symlink.

Along the same lines, if you have several filesystems mounted, you can tell tar to stick to only one filesystem with the option -l. Hopefully, this gives you lots of ideas for ways to archive your files.

Source - http://www.linuxjournal.com/content/stupid-tar-tricks

Windows registry repair

2010-06-28T07:52:00-04:00

I found this wonderful free tool to repair registry problems.. It’s just simply called registry repair! Saw some good reviews and it is dead simple to use.. Just scan your registry and you will see a list of potential problems. You can then fix and repair the problem right there. Here is a link for it - http://www.glarysoft.com/rr.html

Best Gyro in Boston

2010-04-22T12:06:00-04:00

My quest for the best gyro sandwich is over!! Mick from timony.com/mickzblog pointed me to Zo Boston after seeing my yelp postings on other gyro places I had been to.
I fell in love with the first bite. The sandwich is soo tasty!! Don’t know if it is the tzatiki sauce or the fresh warm pita bread or the tasty pork meat.. Maybe it’s the combination of all the ingredients… It doesn’t matter.. What matters is how tasty this sandwich is!!

Zo Boston is different from other gyro places.. The way they are different is because of how they cook and add the meat to the sandwich. Other places have strips of meat in their sandwich whereas Zo shreds it’s pork meat!!

The last time I was there, I found out that they also serve a chicken and pork combination Gyro sandwich… Man!! I love it when places have secret menu items.. It makes the experiance even more fun!! I can’t wait to go back and try the chicken and pork combination Gyro!!

Rotini with cheese and cream

2010-03-29T20:08:00-04:00

Pasta is comfort food for me, specially white sauce pasta. I try to make the cooking exciting by changing the recipe at times. Some times I add tomato sauce to it or there are time when I have even tried to use cooking wine to my white sauce. I came across this recipe in one of the abundant cook books we have at home. I changed a few things on it and added few ingredients like garlic, chilly flakes etc to make it more flavorful.

The recipe goes like this -

Ingredients

Rotini 1 pound
Egg yolks 3
Freshly grated Parmesan cheese 7 tablespoon
Ricotta Cheese 1 cup
Light Cream 4 tablespoon
Butter 3 tablespoon
Onion (sliced) 1
Green peppers 1
Salt, red chilly flakes and pepper per your taste

Method

Cook pasta as per instruction
Meanwhile mix the egg york, grated Parmesan and ricotta cheese in a bowl. Add the cream and mix it with a fork.
Melt butter in a pan. Saute garlic. Add onions and fry till golden brown
Add green pepper and salt. Cover and cook till the green peppers are tender.
Add the cream sauce and then add pasta.
You can add black pepper and red chilly flakes according to your taste.

This pasta has to be eaten hot as soon as it is cooked. The sauce tend to thicken if you wait.

Convert AVCHD video to MPEG-2

2010-03-29T09:00:00-04:00

I have a Canon H100 video camera. This records high definition 1080p videos in a digital video format called AVCHD. It has been a pain to watch these videos. First of all, these videos don’t even play properly on my single core PC. My 3-4 year old PC is just not powerful enough. The video is choppy and jittery.. It is completely unwatachable. I have been watching these by copying the files to my camera and then connect it to my TV. If I had a PS3, I could take my memory card and plug it directly and watch these videos. But I don’t have a PS3. I normally use my legacy xbox which has XBMC installed, to watch and listen to my media. It sucks big time that I can’t watch my AVCHD videos through my xbox.

So the only way I can watch these AVCHD videos is to convert it to either an MPEG or AVI format. I know.. I know.. I am loosing video quality by down converting these videos… but till the time I can find a better way to watch these videos… thats what I am going to do. I did find a script at videohelp.com which was using some freeware tools.. But in my opinion the process was still painfull. My ideal tool would be to take these AVCHD files and either convert them to AVI/MPEG or to burn them to a DVD directly. Did I mention, I want this to be free too ;). You could also pay \$50.00 for ConvertXtoDVDwhich does that for you.

Another option is to create an AVCHD DVD which will play directly on a blue ray player or a PS3. Scott Johnson, CEO of AtTask has a script on his blog which will create the DVD disc which will play on a blue ray player.

I got my camera in 2007 and in 2010 I have finally found a software which does the video conversion for me. It’s called Free HD Converter V1.6.
The software works on my old machine.. It’s not fast but I just batch all the videos and a couple of hours later, I have the converted MpEG files. I wish this tools would stitch the many generated mpeg files together too. Maybe it does and I haven’t found the setting! I can now burn a DVD from the Mpeg files and send the videos of my new born to my Mom or upload to YouTube directly!!

Next step is to convert all my AVCHD videos to MPEG. Now only if somebody could just do that for me too ;)

Wake up your PC from your xbox using xbmc

2010-03-26T08:00:00-04:00

I read an old lifehacker article on wake-on-lan the other day and decided I should try it on my home network. After using the free tool Magic Packet Sender mentioned in the article successfully, I thought how cool would it be to start my media server from my xbox. I use my xbox as my media center. I know… using the xbox is old news and it is not powerful enough for hi-definition videos… but still.. I like my xbox and am not willing to get rid of it yet. Till something better and cheap comes along, I am going to continue using it. My brother bought the Acer Aspire Revo (AR3610-U9022) which comes with the Intel Atom 330 processor. I like it a lot and boxee and xbmc both work really well. Streaming 1080p video works flawlessly. Hulu desktop works too, but you have to use the adobe flash 10.1 beta drivers for the video to work. let’s see what path I take.

Going back to my little project, I thought I will study what is involved in sending a wake-on-lan packet to a PC to wake it up. I read the lifehacker article in detail and then googled for other resourses. I came across a really nice resource which explained how the wake-on-lan packet is constructed along with a sample java program. I also found, I am not the only person who thought about using this technology to start their media server from their xbox. I found this script in the forums at xbmc.org. Here you go -

# Wake-On-LAN  
#  
# Copyright (C) 2002 by Micro Systems Marc Balmer  
# Written by Marc Balmer, marc@msys.ch, http://www.msys.ch/  
# This code is free software under the GPL

import struct, socket

def WakeOnLan(ethernet_address):

# Construct a six-byte hardware address

addr_byte = ethernet_address.split(':')  
hw_addr = struct.pack('BBBBBB', int(addr_byte[0], 16),  
int(addr_byte[1], 16),  
int(addr_byte[2], 16),  
int(addr_byte[3], 16),  
int(addr_byte[4], 16),  
int(addr_byte[5], 16))

# Build the Wake-On-LAN "Magic Packet"...

msg = '\xff' * 6 + hw_addr * 16

# ...and send it to the broadcast address using UDP

s = socket.socket(socket.AF_INET, socket.SOCK_DGRAM)  
s.setsockopt(socket.SOL_SOCKET, socket.SO_BROADCAST, 1)  
s.sendto(msg, ('<broadcast>', 9))  
s.close()

# Example use  
WakeOnLan('aa:bb:cc:dd:ee:ff') # My PC

The last line in the script is calling the WakeOnLan method and passing the MAC address of the PC you are trying to wake up. You can get the MAC address of your PC from a DOS command prompt by the command -

c:\> ipconfig /all

It will show up as Physical Address in the detailed listing. Read the lifehacker article given above for more details. You can wake more than 1 by duplicating the WakeOnLan(MAC) line in the script.

You have a couple of options of executing this script. You could create a sub-folder under the scripts folder in your XBMC install and place this script there. In which case you can call it anything you want. I have created a script for every PC I want to wake up in my network and have the machine name as the script name. You could also start up all your computers as soon as the xbox is switched on. Just name the file autoexec.py and put it in XBMC/scripts. It’s similar to how autoexec.bat works for windows :)

I also put these scripts on every PC in my network, so I can start any of them remotely. The pre-requisite is that you need python installed on your PC for the script to work. You can either install python or just use the python which comes with cygwin. I recommend that if you need a decent command line on windows, just get cygwin and give mintty a try. Andy Koppe the creator of mintty recommended it to me the other day and I love it so far!! I have also further recommended it to my friends at work. I just launch mintty to get a bash shell on my windows box and execute the script this way -

python /path/to/WakeOnLan.py

Same music album showing up multiple times on iPhone

2010-03-25T08:00:00-04:00

I created an album with different artists to listen on my iPhone. To my horror all the songs showed up as separate albums. Cover flow showed up the same album art for each of the songs. I hated this!! The fix is pretty quick. You have to make sure of the following to make your songs show up as one album -

The Album name has to be the same for all the songs
The Album Artist name has to be the same for all the songs
All the songs should be set as a part of a compilation.

Here is how you do it . Select all the songs in itunes you want to show up as an album, then right click on one of the chosen songs and click ‘get info’ -

Type in the Album Name and Album Artist
Under the options tab, make sure the ‘Part of A Compilation’ is checked and set to Yes.

That’s it… That will make sure the album shows up only once for the entire album

Mutton Curry revisited

2010-03-24T08:00:00-04:00

I got about 5 pounds of goat leg from Waltham India Grocer. It is an Indian grocery store which also sells Halal meat. I normally get my meat from Cheema Supermarket but as I was in the neighborhood, I just picked it up from Waltham.

I was making my usual Mutton Curry and half way through making this dish I realized I did not have any crushed tomatoes.. so I had to improvise. According to some purists Goat curry should not have any tomatoes. I have always used tomatoes.. so this was a chance to try make this dish without tomatoes… Unfortunately I don’t have a picture to show on how this looks. It was a big hit at the party and that’s what counts :)

Here is how I did it…

Ingredients

Goat leg - cut into small pieces (1 inch)
Onions chopped - 4 large onions chopped
cinnamon - one 2 inches stick
cardamom - 5-10
cloves - 10-15
curry leaves - 10-12 leaves
green chillies - 4-5 cut into big pieces
garlic - 5-6 pods finely diced
ginger - finely diced - 2 inches
turmeric - 1-2 tsp
garam masala - 3 tsp
dhaniya powder - 4 tsp
lal mirch - 2 tsp
degi mirch - 2 tsp
jeera powder - 2 tsp
MDF Meat masala - 1 tsp
Coconut milk - 1 can

Method

Heat oil and some ghee in a big flat bottom vessel.
Add Cinnamon, cardamom and cloves
Once it sputters, add onions and start frying it. Add the salt to taste and turmeric immediately.
Once it is slightly brown, add the chopped ginger and garlic.
Continue sauteing and add the chopped green chillies and curry leaves.
Add the mixed powdered spices
Make sure you don’t burn the spices. Add a little ghee like maybe half a teaspoon.
Add the mutton and coat the onions and spices to all the meat.
Sear the meat so it cooks and changes color on all sides.
Add a can of coconut milk and mix everything together. After 5 minutes transfer all to a pressure cooker.
Wash the vessel you used and use the water for the pressure cooker. Put enough to cover the meat.
Add a cut potato in the cooker. This helps bring some texture to the gravy.
Add 2 tsp of MDF Meat masala.
Add chopped cilantro and then pressure cook for 30-40 minutes.
Once the pressure comes down, open and check if the meat is tender. If not tender, pressure cook for 10-15 more minutes.
Taste and adjust salt
Serve with Rice

First warm weekend of 2010

2010-03-22T14:13:00-04:00

What a phenomenal weekend! It was nice and sunny and the best part was it was not too hot!

Food wise it was a great weekend too.. Friday night we went to Aegean in Framingham. Their bread basket had fresh baked bread and Pita bread. This came with olive oil and feta cheese. This combination was so tasty. The key here was their Pita Bread… Fresh!! For the main course I order their Lamb Youvetsi. This was a Braised Lamb with Orzo topped with grated cheese. My wife ordered the Mousaka. This was Layered eggplant, seasoned with ground meat, grated cheese and topped with bechamel sauce (My wife is a sucker for white sauce!!). Both dishes were very tasty and very very satisfying. My lamb was very succulent and tender.. This was cooked with tomato sauce, spices and orzo. Did I mention it was very satisfying! My wife’s dish was good too… It was like a lasagna but instead of pasta, made with eggplant.

Saturday morning was bright and started out a little cold but it got warmer and ultimately turned out to be a great day to spend outdoors. We met our cousins in Waltham and spent a couple of hours in Waltham Commons. Got coffee, coffee cake and cranberry scones from Cafe on the Common. A very cute place with some tasty baked goods. The place looked like an old school coffee shop.

For lunch we ended up in Taqueria Mexico. My Bro-in-law recommended it… looked like a whole in the wall.. the Food was very authentic and the prices were not too high. Check out their menu and you will find some traditional mexican dishes not available else where in and around Boston. I wanted to get the Guisado De Carne De Puerco. This was slices of pork cooked in green sauce with onions and peppers. Serverd with rice, beans, salad and tortillas. I got chicken instead of the pork. My Bro-in-law got the original dish and we both agreed the dish was better with the chicken. The corn tortilas were fresh and went good with both our dishes. My wife and my cousin got a chili reileno stuffed with cheese, one green chicken enchilada with melted cheese and sour chream on top. Served with rice, beans, saland. All dishes were yummy.

We then walked down Moody street to do some Indian Grocery Shopping at Patel Brothers. While I was there I got 5 pounds of goat leg from waltham india grocery planning to make mutton curry on sunday (will share the recipe). We then stopped by Lizzy’s Homemade Ice cream :) where I got their orange pineapple ice cream, my Wife got a chocolate frappe, my cousin got a cup of vanilla ice cream whereas my bro-in-law got some green tea.

Sunday was another fabulous day. The Sri Laxmi temple in Ashland was going to Host the Indian festival Holi. I guess the temple chose to celebrate it on Sunday as it was really cold the weekend of the actual day. I celebrated Holi after like 10 years. It was a lot of fun playing with colors with your friends and family. The organizers did a good job of arranging for food and Indian tea for everybody. I then invited my cousin’s family along with other family friends for an impromptu dinner.

Dinner was a big hit! I made mutton curry, Priya made okra fry. My Cousin Vandana made an awesome dish with Cholle and Spinach. The family friends made Channa Dal. It turned out to be a feast and a fitting end to this awesome weekend!!

A better cygwin terminal - rxvt

2010-03-22T09:00:00-04:00

rxvt is so much more better than the default cygwin terminal. You will thank me once you switch to using rxvt. To get rxvt working without starting an X Server, just create a shortcut using the following -

C:\cygwin\bin\rxvt.exe -fg green -bg black -fn "Lucida Console-14" -e bash --login -i

This is assuming that cygwin is installed at c:\cygwin. You will get a foreground of green on a black background. The font used is “Lucida Console-14”.

Grep for multiple stuff in one file

2010-03-19T08:00:00-04:00

Update: See the comments to see how to do this correctly.

the UNIX grep command is very powerful, but the egrep command is even more powerful. You can grep for multiple strings in a file using the egrep command. All you have to do is use the | (or) symbol and add as many strings you want to look for. You have to make sure you escape the pipe symbol. This is especially helpful when you are wading through log files.

egrep "string 1"\|"string 2" file.log

The above command will only display the lines from file.log that have a reference of “string 1” and “string 2”

Find user details on windows user in a domain

2010-03-18T10:05:00-04:00

Here is a quick tip - If you want to find out details on a user account on windows inside a particular domain use the following command on a DOS prompt

net user username /domain

where the username is the username of the account you want more information on. You can get a lot of information with this simple command. Some of the things you can find are -

Full Name
Account information - if active, if expires
Password - last set, expiry date etc
last logon date and time
group memberships

Hit by AV.EXE virus - Cannot run any programs

2010-03-09T20:58:00-05:00

I recently got hit by the AV.EXE virus. This was a really bad virus. I got a pop up which looked as if it was a windows dialogue box telling me to install this anti spyware/virus scanner. I knew I should not even touch the popup box and not even try to close it. The best way was to kill the process. I used System Explorer which is a really good task manager replacement. It is available from http://systemexplorer.mistergroup.org/. Once killed, I wanted to make sure I wasn’t infected any further, so I started my spyware scanner called Malwarebytes’ Anti-Malware.
Surprisingly the AV.EXE popup came back and on top of that, my malwarebytes program didn’t start. I killed the process once again and tried starting malwarebytes again and the av.exe popup came back!! Damn what was going on!! After killing the process once again I tried Superantispyware and it too had the same fate… I was now perplexed… I tried a random program and after seeing that too starting the popup, finally realized that my registry had been compromised. The virus had modified my registry in such a way that it would get enabled whenever I would start any executable. I then thought I better do more research on this before I screw up my machine even further. Here is what I found

I had to right click the Malwarebytes program and use Run as.. to run the program as administrator
Update the program and run it and remove any instance of the virus
Cleaning the system broke my registry. I couldn’t run any programs anymore.

The registry was fixed with the following file. Save the following in a file called fixit.reg

Windows Registry Editor Version 5.00

[HKEY_CLASSES_ROOT\.EXE]  
@="exefile"  
"Content Type"="application/x-msdownload"

[HKEY_CLASSES_ROOT\.EXE\PersistentHandler]  
@="{098f2470-bae0-11cd-b579-08002b30bfeb}"

[HKEY_CLASSES_ROOT\exefile]  
@="Application"  
"EditFlags"=hex:38,07,00,00

"FriendlyTypeName"=hex(2):40,00,25,00,53,00,79,00,73,00,74,00,65,00,6d,00,52,\

00,6f,00,6f,00,74,00,25,00,5c,00,53,00,79,00,73,00,74,00,65,00,6d,00,33,00,\

32,00,5c,00,73,00,68,00,65,00,6c,00,6c,00,33,00,32,00,2e,00,64,00,6c,00,6c,\  
00,2c,00,2d,00,31,00,30,00,31,00,35,00,36,00,00,00

[HKEY_CLASSES_ROOT\exefile\DefaultIcon]  
@="%1"

[HKEY_CLASSES_ROOT\exefile\shell]

[HKEY_CLASSES_ROOT\exefile\shell\open]  
"EditFlags"=hex:00,00,00,00

[HKEY_CLASSES_ROOT\exefile\shell\open\command]  
@="\"%1\" %*"  
"IsolatedCommand"="\"%1\" %*"

[HKEY_CLASSES_ROOT\exefile\shell\runas]

[HKEY_CLASSES_ROOT\exefile\shell\runas\command]  
@="\"%1\" %*"  
"IsolatedCommand"="\"%1\" %*"

[HKEY_CLASSES_ROOT\exefile\shellex]

[HKEY_CLASSES_ROOT\exefile\shellex\DropHandler]  
@="{86C86720-42A0-1069-A2E8-08002B30309D}"

[-HKEY_CURRENT_USER\Software\Microsoft\Windows\CurrentVersion\Explorer\FileExts\.exe\UserChoice]

Double click and execute this file. This fixes the registry

I think I should start browsing in a sandbox so I don’t infect my pc. I have a couple of options, but I think the easiest is to just use Sandboxie which is available at http://www.sandboxie.com/. Other ways I can protect myself is to use a limited user account instead of using an admin account. Ohh well…

How to go from Oracle to CSV

2010-03-04T02:37:00-05:00

Here is a trick to to create a comma delimited file from oracle. You can create a csv file simply using SQL*Plus with any sql statement. Here it is -

set feedback off  
set heading off  
set underline off  
set colsep ','  
spool /path/to/file/mysheet.csv  
select * from table  
spool off

Here are some other ways I found online

http://kpanchan.blogspot.com/2007/05/extract-csv-file-format-in-oracle-sql.html

http://www.oracle-base.com/articles/9i/GeneratingCSVFiles.php

http://asktom.oracle.com/pls/asktom/f?p=100:11:0::::P11_QUESTION_ID:68212348056

Sending email with a body and attachment from UNIX

2010-03-03T03:40:00-05:00

It is easy to send a text file as the body of an email from UNIX. You can do it using the command below

mail -s "Email with file as body" email@domain.com < message.txt

Now if you want to send a file as an attachment, you would need to use uuencode piped to mail. This file can be a binary file or a text file. The command is

uuencode file.dat file.dat | mail -s "Email with attachment" email@domain.com

Suppose you want to send an email with a body and an attachement. Here is the trick -

First create the attachment file by using uuencode. Then join the message file which is the body of the email along with the uuencoded file which is going to be an attachment by using the cat command.

uuencode file.dat file.dat > attachment.txt  
cat message.txt attachment.txt > combined.txt

Now all you need do is to send the email using the first method shown here

mail -s "Email with message body and attachment" email@domain.com < combined.txt

*** Updated Feb 22, 2013
Here is a script I wrote a while ago using the same concepts as above… I call it mailfile.ksh

#! /usr/bin/ksh

A_FLAG=FALSE

while getopts :a OPTION  
do  
    case ${OPTION} in  
    # -a as attachment  
        a) shift;A_FLAG=TRUE;;  
        \?) print -u2 "Usage: ${PROG_NAME} [-a] EMAIL MAILFILE SUBJECT"  
            exit 2;;  
    esac  
done

EMAIL=$1  
# shift param one position  
shift  
MAILFILE=$1  
# shift param one position  
shift  
# copy rest of params to SUBJECT  
SUBJECT=$*  
# get rid of quotes  
SUBJECT=$(echo $SUBJECT| sed 's/"//g')

if [ $A_FLAG = TRUE ]  
then  
    ATTACHEDFILE=`basename $MAILFILE`  
    /usr/bin/uuencode $MAILFILE $ATTACHEDFILE > /tmp/$ATTACHEDFILE$$  
    MAILFILE=/tmp/$ATTACHEDFILE$$  
fi

/usr/ucb/mail -s"$SUBJECT" $EMAIL < $MAILFILE

# Cleanup  
if [ $A_FLAG = TRUE ]  
then  
    rm /tmp/$ATTACHEDFILE$$  
fi

The usage is the following if you want to attach a file as an attachment

$ mailfile.ksh -a email@example.com yourfile.gz Your Subject

And if you want the file as the body, the usage is -

$ mailfile.ksh email@example.com SomeFile.csv Your Subject

Chicken 65

2010-02-28T05:02:00-05:00

Chicken 65 - Any Hyderabadi’s favorite dish! Well, I guess maybe after the hyderabadi Biryani :) I dont want to start any holy wars here but today my post is about Chicken 65. A lot has been talked about this dish’s origin which can read on wikipedia. I believe Sanjay Thumma from vahrehvah.com got it right by saying this chicken dish was the 65th item on a menu. As popularity rose in the restaurant, it started being called Chicken 65.

I was browsing the internet the other day and came across a friends photo blog of his recent visit to India on facebook. Being a self proclaimed foodie, watching the slideshow was very torturous, but it got some really good memories of places to eat in New Delhi and Hyderabad. I thought the best way to deal with this would be to cook this dish. I called my cousins and we decided to meet up that day. I offered to make chicken 65 as an appetizer and my cousin offered to make spaghetti in a turkey meat sauce(yum) for dinner. Ultimately I was very glad with the results of my dish. I think this has been my best attempt so far in making this dish. Here is the recipe -

Ingredients

Boneless skinless Chicken thighs -3 pounds
Green chillies - 5-6 slit and the cut in half
Ginger - grated 3-4 tsp
Garlic- grated 3-4 tsp
Curry leaves -10-15 leaves
One egg
Corn starch - 4-5 tsp
2 cups yogurt

Spices

Salt - according to taste
Pepper - 1 tsp
Deghi mirch - 3 tsp
Lal mirch - 3 tsp (depending on how spicy you want it)
Garam masala - 1.5 tsp
Haldi - 1/2 tsp
MSG - 1/2 tsp

Method

Cut chicken in small cubes
Marinate chicken for an hour in salt, pepper, deghi mirch 1.5 tsp, lal mirch 1.5 tsp,.5 tsp of MSG
Mix the chicken in the marinate and while mixing add an egg to the chicken and then mix in corn starch
Whisk the rest of the spices in yogurt and keep aside
Fry the chicken and make pakodas and keep aside
Add 2 tbl spoon of oil in a wok, heat and then add Mustard seeds - Add Ginger, garlic, green chillies and curry leaves.
Once a little brown and dry, add the fried chicken and mix it well
Add the whisked yogurt
Cook till yogurt is dried up

Here is what it looked like after the pakoda’s were made -

Here is the finished product -

The reason this does not look red in color was because I did not use any food coloring.

Enjoy