First, clone this project and change to its directory.

git clone <repo_location>/se-challenge.git
cd se-challenge

Install Homebrew from http://brew.sh/

Install Python:

brew install python

Install Postgres:

brew install postgres

Configure Postgres:

initdb /usr/local/var/postgres

createdb csvstorage

Start Postgres:

postgres -D /usr/local/var/postgres

Set the connection string environment variable. Mine looks like this:

export DATABASE_URL=postgres://alex:@localhost:5432/csvstorage

Create a virtualenv:

virtualenv venv

Activate virutualenv

source venv/bin/activate

Install dependencies:

pip install -r requirements.txt

Start the server:

python src/server.py

Navigate to http://localhost:5000 in Chrome or Fox and follow the on-page instructions.

This application also runs in the cloud on Heroku.

First, get started with Heroku at: https://devcenter.heroku.com/articles/getting-started-with-python#introduction

In the project directory provision a dyno and add postgres to it:

heroku create 

heroku addons:add heroku-postgresql:hobby-dev 

Run locally with:

foreman start

Run in the cloud with:

git push heroku master

heroku open

##Repurposed UI

I repurposed code from another project to quickly iterate on the UI for this one. I am proud of the speedup that allowed.

##Unicode

This project has skeletal unicode support. CSV files are often exported from Microsoft Excel, and nonexpert users can easily end up with files encoded, for example, as UTF-16BE. There is skeletal support for detecting and parsing these encodings included, but I abandoned it when I decided to use PG to ingest the CSVs directly (in the interests of speed).

Add a fade to a spinner image to show the upload is working when there's a slight delay (eg, on Heroku).