ericbarnhill

Building an Insight project with Python, Flask, and Heroku

A common goal in Insight data science projects is to deploy one's work as an app. The recommended app is Flask due to its dynamic page construction and easy interoperability with SQL databases. Heroku is then a great choice for deployment as it interacts easily with Python and Flask.

However a lot of information is out-of-date or suboptimal. This project will document how I built my Insight project app using Python, Flask, and Heroku.

*This project is adapted from and enormously indebted to the excellent "Flask by Example" tutorial found here. Basically I take this tutorial a little less far, fix some errors and adapt it to the requirements of the Insight project.

Contents

sim642

Installing boost libraries for GCC (MinGW) on Windows

Folder setup

1. Extract downloaded boost source, e.g. C:\Program Files\boost_1_59_0.
2. Create a folder for Boost.Build installation, e.g. C:\Program Files\boost-build.
3. Create a folder within for building, i.e. C:\Program Files\boost_1_59_0\build.
4. Create a folder for installation, e.g. C:\Program Files\boost.

GCC setup

1. Open Command Prompt.

TerryMooreII

import cx_Oracle
 
conn = cx_Oracle.Connection("userid/password@tns")
curs = conn.cursor()
curs.callproc("dbms_output.enable")
 
sqlCode = """
some long
sql code 
with dbms_output