LSST data release production (drp) analysis

• Purpose: A 2-days tutorial on how to analyze LSST-stack output data using python
• For: Someone who wants to use the LSSt stack output and start analyzing it with Python
• Where: CC-IN2P3, Lyon, France
• When: Octobre 3-4, 2017

Outline

1 Introduction ------------

• Short reminder about Python
• What is the LSST stack, and what is it for
• What tools will we be using today

2 The LSST stack --------------

2.1 How to install and setup the LSST stack ` - Stable version and weeklies - Setup of stack-install packages - Install and setup of a non-stack packages 2.2 Overview of the data processing`

• My input data, and what obs_* should I be using
• Tasks and command line tools
• A complete data reprocessing work-flow
• What catalogs are produced and from which step of the pipeline

2.3 Access the data ` - What is a data dataIds - an example with CFHT data - Get and open images - Get and load catalogs - First step to analysis 3 Python libraries for data analysis ---------------------------------- 3.1 Overview of useful python libraries: a non-exausthive list

• useful native functionalities
• numpy
• scipy
• math
• pandas
• matplotlib, seaborn
• astropy
• astroquery
• pyfits, h5py
• yaml, json, (c)Pickle
• healpy

3.2 In more details ` - `numpy` - `astropy` - `scipy` - `matplotlib` - `other` 4 Build a python package for data analysis ---------------------------------------- 4.1 Short tutotial to build a python package- setup.py - pypy - libraries - notebooks - install and test your code localy 4.2 Share your work and make it useful

• git / github: basic functionnalities
• continuous integration: Travis-CI
• documentation: sphinx and readthedoc
• static code analysis (how well my code is written): landscape
• "dynamic" code analysis (make and run my unit/integration tests): codecov

5 Conclusion ----------

TBD

Requirements

Install

• Python 3 (conda install is the easiest way)
• Python libraries from the requirements.txt
• git + a github account

Knowledge

• install python - a lot of way to do that, and that could be a mess
• install a python package
• ipython
• jupyter notebook
• basis knowledge on python