The purpose of this repo is to give you the opportunity to set up a collaborative workflow and practice using it without polluting your actual project repository for the class. You will also be using this as an opportunity to implement programatic downloading and verification of your data. These are important for reproducibility and will be required of your final project; so we figured it be good to get you in the habit of practicing it now.
Do not clone this directly from the berkeley-stat159 organization!
First, break up into your teams. If your actual project team mates are not in the same lab session as you, just join the others sitting around you. You won't be working on your projects directly, so you don't have to be in your project teams.
Now, select one member from your group to be your "collaboration host". This
person will fork from https://github.com/berkeley-stat159/lab6_exercise.git
Everyone else in the group will now fork from your collaboration host's fork. We're doing it this way so each group can have the same starting point, but all of the practice work from each group is contained within that group.
Once everyone has forked correctly, every group member should clone from their fork and set up their remotes like we did in class on Thursday. You should have two remotes, one pointing to your fork and another pointing to your collaboration host's fork (you don't need to have a remote pointing to the original berkeley-stat159 repo for the exercise).
-
Add a
datadirectory to the project. Create a.gitignoreand add thedatadirectory to it so any data files stored in that directory will be ignored by github -
Implement functions to download and verify data. The prototypes for these functions can be found in
src- You may want to split your group into subgroups and have each subgroup work on a different task. Don't forget about branching so youcan have multiple parallel lines of development.
-
Practice code review via pull requests. Make sure that your implementations run on every group memeber's system (cross-platform compatibility is important)
If you manage to implement and review all of the functions in
src/data_acquisition.py, then move on to the in class example on
hemodynamic response functions.
Start by using your data_acquisition module to download and save the
necessary modules to your src/ directiory. You might start by downloading the
stimuli.pyand
pearson.py
modules. You will also need to download the
condition file
to your data/ directory.
Once you have all the tools you need, collaborate with your group to follow the convolution lecture and create an HRF. Having constructed the HRF, apply the pearson correlation analysis using the HRF as the correlate rather than the binary task on/off condition.