Algorithm development repository for Cartograph team.
Team members: Lily Irvin, Lu Li, Yuren "Rock" Pang, Jonathan Scott, Shilad Sen.
- Python 3
- Python modules in requirements.txt
- Git LFS. Instructions on OS-X:
- brew install git-lfs
- sudo git lfs install
- sudo git lfs install --system
- PyCharm Ultimate
- iGraph. Instructions on OS-X:
- brew install cairo
- brew install pkg-config
- brew install igraph
- On OS-X Mojave, Shilad needed to run
sudo installer -pkg /Library/Developer/CommandLineTools/Packages/macOS_SDK_headers_for_macOS_10.14.pkg -target /
- Clone this repo
- In the repo, run git lfs install
- Launch PyCharm, select "open" and the cloned directory.
- In preferences, go to the project settings, add interpreter -> New virtualenv environment with default settings.
- Open requirements.txt. PyCharm will ask you to install everything. Do it. It will take some time.
- Maybe necessary, but we aren't sure: Mark some of the top-level directories as src directories.
-
Add a new shell script for creating map input data. See
bin/make-food-map.sh
for an example template. Run the script. -
Make sure you have git LFS installed (see above). Check in your new map data files, which should live in data/map_name .
-
If you need to update the files in the future, change your shell script, re-run it, and check in the new files.
- Create an experiment script, similar to example-experiment.sh
- Run the script (e.g.
./bin/run-joint.sh
. - The script will tell you where it sends output.
- Only check in "final" experiments.
- You need to locate your
experiment
folder and copydomain_concept.csv
andpopularity_score.csv
fromdata/[you topic]
toexperiment/[your topic]/[your experiment]
. - Copy this experiment folder to catograph-server under the
data
directory. See instruction at catograph-server to proceed.