li-lab-mcgill / covid19_media Goto Github PK

For the new GPHIN data created by the GPHIN scraper, experiment on topic mixture ETM, METM, D-ETM and DM-ETM

Travis-CI configuration link : https://travis-ci.org/

Virtual environments are a great way to have all dependencies from a python code using only one command, pip freeze. I'll also add in the readme how to create your own local virtual environment before cloning this repo.

Expected Behavior

The DMETM code by running python main.py --mode train --dataset "GPHIN" --datapath data/GPHIN --num_topics 10 --train_source_embeddings 1 --epochs 10 --min_df 10 should start running and giving epoch, since the data in min_df_10 has timestamps now.

Current Behavior

Index out of bound error is appearing under Getting training data ... for file data.py line 104.

Possible Solution

Add piece of code under a try catch exception that continues the for loop in case this index out of bound error appears.

Steps to Reproduce

1. Add either the official csv or the online parsed csv from the data folder under DMETM/data/GPHIN and rename the csv file to gphin.csv
2. Navigate to the folder DMETM/scripts and run python data_gphin.py. You should a new folder titled min_df_10 appearing under DMETM/data/GPHIN
3. Go back one directory to DMETM and run python main.py --mode train --dataset "GPHIN" --datapath data/GPHIN --num_topics 10 --train_source_embeddings 1 --epochs 10 --min_df 10

Update readme with information on how to run the updated main.py script