code-classifier-dataset

This repo contains dataset for code classifier to be implemented in oppia-ml.

The dataset is raw and hasn't been tagged yet. These programs are extracted from answer history of euler exploration from Oppia.org.

Datasets

• dataset.json: contains all the programs which were extracted from euler exploration from Oppia (~11000 programs).
• syntax_dataset.json: contains all programs which are syntactically correct (7092).
• compiled_dataset.json contains all programs which are getting ecompiled (7061).
• execd_dataset.json: contains all programs which are getting executed and don’t raise exception during execution (63).
• execd_dataset_with_classes.json: contains all programs in execd_dataset.json with feedback class assigned to them.
• execd_dataset_with_classes.csv: execd_dataset_with_classes.json in CSV format.

Scripts

• data_tag.py: python script to tag source programs stored in above json files.