PROMISE dataset

folder usage

All the data is placed under the PROMISE/ folder.

source-code/: The source code files of the projects.
labeled-data/: The label csv files of the corresponding source code. name and version describe the exact project and version of the code. The name1 indicates the actual file path of the code. The last bug column indicates the file-wise label. And the remainder columns are the provided hand-crafted features.
embedded-data/: The extracted AST tokens and token embeddings of each code file. The token nodes are selected. Provided by the PROMISE dataset.
token-data/: The AST tokens extracted by me without selection, which preserve as much information as possible.
resources/: Some useful information. code-prefix.json records the actual source code root folders. token-count.json records the number of token-extracted files of each project within the token-data/ folder.

The codes I used to extract AST tokens are placed under preprossed/ folder.

If you want to filter out some token type, go to preprossed/src/get-ast/ast_utils.py and uncomment or add the token type you want to filter out. Then run the preprossed/src/get-ast/main.py. The extracted tokens will be placed under the PROMISE/token-data/ by default (and override the old files).

Name		Name	Last commit message	Last commit date
Latest commit History 4 Commits
PROMISE		PROMISE
preprocessed		preprocessed
promise-dataset-hf		promise-dataset-hf
.gitignore		.gitignore
README.md		README.md