bibdl
📃 Automatically generate BibTeX style references for ml/dl papers.
bibdl
Automatically generate BibTeX style references for ml/dl/cv papers.
Installation
Build locally. You can easily customize the code.
git clone https://github.com/bchao1/bibdl.git
cd bibdl
python3 -m pip install .
Usage
bibdl --list list.txt --bib refs.bib
List all paper titles in list.txt. bibdl will spawn multiple processes to search for the papers in pre-defined sources and generate BibTeX style references in refs.bib.
Papers that are not found will be listed in not_found.txt. You'll have to manually find references for those papers.
For more help, type
bidbl -h
or check the examples/ folder.
Disclaimer
This is not expected to be a search engine, so make sure you get your paper titles exactly right, or you might not be able to find anything. It's an ordeal looking for which conference a paper is submitted to, manually copying references, and finally formatting them, so just let the computer do their work.
If you get most of your paper titles right, the script usually has a hit rate of > 85% . For 100 papers, the search process usually finishes in less than 30 seconds.
For Speech/NLP papers, the hit rate might be lower since I did not implement search for Speech/NLP-specific conferences.
Sources
- OpenReview
- arXiv
- NIPS Proceedings (optional)
- Papers with Code (optional)
TODO
- Search Sources
- Semantics Scholar
- IEEE Explore (need credentials)
- Multiprocessing to threading
- SoupStrainer for html parsing