Awesome

Rebiber: A tool for normalizing bibtex with official info.

We often cite papers using their arXiv versions without noting that they are already PUBLISHED in some conferences. These unofficial bib entries might violate rules about submissions or camera-ready versions for some conferences. We introduce Rebiber, a simple tool in Python to fix them automatically. It is based on the official conference information from the DBLP or the ACL anthology (for NLP conferences)! You can check the list of supported conferences here. Apart from handling outdated arXiv citations, Rebiber also normalizes citations in a unified way (DBLP-style), supporting abbreviation and value selection.

Demo on Huggingface Space https://huggingface.co/spaces/yuchenlin/Rebiber (recommended)

Colab notebook: here

Changelog

• 2024.7 Version 1.2.0. added automatic script to download bib files for recent conferences from dblp.

• 2023.06.01 New demo ready to use on Huggingface's Space via Gradio. Also, a few conferences are added.

• 2021.09.06 We fixed a few minor bugs and added features such as sorting and urls to arXiv (if the paper is not in any conferences; thanks to @nicola-decao). We also updated the ACL anthology bib/json to the latest version as well as other conferences.

• 2021.05.30 We build a beta version of our web app for Rebiber; add new conferences to our dataset; fix a few minor bugs. (It is not working anymore. Please use the new huggingface space demo.)

• 2021.02.08 We now support multiple useful features: 1) turning off some certain values, e.g., "-r url,pages,address" for removing the values from the output, 2) using abbr. to shorten the booktitle values, e.g., Proceedings of the .* Annual Meeting of the Association for Computational Linguistics --> Proc. of ACL. More examples are here.

• 2021.01.30 We build a colab notebook as a simple web demo. link

Installation

# pip install rebiber -U # for the stable version
pip install -e git+https://github.com/yuchenlin/rebiber.git#egg=rebiber -U
# rebiber --update  # (optional) update the bib data and the abbr. info  (using wget)

OR

git clone https://github.com/yuchenlin/rebiber.git
cd rebiber/
pip install -e .

If you would like to use the latest github version with more bug fixes, please use the second installation method.

Usage（v1.1.3 and v1.2.0）

Normalize your bibtex file with the official conference information:

rebiber -i /path/to/input.bib -o /path/to/output.bib

You can find a pair of example input and output files in rebiber/example_input.bib and rebiber/example_output.bib.

Example Input and Output

An example input entry with the arXiv information (from Google Scholar or somewhere):

@article{lin2020birds,
	title={Birds have four legs?! NumerSense: Probing Numerical Commonsense Knowledge of Pre-trained Language Models},
	author={Lin, Bill Yuchen and Lee, Seyeon and Khanna, Rahul and Ren, Xiang},
	journal={arXiv preprint arXiv:2005.00683},
	year={2020}
}

An example normalized output entry with the official information:

@inproceedings{lin2020birds,
    title = "{B}irds have four legs?! {N}umer{S}ense: {P}robing {N}umerical {C}ommonsense {K}nowledge of {P}re-{T}rained {L}anguage {M}odels",
    author = "Lin, Bill Yuchen  and
      Lee, Seyeon  and
      Khanna, Rahul  and
      Ren, Xiang",
    booktitle = "Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP)",
    month = nov,
    year = "2020",
    address = "Online",
    publisher = "Association for Computational Linguistics",
    url = "https://www.aclweb.org/anthology/2020.emnlp-main.557",
    doi = "10.18653/v1/2020.emnlp-main.557",
    pages = "6862--6868",
}

Supported Conferences

The bib_list.txt contains a list of converted json files of the official bib data. In this repo, we now support the full ACL anthology, i.e., all papers that are published at *CL conferences (ACL, EMNLP, NAACL, etc.) as well as workshops. Also, we support any conference proceedings that can be downloaded from DBLP, for example, ICLR2020.

Note that to DBLP only allows you to download in batches of 1000 using &h=1000&f=0, where f=0|1000|2000 specifies the starting index. So we have to manually download the bib files of each conference and concatenate them together. add_conf.sh takes care of that, too.

The following conferences are supported and their bib/json files are in our data folder. You can turn each item on/off in bib_list.txt. Please feel free to create PR for adding new conferences following this!

Thanks for Anton Tsitsulin's great work on collecting such a complete set bib files!

Adding a new conference

• Step 1, obtain the bib files for the conferences

You may download the bib files of recent conferences from dblp by running the code:

python download_dblp.py

Notice that ECCV and ECML does not work for the automatic download.

Alternatively, you can manually add any conferences from DBLP by downloading their bib files to our raw_data folder, and run a prepared script add_conf.sh. Take ICLR2020 and ICLR2019 as an example:

• Go to DBLP and Download the bib files, and put them here as raw_data/iclr2020.bib and raw_data/iclr2019.bib (name should be in the format as {conf_name}{year}.bib)

• Step 2: Run script to add the conferences

bash add_conf.sh iclr 2019 2020

Particularly, to update *CL conference, we can

python bib2json.py -i raw_data/anthology.bib -o data/acl.json

Star History

Contact

Please email yuchen.lin@usc.edu or create Github issues here if you have any questions or suggestions.