A python module to scrape arxiv.org for specific date range and categories
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Type Name Latest commit message Commit time
Failed to load latest commit information.
arxivscraper Fixing counter and string bugs May 24, 2018
.gitignore first commit Jun 7, 2017
setup.py update dl url Sep 12, 2017
todo.md update README Jun 8, 2017




An ArXiV scraper to retrieve records from given categories and date range.


Use pip (or pip3 for python3):

$ pip install arxivscraper

or download the source and use setup.py:

$ python setup.py install

or if you do not want to install the module, copy arxivscraper.py into your working directory.

To update the module using pip:

pip install arxivscraper --upgrade


Without filtering

You can directly use arxivscraper in your scripts. Let's import arxivscraper and create a scraper to fetch all preprints in condensed matter physics category from 27 May 2017 until 7 June 2017 (for other categories, see below):

import arxivscraper
scraper = arxivscraper.Scraper(category='physics:cond-mat', date_from='2017-05-27',date_until='2017-06-07')

Once we built an instance of the scraper, we can start the scraping:

output = scraper.scrape()

While scraper is running, it prints its status:

fetching up to  1000 records...
fetching up to  2000 records...
Got 503. Retrying after 30 seconds.
fetching up to  3000 records...
fetching is complete.

Finally you can save the output in your favorite format or readily convert it into a pandas dataframe:

import pandas as pd
cols = ('id', 'title', 'categories', 'abstract', 'doi', 'created', 'updated', 'authors')
df = pd.DataFrame(output,columns=cols)

With filtering

To have more control over the output, you could supply a dictionary to filter out the results. As an example, let's collect all preprints related to machine learning. This subcategory (stat.ML) is part of the statistics (stat) category. In addition, we want those preprints that word learning appears in their abstract.

import arxivscraper.arxivscraper as ax
scraper = ax.Scraper(category='stat',date_from='2017-08-01',date_until='2017-08-10',t=10, filters={'categories':['stat.ml'],'abstract':['learning']})
output = scraper.scrape()

In addition to categories and abstract, other available keys for filters are: author and title.


Here is a list of all categories available on ArXiv. For a complete list of subcategories, see categories.md.

Category Code
Computer Science cs
Economics econ
Electrical Engineering and Systems Science eess
Mathematics math
Physics physics
Astrophysics physics:astro-ph
Condensed Matter physics:cond-mat
General Relativity and Quantum Cosmology physics:gr-qc
High Energy Physics - Experiment physics:hep-ex
High Energy Physics - Lattice physics:hep-lat
High Energy Physics - Phenomenology physics:hep-ph
High Energy Physics - Theory physics:hep-th
Mathematical Physics physics:math-ph
Nonlinear Sciences physics:nlin
Nuclear Experiment physics:nucl-ex
Nuclear Theory physics:nucl-th
Physics (Other) physics:physics
Quantum Physics physics:quant-ph
Quantitative Biology q-bio
Quantitative Finance q-fin
Statistics stat


Ideas/bugs/comments? Please open an issue or submit a pull request on Github.

How to cite

If arxivscraper was useful in your work/research, please consider to cite it as :

Mahdi Sadjadi (2017). arxivscraper: Zenodo. http://doi.org/10.5281/zenodo.889853


  author       = {Mahdi Sadjadi},
  title        = {arxivscraper},
  year         = 2017,
  doi          = {10.5281/zenodo.889853},
  url          = {https://doi.org/10.5281/zenodo.889853}



This project is licensed under the MIT License - see the LICENSE file for details.