arxiv.py

Python wrapper for the arXiv API.

arXiv is a project by the Cornell University Library that provides open access to 1,000,000+ articles in Physics, Mathematics, Computer Science, Quantitative Biology, Quantitative Finance, and Statistics.

Usage

Install the package:

$ pip install arxiv   # Or `uv add arxiv` or similar.

In your Python code, include the line:

import arxiv

Examples

[!TIP] arxivql may simplify constructing complex query strings.

Fetching results

import arxiv

# Construct the default API client.
client = arxiv.Client()

# Search for the 10 most recent articles matching the keyword "quantum."
search = arxiv.Search(
  query = "quantum",
  max_results = 10,
  sort_by = arxiv.SortCriterion.SubmittedDate
)

results = client.results(search)

# `results` is a generator; you can iterate over its elements one by one...
for r in client.results(search):
  print(r.title)
# ...or exhaust it into a list. Careful: this is slow for large results sets.
all_results = list(results)
print([r.title for r in all_results])

# For advanced query syntax documentation, see the arXiv API User Manual:
# https://arxiv.org/help/api/user-manual#query_details
search = arxiv.Search(query = "au:del_maestro AND ti:checkerboard")
first_result = next(client.results(search))
print(first_result)

# Search for the paper with ID "1605.08386v1"
search_by_id = arxiv.Search(id_list=["1605.08386v1"])
# Reuse client to fetch the paper, then print its title.
first_result = next(client.results(search_by_id))
print(first_result.title)

Fetching results with a custom client

import arxiv

big_slow_client = arxiv.Client(
  page_size = 1000,
  delay_seconds = 10.0,
  num_retries = 5
)

# Prints 1000 titles before needing to make another request.
for result in big_slow_client.results(arxiv.Search(query="quantum")):
  print(result.title)

Logging

To inspect this package's network behavior and API logic, configure a DEBUG-level logger.

>>> import logging, arxiv
>>> logging.basicConfig(level=logging.DEBUG)
>>> client = arxiv.Client()
>>> paper = next(client.results(arxiv.Search(id_list=["1605.08386v1"])))
INFO:arxiv.arxiv:Requesting 100 results at offset 0
INFO:arxiv.arxiv:Requesting page (first: False, try: 0): https://export.arxiv.org/api/query?search_query=&id_list=1605.08386v1&sortBy=relevance&sortOrder=descending&start=0&max_results=100
DEBUG:urllib3.connectionpool:Starting new HTTPS connection (1): export.arxiv.org:443
DEBUG:urllib3.connectionpool:https://export.arxiv.org:443 "GET /api/query?search_query=&id_list=1605.08386v1&sortBy=relevance&sortOrder=descending&start=0&max_results=100&user-agent=arxiv.py%2F1.4.8 HTTP/1.1" 200 979

Types

Client

A Client specifies a reusable strategy for fetching results from arXiv's API. For most use cases the default client should suffice.

Clients configurations specify pagination and retry logic. Reusing a client allows successive API calls to use the same connection pool and ensures they abide by the rate limit you set.

Search

A Search specifies a search of arXiv's database. Use Client.results to get a generator yielding Results.

Result

The Result objects yielded by Client.results include metadata about each paper and helper methods for downloading their content.

The meaning of the underlying raw data is documented in the arXiv API User Manual: Details of Atom Results Returned.

Result also exposes helper methods for downloading papers: Result.download_pdf and Result.download_source.

Development

This project uses UV for development, while maintaining compatibility with traditional pip installation for end users.

Development Setup

1. Install UV (if you haven't already):

   curl -LsSf https://astral.sh/uv/install.sh | sh
   

2. Clone and setup:

   git clone https://github.com/lukasschwab/arxiv.py
   cd arxiv.py
   make dev-setup