Application Info

This application twXiv gives arXiv daily new submissions by tweets. We use python3 scripts. twXiv is not affiliated with arXiv.

Setup

• Install pandas, ratelimit, tweepy, twitter-text-parser, nameparser, and feedparser.

   % pip3 install pandas ratelimit tweepy twitter-text-parser nameparser feedparser 
  

• Let twXiv.py be executable.

   % chmod +x twXiv.py
  

• Put the following python scripts in the same directory.

  • twXiv.py
  • twXiv_post.py
  • twXiv_format.py
  • twXiv_daily_feed.py
  • arXiv_feed_parser.py
  • variables.py

• Configure switches.json, logfiles.json, and aliases.json in the tests directory for your settings.

  • accesses.json specifies twitter access keys and whether to put new submissions for each category.

  • logfiles.json indicates log file locations for tweet summaries and tweets. In the tests/logfiles director, you have the sample log files mathACb_tweet_summaries.csv and mathACb_tweets.csv.

  • aliases.json tells twXiv aliases of arXiv category names. For example, math.IT is an alias of cs.IT. If provided, twXiv replaces category names by their aliases.

• Configure variables.py for your settings.

  • variables.py assigns format parameters for twXiv tweets and access frequencies for arXiv and twitter.

Notes

• arXiv_feed_parser.py is a simple arXiv feed parser for twXiv. We use this via twXiv_daily_feed.py to regularly obtain data.

• Outputs of twXiv can differ from arXiv new submission web pages. First, this can be due to bugs in my scripts or connection errors. Second, items of arXiv rss feeds are not necessarily the same as those of arXiv new submission web pages (see https://mastoxiv.page/@vela/109829294232368163). Third, arXiv_feed_parser for an arXiv category C gives new submissions whose primary subjects are the category C. Then, twXiv for the category C counts and tweets a new paper whose principal subject matches the category C.

  • For example, let us assume that there is no new paper whose principal subject matches the category C, but there is a new paper P whose non-principal subject matches the category C. Then, the arXiv new submission web page of the category C lists the paper P as a new submission of the category C, not as a cross-list. However, twXiv keeps considering the paper P as a cross-list for the category C. Then, the output of twXiv for the category C differs from the arXiv new submission web page of the category C. One reason for this is to clarify what twXiv tweets across categories. Another reason is that twitter does not like duplicate or substantially similar tweets for bots.

• On the use of metadata of arXiv articles, there is the web page of Terms of Use for arXiv APIs. This says that "You are free to use descriptive metadata about arXiv e-prints under the terms of the Creative Commons Universal (CC0 1.0) Public Domain Declaration." and "Descriptive metadata includes information for discovery and identification purposes, and includes fields such as title, abstract, authors, identifiers, and classification terms."

Usage

% ./twXiv.py -h
usage: twXiv.py [-h] --switches_keys SWITCHES_KEYS
                [--logfiles LOGFILES] [--aliases ALIASES]
                [--captions CAPTIONS] [--mode {0,1}]

arXiv daily new submissions by tweets.

optional arguments:
  -h, --help            show this help message and exit
  --switches_keys SWITCHES_KEYS, -s SWITCHES_KEYS
                        output switches and api keys in
                        json
  --logfiles LOGFILES, -l LOGFILES
                        log file names in json
  --aliases ALIASES, -a ALIASES
                        aliases of arXiv categories in
                        json
  --captions CAPTIONS, -c CAPTIONS
                        captions of arXiv categories in
                        json
  --mode {0,1}, -m {0,1}
                        1 for twitter and 0 for stdout
                        only

Sample stdouts

• New submissions (if any) for math.AC and math.AG:

   % ./twXiv.py -s tests/switches.json -l tests/logfiles.json -a tests/aliases.json -c tests/captions.json -m 1
   **process started at xxxx-xx-xx xx:xx:xx (UTC)
   starting a thread of retrieval/new submissions for math.AC
   getting daily entries for math.AC
   waiting for a next thread of retrieval/new submissions
   new submissions for math.AC
  
   utc: xxxx-xx-xx xx:xx:xx 
   thread arXiv category: math.AC
   arXiv id: 
   url: https://x.com/user/status/
   post method: tweet
   post mode: 1
   url: https://x.com/user/status/xxxxxxxxxxxxxxxxxxxxxxxxx
   text: [xxxx-xx-xx Sun (UTC), 2 new articles found for mathAC Commutative Algebra]
  
   starting a thread of retrieval/new submissions for math.AG
   getting daily entries for math.AG
   joining threads of retrieval/new submissions
  
   utc: xxxx-xx-xx xx:xx:xx
   thread arXiv category: math.AC
   arXiv id: 2007.xxxxxx
   url: https://x.com/user/status/
   post method: tweet
   post mode: 1
   url: xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
  
   new submissions for math.AG
  
   utc: xxxx-xx-xx xx:xx:xx
   thread arXiv category: math.AG
   arXiv id: 
   url: https://x.com/user/status/
   post method: tweet
   post mode: 1
   url: https://x.com/user/status/xxxxxxxxxxxxxxxxxxxxxxxxxxxx
   text: [xxxx-xx-xx Sun (UTC), 4 new articles found for mathAG Algebraic Geometry]
   .....
  
   ** process ended at xxxx-xx-xx xx:xx:xx (UTC)
   ** elapsed time from the start: xx:xx:xx
  

Versions

• 0.0.1

  • 2020-08-08, initial release.

• 0.0.2

  • 2020-08-15, added the -c option.

• 0.1.0

  • 2021-01-31, text counting by twitter-text-parser, replacements by quotes and retweets, and links to research tools.

• 0.1.1

  • 2021-03-02, minor fixes.

• 0.1.2

  • 2021-06-03, a fix for semantic scholar api update.

• 0.1.3

  • 2021-10-26, a fix for tweepy v4.

• 0.1.4

  • 2022-08-20, minor fixes.

• 0.1.5

  • 2022-01-09, a fix for semanticscholar 0.3.2.

• 0.2.0

  • 2023-06-25, a fix for twitter API V2 endpoints and a daily rate limit.

• 0.3.0

  • 2024-02-04, updated arXiv_feed_parser for arXiv rss re-implemented.

• 0.3.1

  • 2024-02-07, a minor update.

• 0.4.1

  • 2025-01-12, deleted functions of retweets and replies for the X free api limits, updated for the X api "17 posts a day" limit by a tweet of two papers, added https://en.wikipedia.org/wiki/Mathematics as a generic math page for an automatic url preview of the two paper tweets, and added html paper links.

List of Bots

• https://x.com/mathACb: Commutative Algebra
• https://x.com/mathAGb: Algebraic Geometry
• https://x.com/mathAPb: Analysis of PDEs
• https://x.com/mathATb: Algebraic Topology
• https://x.com/mathCAbot: Classical Analysis and ODEs
• https://x.com/mathCObot: Combinatorics
• https://x.com/mathCTbot: Category Theory
• https://x.com/mathCVb: Complex Variables
• https://x.com/mathDGb: Differential Geometry
• https://x.com/mathDSb: Dynamical Systems
• https://x.com/mathFAbot: Functional Analysis
• https://x.com/mathGMb: General Mathematics
• https://x.com/mathGNb: General Topology
• https://x.com/mathGRbot: Group Theory
• https://x.com/mathGTb: Geometric Topology
• https://x.com/mathHOb: History and Overview
• https://x.com/mathITbot: Information Theory
• https://x.com/mathKTb: K-Theory and Homology
• https://x.com/mathLOb: Logic
• https://x.com/mathMGb: Metric Geometry
• https://x.com/mathMPb: Mathematical Physics
• https://x.com/mathNAb: Numerical Analysis
• https://x.com/mathNTb: Number Theory
• https://x.com/mathOAb: Operator Algebras
• https://x.com/mathOCb: Optimization and Control
• https://x.com/mathPRb: Probability
• https://x.com/mathQAb: Quantum Algebra
• https://x.com/mathRAb: Probability
• https://x.com/mathRTb: Representation Theory
• https://x.com/mathSGb: Symplectic Geometry
• https://x.com/mathSPb: Spectral Theory
• https://x.com/mathSTb: Statistics Theory

Author

So Okada, so.okada@gmail.com, https://so-okada.github.io/

Motivation

Since 2013-04, the author has been running the bots above for all arXiv math categories. Until 2020-08, the author used arxiv_speaker https://github.com/misho104-obsolete/arxiv_speaker. Sho Iwamoto wrote it by ruby for tweeting new submissions with titles, authors, and ids. However, it was discontinued. Therefore, the author has written twXiv by python3, adding functions such as abstracts by replies, cross-lists by retweets, and replacements by quotes and retweets. Since 2020-08, the bots above have used twXiv on gcp free tier. Since 2025-01, the author has added the function of two-new-submissions-in-one-tweet and abondoned the functions of abstracts, cross-lists, and replacements for the limitations of X api free tire.

For more background, see a paper on these bots, bots of Notionsandnotes, and the author's comments:

I am running these bots mainly just for convenience of sharing wisdom of arXiv papers for each arXiv category.

— So Okada (@vela) May 9, 2016

This would end my 32 math arXiv twitter bots such as https://t.co/w8cCIaUlBM, https://t.co/2j62N9Vk8a, ... They have shared wisdom of arXiv papers with comments, discussions, or jokes not just to academics but to the public. I have maintained them nearly 10 years from 2013-04. https://t.co/J0WUSSTlwZ

— So Okada (@vela) February 5, 2023

Contributing

Pull requests are welcome. For major changes, please open an issue first to discuss what you would like to change.

License

AGPLv3