twXiv icon indicating copy to clipboard operation
twXiv copied to clipboard

Published 20 hours ago • verified publisher icon so-okada
→

Metadata

This app twXiv gives arXiv daily new submissions by tweets, abstracts by replies, cross-lists by retweets, and replacements by quotes and retweets. We use python3 scripts. twXiv is not affiliated with...

Application Info

This application twXiv gives arXiv daily new submissions by tweets, abstracts by replies, cross-lists by retweets, and replacements by quotes and retweets. We use python3 scripts. twXiv is not affiliated with arXiv.

Setup

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

     pip3 install pandas ratelimit semanticscholar 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
    • twXiv_semantic.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 use new submissions/abstracts/cross-lists/replacements by twXiv.

    • logfiles.json indicates log file locations for tweet summaries, tweets, retweets, unretweets, and replies. You can check their formats by mathACb_tweet_summaries.csv, mathACb_tweets.csv, mathACb_retweets.csv, mathACb_unretweets.csv, mathACb_replies.csv, and mathACb_quotes.csv in the tests/logfiles director. twXiv needs a tweet log file for cross-lists and replacements. Other log files are useful to avoid duplication errors of tweets and retweets.

    • aliases.json tells twXiv aliases of arXiv category names. For example, math.IT is an alias of cs.IT. Without this file, twXiv of rss feeds returns no new submissions, when you take the category name math.IT. If provided, twXiv replaces category names by their aliases for new submissions, cross-lists, and replacements.

  • 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, retweets, and quotes across categories. Another reason is that twitter does not like duplicate or substantially similar tweets from bots. So, twXiv maintains a single tweet for the title, authors, and abs/pdf identifiers of a new paper across bots in access keys. Furthermore, if configured, twXiv tries to retweet and quote for cross-lists and replacements from bots in access keys.

  • On the use of metadata of arXiv articles, there is the web page of Terms of Use for arXiv APIs. As of the revision 0.8.0, 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, abstracts by
replies, cross-lists by retweets, and replacements by
quotes and retweets.

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 for math.AC and math.AG with no log files:

     ./twXiv.py -s tests/switches.json -m 1
*process started at xxxx-xx-xx xx:xx:xx (UTC)
tarting a thread of retrieval/new submissions/abstracts for math.AC
etting daily entries for math.AC
aiting for a next thread of retrieval/new submissions/abstracts
ew submissions for math.AC
o log files
o log files

tc: xxxx-xx-xx xx:xx:xx 
hread arXiv category: math.AC
rXiv id: 
rl: https://twitter.com/user/status/
ost method: tweet
ost mode: 1
rl: https://twitter.com/user/status/xxxxxxxxxxxxxxxxxxxxxxxx
ext: [xxxx-xx-xx Sun (UTC), 2 new articles found for mathAC]

tarting a thread of retrieval/new submissions/abstracts for math.AG
etting daily entries for math.AG
oining threads of retrieval/new submissions/abstracts

tc: xxxx-xx-xx xx:xx:xx
hread arXiv category: math.AC
rXiv id: xxxx.xxxxx
rl: https://twitter.com/user/status/
ost method: tweet
ost mode: 1
rl: https://twitter.com/user/status/xxxxxxxxxxxxxxxxxxxxxxxx
ext: xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx

...

*process ended at xxxx-xx-xx xx:xx:xx (UTC)
*elapsed time from the start: xx:xx:xx

  • New submissions, abstracts, cross-lists and replacements (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)
tarting a thread of retrieval/new submissions/abstracts for math.AC
etting daily entries for math.AC
aiting for a next thread of retrieval/new submissions/abstracts
ew submissions for math.AC

tc: xxxx-xx-xx xx:xx:xx 
hread arXiv category: math.AC
rXiv id: 
rl: https://twitter.com/user/status/
ost method: tweet
ost mode: 1
rl: https://twitter.com/user/status/xxxxxxxxxxxxxxxxxxxxxxxxx
ext: [xxxx-xx-xx Sun (UTC), 2 new articles found for mathAC Commutative Algebra]

tarting a thread of retrieval/new submissions/abstracts for math.AG
etting daily entries for math.AG
oining threads of retrieval/new submissions/abstracts

tc: xxxx-xx-xx xx:xx:xx
hread arXiv category: math.AC
rXiv id: 2007.xxxxxx
rl: https://twitter.com/user/status/
ost method: tweet
ost mode: 1
rl: xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx

ew submissions for math.AG

tc: xxxx-xx-xx xx:xx:xx
hread arXiv category: math.AG
rXiv id: 
rl: https://twitter.com/user/status/
ost method: tweet
ost mode: 1
rl: https://twitter.com/user/status/xxxxxxxxxxxxxxxxxxxxxxxxxxxx
ext: [xxxx-xx-xx Sun (UTC), 4 new articles found for mathAG Algebraic Geometry]

tc: xxxx-xx-xx xx:xx:xx
hread arXiv category: math.AC
rXiv id: 2007.xxxxxx
rl: https://twitter.com/user/status/xxxxxxxxxxxxxxxxxxxx
ost method: reply
ost mode: 1
rl: https://twitter.com/user/status/xxxxxxxxxxxxxxxxxxxx
ext:  xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx [1/4 of https://arxiv.org/abs/2007.xxxxxv1]

....

*crosslist process started at xxxx-xx-xx xx:xx:xx (UTC) 
*elapsed time from the start: xx:xx:xx
tart a crosslist thread of math.AC
aiting for a next crosslist thread
tart a crosslist thread of math.AG
oining crosslist threads

*replacement process started at xxxx-xx-xx xx:xx:xx (UTC)
*elapsed time from the start: xx:xx:xx
*elapsed time from the cross-list start: xx:xx:xx
uote-replacement starts
tart a quote-replacement thread of math.AC
aiting for a next quote-replacement thread
tart a quote-replacement thread of math.AG

....

tc: xxxx-xx-xx xx:xx:xx
hread arXiv category: math.AG
rXiv id: 2010.xxxx
rl: https://twitter.com/user/status/xxxxxxxxxxxxxxxxxxxx
ost method: quote
ost mode: 0
rl: https://twitter.com/user/status/
ext: This https://arxiv.org/abs/xxxx.xxxx has been replaced.....

....

etweet-replacement starts
tart a retweet-replacement thread of math.AC
aiting for a next quote-replacement thread
tart a retweet-replacement thread of math.AG

tc: xxxx-xx-xx xx:xx:xx 
hread arXiv category: math.AG
rXiv id: 1900.xxxxx
rl: https://twitter.com/user/status/xxxxxxxxxxxxxxxxxxxx
ost method: unretweet
ost mode: 1
rl: https://twitter.com/user/status/xxxxxxxxxxxxxxxxxxxx
ext: 

tc: xxxx-xx-xx xx:xx:xx 
hread arXiv category: math.AG
rXiv id: 1900.xxxxx
rl: https://twitter.com/user/status/xxxxxxxxxxxxxxxxxxxx
ost method: retweet
ost mode: 1
rl: https://twitter.com/user/status/xxxxxxxxxxxxxxxxxxxx
ext: 

....

*process ended at xxxx-xx-xx xx:xx:xx (UTC)
*elapsed time from the start: xx:xx:xx
*elapsed time from the cross-list start: xx:xx:xx
*elapsed time from the replacement start: xx:xx:xx

  • Without the option -c tests/captions.jsonabove, you get

    ext: [xxxx-xx-xx Sun (UTC), 4 new articles found for mathAG]

    instead of

    ext: [xxxx-xx-xx Sun (UTC), 4 new articles found for mathAG Algebraic Geometry]

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.

List of Bots

Author

So Okada, [email protected], 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.

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

Contributing

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

License

AGPLv3

← Metadata

15
Stars
3
Forks
Watchers

Owner

verified publisher icon so-okada

Metadata

This app twXiv gives arXiv daily new submissions by tweets, abstracts by replies, cross-lists by retweets, and replacements by quotes and retweets. We use python3 scripts. twXiv is not affiliated with...

Back