Which posts make it into the papers feed:

Rules for tagging:

• a link is a "paper link" if it either matches the list of domains or the DOI number regular expression
• the top level post is treated as a "paper post" if it contains a a paper link, or has a self-reply with a paper link.

DOI regexp

10\.[0-9]{4,}(?:\.[0-9]+)*/[^\s"#?%]+

Domain list

arxiv.org

biorxiv.org

medrxiv.org

chemrxiv.org

doi.org

philpapers.org

openreview.net

aclanthology.org

ncbi.nlm.nih.gov

ssrn.com

jstor.org

nature.com

sciencedirect.com

springer.com

onlinelibrary.wiley.com

tandfonline.com

academic.oup.com

science.org

sciencemag.org

sagepub.com

cell.com

osf.io

mdpi.com

cambridge.org

frontiersin.org

pnas.org

zenodo.org

researchsquare.com

plos.org

plos.io

dl.acm.org

ieeexplore.ieee.org

semanticscholar.org

researchgate.net

elifesciences.org

preprints.org

thelancet.com

jamanetwork.com

hal.science

nber.org

bmj.com

eprint.iacr.org

ia.cr

alphaxiv.org

philsci-archive.pitt.edu

muse.jhu.edu

persee.fr

cairn.info

repec.org

psycnet.apa.org

diabetesjournals.org

journals.ametsoc.org

academia.edu

rlj.cs.umass.edu

Stats from ~90 days of Bluesky posts with links

total links scanned:              16,959,006
distinct hosts seen:                 772,123

URLS matched as paper
  by allowlist (host):               217,332  (1.282%)
  by DOI regex:                       82,307  (0.485%)
  by BOTH:                            73,846
  allowlist ONLY (no DOI):           143,486
  regex ONLY (host not listed):        8,461
  EITHER (= total paper urls):       225,793  (1.331%)

DISTINCT DOMAINS matched as paper
  by allowlist:                          372
  by DOI regex (>=1 DOI url):            511
  BOTH:                                  115
  allowlist ONLY:                        257
  regex ONLY:                            396
  EITHER (= total paper hosts):          768