Project:Analytics/PubPeer

API

https://dashboards.pubpeer.com/docs/api#/operations/partner

Relevant parameters:

• page: start with 1 then iterate based on whether there are more results
• per_page: set at maximum value 300
• sort:
• published_at: concerns when the document was published; I only care about comments

Resources

• Wikimedia Cloud Services
  • Toolforge: project "pubpeer"
  • Cloud VPS: project "wikicite", Trove DB instance

Process

• Initial seed:
  • Build pageset
    • Pull usages of identifiers from Wikimedia Cloud DB Replicas
    • Create database table (wikipedia):
      • id (incrementing key)
      • language_code
      • mw_page_id
      • mw_page_title (probably should have a process to refresh this before the full process runs)
  • API query: 2006-01-01..2025-12-31
    • To prevent overloading the script, I'll probably end up going one month at a time, 2006-01-01..2006-01-31
  • Iterate through as many pages as needed to get to the end
  • Build internal database (pubpeer_articles table):
    • id_pubpeer (key)
    • id_doi (update on conflict)
    • id_pubmed (update on conflict)
    • id_arxiv (update on conflict)
  • Build minimal citations database (citations table):
    • id_pubpeer (key to pubpeer_articles table)
    • id_wiki_page (key to wikipedia table)
    • cited_id_type (integer)
      • 0 = unknown/other
      • 1 = doi
      • 2 = pubmed
      • 3 = arxiv
      • Why not use enum? It's easier to add a new value Python-side than it is to carry out a database schema migration
    • cited_id_value (string)
    • time_last_updated_table (null when created)
    • time_last_talk_page_post (null when created)
    • time_most_recent_comment (on conflict, update if submitted > stored)
  • Post initial report to wiki as a table
  • Post initial notification that the report is posted
• Subsequent builds:
  • Get most recent time_most_recent_comment from database
  • API query: that date...present day
  • Iterate through as many result pages as needed (probably only one page)
  • Submit into database, which should transparently handle conflicts
  • Build new wiki table based on citations database table
    • Check database for
      • null time_last_updated_table
      • time_most_recent_comment > time_last_updated_table
  • Come up with alerts describing changes to table.
    • Retire old notifications to a subpage
• If/when talk page notifications are approved:
  • Check database for
    • null time_last_talk_page_post
    • time_most_recent_comment > time_last_talk_page_post
  • Queue up talk pages to notify
  • Check presence for message already on talk page.
    • No message comment: add post to talk page
    • Presence of message comment: skip over talk page