ConvexPi Content

LLM-generated research content for the ConvexPi quantitative finance education platform. Generated by the convexpi/lab literature pipeline and consumed by convexpi/platform at build time.

Structure

wikis/
  {paper_id}.md       — one wiki per paper, keyed by Supabase papers.id
  ...
surveys/
  {topic}.md          — topic-level survey essays
  ...
index.json            — lightweight metadata for all wikis (title, authors, year,
                        topics, updated_at); loaded by the platform to drive
                        /papers listing and anomaly page cross-links

How content gets here

1. pipeline/ingest_finance_papers.py fetches arXiv q-fin papers → Supabase papers table
2. pipeline/generate_factor_wiki.py generates wiki markdown → local CONVEXPI_DATA_DIR/wiki/
3. pipeline/publish_content.py commits new/changed files to this repo and pushes

Wikis are also written back to papers.wiki_markdown in Supabase for DB-driven search.

Consuming in Next.js

The platform fetches wikis via the GitHub raw content URL at build time, or reads them from a local clone when running anomalies.yml.

Individual paper pages (/papers/[id]) load wikis/{paper_id}.md from this repo. The listing page (/papers) reads index.json.

Not financial advice

All content is generated from academic papers for educational purposes only.