Skip to content

index.md

Latest commit

 

History

History
88 lines (77 loc) · 4.8 KB

index.md

File metadata and controls

88 lines (77 loc) · 4.8 KB
Raw

PyAutoMemory — top-level index

The PyAuto organism's long-term knowledge: a multi-domain knowledge base of accumulated scientific and project knowledge, organised as several Karpathy-style LLM wikis over a shared canonical bibliography. Each sub-wiki uses the schema in wiki/AGENTS.md.

Sub-wikis

  • wiki/lensing/ — strong gravitational lensing. The primary sub-wiki. Lens equation, mass models, source reconstruction, degeneracies, time-delay cosmography, dark-matter substructure, surveys (SLACS, BELLS, H0liCOW, TDCOSMO, Euclid Q1, HFF), software (PyAutoLens, lenstronomy).
  • wiki/smbh/ — supermassive black holes. Binaries, recoil, scouring, mass functions, GW background (NANOGrav), seeds and growth.
  • wiki/cti/ — Charge Transfer Inefficiency. CCD trap physics, forward-model correction (arctic), trap pumping, Euclid VIS calibration, HST ACS, Gaia.
  • wiki/methods/ — statistical and computational methods. Bayesian inference, samplers, deep learning, probabilistic programming, NUFFT, simulations, scientific software.
  • wiki/galaxies/ — galaxy formation and evolution. Massive ellipticals (MASSIVE survey), bulge/disk decomposition, IFU spectroscopy, stellar halos, COSMOS, high-redshift, halo-galaxy connection, galaxy-scale DM geometry.

Reading queue

  • arxiv-inbox.md — the tier in front of the queue: papers the nightly arXiv digest suggested, one per line as <YYYY-MM-DD> — <title>[ — <ref>], where the date is the arXiv announcement date. The dashboard renders each with four one-tap actions — ➕ add to the queue, 📥 intake, 📑 cite, ✖️ dismiss — and un-acted lines lapse after seven days (INBOX_WINDOW_DAYS in scripts/inbox_actions.py). Nothing here has been chosen yet: a suggestion is not reading history, so a lapsed line is dropped rather than DONE-marked, and git history keeps it.
  • arxiv-interests.md — the same tier for everything that is not strong lensing: black holes, dark matter, galaxy formation, statistics. One line per paper as <YYYY-MM-DD> — [<Topic>] <title>[ — <ref>], where the topic names the reading-queue section ➕ files it into. It differs from the inbox in one way: it is day-batched. Each run appends that day's ten, the dashboard shows the oldest un-cleared batch only, and one 🧹 clear button drops that whole day and reveals the next. The seven-day window is the inbox's: a batch nobody clears lapses whole (sweep in scripts/interests_actions.py, on INBOX_WINDOW_DAYS), because at a measured 1.6 % clearance rate an unlapsing backlog just makes the oldest batch — the one on the board — the least relevant thing on it. Same five per-paper actions as the inbox. Format and transitions live in scripts/interests_actions.py.
  • reading-queue.md — the paper queue AND the reading history: ## section headers group papers by domain, one title per line, and a read paper is never deleted — its line gains a DONE <date> — prefix. When a queued paper is read, add its canonical entry to bibliography/, stub it in the matching sub-wiki sources/*.md, mark the line DONE, and run make validate — the PDF itself stays off-repo, never committed. Per-section waiting/read counts live on the dashboard.

Schema

The schema is defined once in wiki/AGENTS.md and inherited by all sub-wikis. Each sub-wiki has its own AGENTS.md that diverges only on the ## Scope section. Per-page frontmatter, naming conventions, [[wiki-links]], status flags (stub | drafted | reviewed), and source-page layout are uniform across sub-wikis.

Citation metadata

The sub-wikis explain which claims papers support. Canonical BibTeX metadata, downstream-project resolution rules, and validation live in bibliography/.

Provenance

The wikis were seeded in 2026-05 from a ~615-paper personal PDF library (the PDFs themselves were deleted from the repo the same day, backed up externally). Most legacy per-paper entries remain filename-inferred stubs: upgrade them to compact, claim-oriented drafted entries only after verifying the paper, and log the change in the relevant sub-wiki. The eight pages on which every entry is still such a stub were split out of sources/ into wiki/<domain>/seed/ — same shape, nothing citeable, promoted a paper at a time as each is verified. The live counts — pages, maturity, unresolved citation keys, reading queue — are on the dashboard.