w0lph/dog-geroscience-mcp-data

dog-geroscience-mcp database

dog_geroscience.sqlite is the prebuilt database that the dog-geroscience-mcp server downloads on first run (uvx dog-geroscience-mcp, or dog-geroscience-mcp fetch-data). It is not meant to be browsed here; it exists so that installing the server needs no build step.

Table	Source	Terms
anage, drugage, genage_human, genage_models	HAGR	CC BY 3.0
dap_codebook, dap_codebook_fts	Dog Aging Project codebooks	as published on GitHub
corpus_records, corpus_fts	Canine Aging Corpus (Europe PMC)	Europe PMC terms
corpus_fulltext	the corpus's CC BY / CC0 full text only	per-article licence in each row
foi_dog, foi_records, foi_structured	FOI summaries, dog products	public domain; extraction CC BY 4.0
meta	build summary: every source URL, fetch time, counts
ensembl_cache	empty; filled at run time by the dog_ortholog tool

build_summary.json is the same build record as meta.build.

Rebuilding it

cd mcp && uv sync
uv run dog-geroscience-mcp build --fulltext-licences "cc by,cc0" --out dog_geroscience.sqlite

build downloads the HAGR files and the DAP codebooks, indexes ../corpus/data and ../foi/data, and writes a single file. Without --fulltext-licences it stores every full text on disk (for local use); the published file stores only the redistributable subset.

Licence

Mixed, per table, as listed above and in LICENSE. The server code is MIT.