manasource

Corpus coverage

Imported records

This build ships a compiled copy of the imported records corpus, fixed at the source pin it was built from. Every number, source and notice below is read from that packaged copy, so this page describes the records this deploy actually serves rather than whatever an upstream database holds today.

What this build carries

Aggregates only. This page lists no records and links to none; the records themselves are reached through Library search and the public records API.

Imported records
33,317
Cited source rows
33,317
Links to curated resources
29
Upstream sources
4
Records by the corpus record type they were imported as
Record type Records
Compound 27,232
Supplement ingredient 3,160
Supplement product 2,562
Food 363

What these numbers are not

Breadth is not evidence. How many records this corpus carries, how many source rows they cite, and how precisely a source can be named say nothing about whether anything works. Reported effect size and evidence strength are separate values, and only curated Library resources carry them.

Imported records and their attributed facts receive no score, no grade, no recommendation, no evidence association and no promotion into evidence. A fact an importer carries — a dose range a licence holder stated on a product, for example — is reproduced with its source's own attribution and stays reference data. It is not a recommendation, an intake target or an upper limit.

A link to a curated resource means an imported record names the same thing that resource is about. It is an identity match, ranked by match quality so results can be ordered, and match quality is not an evidence score.

Where these records came from

One entry per source namespace in the packaged corpus, named exactly as the importer that wrote those rows attributes them. Each notice is reproduced as that importer records it, including what the import changed and what it does not cover.

Which corpus this is

The hash names the compiled corpus this page describes. The public records API derives its cache tags from the same string, so a reader can tell whether the catalog they paged through is the one these counts describe.

Corpus hash
c621a2d99639f4cd02cf6dae53607ba5c05799014d39876fa7f207ef20139506
Corpus version
1
Artifact format version
3

The same values are served as JSON at /api/library/records/meta.