Elamur Groupantsmova-topic-corpus-missing
The groupant beside tools/topic-corpus/src/missing.rs in the Mova repository: what the code computes, what it never does, how to go back from its outputs, its invariants, cost and checks.
Facts
Implementation material. A Mova code model describes this code at one source revision: partial models, finite checks and their stated scope. It is not a reviewed general exposition, and its labels are shown as written, never raised.
Components
| component | label | where |
|---|---|---|
| approximant | partial | Forward approximant |
| ast | partial | syntax.json |
| composition | partial | Composition |
| cost | partial | Cost |
| descriptant | present | Descriptant |
| invariant | partial | Invariant |
| tests | finite evidence | tests.md |
Obernants
The typed relations around the forward model: the behaviour complement (what the code never does), the reverse reconstruction (every input compatible with an output) and specialised views.
| id | type | label | where |
|---|---|---|---|
behavior_complement | behavior_complement | partial source derived model | behavior_complement |
reverse_reconstruction | reverse_reconstruction | partial source derived model | reverse_reconstruction |
Model text
Descriptant
Stage 2: the cited works the library lacks (ids, else titles), looked up in the local OpenAlex snapshot by the library's papers openalex lookup, ranked by citations, routed to a legal open copy (ACL Anthology PDF, arxiv.org/pdf, OpenAlex best-OA PDF outside shadow libraries) and written as a new request run directory for the library's stream driver (manifest chunk, list-only works, ranking, run.sh with ROBOTS=1 HOST_SKIP=1 PDFDROP=1). Nothing is started. Source: tools/topic-corpus/src/missing.rs.
Forward approximant
external(paper bibliography + bibliographies of library references and similar works) -> keys acl/arxiv/doi/title -> one lookup call -> title keys re-keyed by the record's ids when the year agrees -> route -> sort by (cited_by, neighbourhood count, by paper, key) -> write_request (disk guard >= 300 GB, new dir only, at most max_requests rows).
Invariant
No shadow-library host is ever queued; a request directory is never rewritten; ACL licence by year (CC BY 4.0 from 2016, CC BY-NC-SA 3.0 before).
Cost
One snapshot scan (~10-20 s, 12 threads in the papers binary) plus linear work in candidates.
Composition
Parent: topic-corpus.
behavior_complement
Does not download, does not start or resume streams, does not touch stopped streams, does not use the OpenAlex API. Evidence status: source-derived model with finite checks, not a proof.
reverse_reconstruction
request.json keeps every candidate with its route and reason; the original bibliography lines are in the library. No unique inverse is claimed.
tests.md
missing::tests (ranking, routes incl. a title-keyed arXiv record, request dir, never rewritten, disk guard refusal; a shadow-library copy never queued; civil date). Negative control 2026-11-04: shadow() always false turns two tests red. Live 2026-11-04 bryant-2017: 201 candidates, 23 queued, 174 list-only, two identical dry runs.
cargo test --manifest-path tools/topic-corpus/Cargo.toml --offline with WORLD_EN_MODEL=$MOVA_DATA/download/0/0.4/en.bin (18 unit tests; fixture library, fake papers binary, fake terminal, local Sea Spider store; two identical runs on 2026-11-04). Without the model the model-dependent tests report a skip.
Negative controls 2026-11-04 (each red, then restored): public admission without the origin condition; assert_public without the Sea Spider/draft/note refusals; seeding without the learner lock; shadow-library filter off; public citing that ignores refusals; cache eviction ignoring the hot pin.
These are finite fixtures and live samples; not a proof.
This page in Markdown → https://md.groupants.elamur.ai/free/en/mova/tools/topic-corpus/src/missing.rs.groupant/