Repository navigation
Conversation
- test_clean_duckdb_read: i 3 test ODS usano importorskip("odf") —
fallivano in env senza extra pandas (CI resta verde grazie a [dev,pandas])
- test_smoke_scout: i 4 test con URL reali (finanze.gov.it, consip.it)
dietro RUN_NETWORK_SMOoke=1 come RUN_S3_SMOKE — lo step smoke CI
(-m smoke) li includeva nonostante il docstring dicesse il contrario,
creando dipendenza da endpoint esterni; il test puro infer_years
resta smoke e gira sempre
L'audit markers richiede un marker per ogni test nei file toccati; i 4 test di rete avevano solo skipif senza marker. Stesso pattern di test_s3_parquet_smoke: @pytest.mark.smoke + skipif env var — restano collezionabili da -m smoke ma saltano senza RUN_NETWORK_SMOKE=1.
… ACB toolkit_dataset find restava indietro su query naturali: la semantica matchava solo la frase intera come substring continua. Ora tokenizza le query (primitive condivise lab_connectors.textmatch) e fa AND cross-part tra meta e colonne. - meta hit: tutti i token sparsi tra slug + name/description/source/ tags/category - colonna hit: tutti i token dentro nome/descrizione colonna - stopword italiane escluse dai token richiesti - morfologia leggera + prefissi (sanitaria~=sanita, import→importo) - split per parola: 'appalti' non matcha più dentro 'n_subappalti' - description tool MCP: find = slug/colonne live; discovery org-wide → dataciviclab-context find (DEC #11) Depende da lab-connectors#92 (textmatch) — mergiare prima quella.
Gabrymi93
force-pushed
the
feat/find-token-minimal
branch
from
October 14, 2026 15:36
7a740af to
c31da84
Compare
Member
Author
|
Chiusa: scope ridotto. Il matcher tokenizzato non serve se la discovery naturale va ad ACB find (DEC #11). Rimane solo il cambio description/routing, aperto in una PR dedicata minima. |
5 tasks done
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Sintesi
toolkit_dataset findnon trovava dataset su query naturali: la semantica matchava solo la frase intera come substring continua (appalti pubblici anac→ 0). Ora tokenizza le query e fa AND cross-part tra meta e colonne, usando le primitive condiviselab_connectors.textmatch. La description del tool dichiara il ruolo: find live = slug/colonne; discovery org-wide →dataciviclab-context find(DEC #11: ACB = dove cerco, toolkit = cosa c'è dentro).Contesto collegato
Depende da lab-connectors#92 — mergiare prima quella (porta
lab_connectors.textmatch). Segue la fix ACB agent-context-builder#90.Cosa cambia
Dettaglio
toolkit/domain/catalog.py—_semantic_query_hitslab_connectors.textmatch(stopword italiane escluse)sanitaria≈sanita,import→importo)appaltinon matcha più dentron_subappalti(falso positivo rimosso)toolkit/mcp/server.py— descriptiontoolkit_datasetdataciviclab-context findImpatto su artifact upstream
Verifica
pytest tests/passa (1322 passed, 1 skipped)ruff checkpassamypy toolkit/domain/catalog.pypassaTestSemanticFind(cross-part, stopword, morfologia, prefisso colonna, no-falso-positivo subappalti)@main). Se resta rosso su import ditextmatch, rifare merge/re-run dopo quel merge.Comportamento atteso vs prima (fixture semantica)
appalti garaanac_bandi_gara(tokens sparsi tra tags e name)bandi di garadidroppata)gare appaltianac_bandi_gara(morfologia)importimportodi dei delleChecklist PR
Note per chi revisiona
lab_connectors.textmatchè puramente funzionale — nessuna dipendenza di rete o stato.