aeat.application.corpus_search package

On-host hybrid corpus-search grounding surface.

The console’s R3 grounding surface: a licence-clean, on-host retrieval stack over the bundled BOE/AEAT corpus, built from the already-shipped *.extracted triples. Three cooperating pieces live here:

• the FTS5 lexical index (_lexical_index) — standard-library SQLite plus a Spanish Snowball stemmed column, always importable;

• the structured citation lookup (_citation_lookup) — an exact legal_refs id resolver over the registry legal catalogue that returns verbatim authoritative text;

• the build-time embedding precompute (_embed_build) — the semantic half, gated behind the aeat-cli[search] extra, that ships its corpus vectors as plain data.

The lexical index and the citation lookup carry the degraded, no-download mode: they refuse nothing for want of a model. Only the embedding precompute and the runtime query embedder need the search extra, and they refuse with an install hint when it is absent.

Submodules

• aeat.application.corpus_search._citation_lookup module
  • CitationLookup
    • CitationLookup.citation_ids
    • CitationLookup.resolve()
    • CitationLookup.resolve_corpus_text()
  • bundled_citation_lookup()
• aeat.application.corpus_search._embed_build module
  • embed_corpus()
  • load_embeddings()
  • more_like_this()
• aeat.application.corpus_search._errors module
  • CorpusSearchError
    • CorpusSearchError.code
  • CorpusSearchInputError
    • CorpusSearchInputError.code
  • CorpusSearchDependencyError
    • CorpusSearchDependencyError.code
• aeat.application.corpus_search._lexical_index module
  • bundled_corpus_html_root()
  • iter_corpus_chunks()
  • build_lexical_index()
  • search_lexical()
• aeat.application.corpus_search._model_loader module
  • SEARCH_EXTRA_HINT
  • load_static_model()
  • model_dimensions()
• aeat.application.corpus_search._models module
  • RetrievalMode
    • RetrievalMode.CITATION
    • RetrievalMode.HYBRID
    • RetrievalMode.LEXICAL_ONLY
  • CorpusChunk
    • CorpusChunk.chunk_id
    • CorpusChunk.corpus_ref
    • CorpusChunk.source_path
    • CorpusChunk.doc_title
    • CorpusChunk.section
    • CorpusChunk.anchor
    • CorpusChunk.ordinal
    • CorpusChunk.text
  • CorpusDocument
    • CorpusDocument.corpus_ref
    • CorpusDocument.source_path
    • CorpusDocument.title
    • CorpusDocument.chunk_count
  • LexicalSearchHit
    • LexicalSearchHit.chunk_id
    • LexicalSearchHit.corpus_ref
    • LexicalSearchHit.doc_title
    • LexicalSearchHit.section
    • LexicalSearchHit.anchor
    • LexicalSearchHit.rank
    • LexicalSearchHit.score
    • LexicalSearchHit.text
  • CitationResolution
    • CitationResolution.citation_id
    • CitationResolution.document_id
    • CitationResolution.kind
    • CitationResolution.corpus_ref
    • CitationResolution.permalink
    • CitationResolution.article
    • CitationResolution.section
    • CitationResolution.anchor
    • CitationResolution.verbatim_text
  • SimilarChunk
    • SimilarChunk.chunk_id
    • SimilarChunk.rank
    • SimilarChunk.similarity
  • CorpusIndexBuildResult
    • CorpusIndexBuildResult.database_path
    • CorpusIndexBuildResult.document_count
    • CorpusIndexBuildResult.chunk_count
  • CorpusEmbeddingBuildResult
    • CorpusEmbeddingBuildResult.matrix_path
    • CorpusEmbeddingBuildResult.chunk_ids_path
    • CorpusEmbeddingBuildResult.chunk_count
    • CorpusEmbeddingBuildResult.dimensions
    • CorpusEmbeddingBuildResult.embedding_model_id
    • CorpusEmbeddingBuildResult.embedding_model_revision
  • RetrievalHit
    • RetrievalHit.chunk_id
    • RetrievalHit.corpus_ref
    • RetrievalHit.doc_title
    • RetrievalHit.text
    • RetrievalHit.score
    • RetrievalHit.rank
    • RetrievalHit.lexical_rank
    • RetrievalHit.semantic_rank
  • RetrievalResponse
    • RetrievalResponse.query
    • RetrievalResponse.mode
    • RetrievalResponse.hits
    • RetrievalResponse.citation
• aeat.application.corpus_search._query_embed module
  • search_extra_available()
  • search_model_cache_dir()
  • QueryEmbedder
    • QueryEmbedder.model_id
    • QueryEmbedder.revision
    • QueryEmbedder.cache_dir
    • QueryEmbedder.embed_query()
• aeat.application.corpus_search._retrieval module
  • RRF_K
  • PER_SIDE_CAP
  • hybrid_search()
• aeat.application.corpus_search._runtime module
  • corpus_search_dir()
  • corpus_index_path()
  • ensure_corpus_index()
  • load_corpus_embeddings()
  • ensure_corpus_embeddings()
  • search_corpus()
• aeat.application.corpus_search._terminology module
  • TerminologyConcept
    • TerminologyConcept.concept_id
    • TerminologyConcept.domain
    • TerminologyConcept.lifecycle
    • TerminologyConcept.locale
    • TerminologyConcept.preferred_label
    • TerminologyConcept.terms
    • TerminologyConcept.short_description
    • TerminologyConcept.definition
    • TerminologyConcept.scope_note
    • TerminologyConcept.legal_refs
  • TerminologyHit
    • TerminologyHit.concept_id
    • TerminologyHit.preferred_label
    • TerminologyHit.short_description
    • TerminologyHit.score
    • TerminologyHit.matched_on
  • load_terminology_concepts()
  • search_terminology()
  • lookup_terminology()