refactor(evidence): migrate acquisition and preprocessing (#30)

harness/tht/evidence/__init__.py

@@ -7,33 +7,69 @@ and exceptions; they do not define a cross-domain service protocol.
 
 from __future__ import annotations
 
-from collections.abc import Iterable
 from pathlib import Path
 from typing import TYPE_CHECKING
 
-from tht.ports.evidence import (
+from tht.evidence.acquisition import acquire, discover
+from tht.evidence.contracts import (
     AcquiredDocument,
     EvidenceSource,
     EvidenceSourceError,
     EvidenceSourceErrorCategory,
     SourceObject,
+    canonical_provenance_uri,
+    normalize_aware_datetime,
+    validate_namespaced_value,
+    validate_safe_metadata,
 )
 
 if TYPE_CHECKING:
+    from tht.config import EvidenceSourcesConfig
+    from tht.corpus.chunk import ChunkPolicy
+    from tht.corpus.pipeline import CorpusPipeline
     from tht.corpus.store import CorpusStore
     from tht.decisions import DecisionRecord
+    from tht.evidence.preprocessing import EvidenceEmbedder
+    from tht.ports.vector import VectorStore
     from tht.search.evidence import ActiveEvidenceSearcher
     from tht.session.models import SchemaLinking
 
 
-def discover(source: EvidenceSource) -> Iterable[SourceObject]:
-    """Discover source objects without changing source-defined laziness or ordering."""
-    return source.discover()
+def build_sources(evidence: "EvidenceSourcesConfig | None") -> list[EvidenceSource]:
+    """Build configured Evidence source adapters in the existing deterministic order."""
+    from tht.evidence.sources import build_sources as build_configured_sources
+
+    return build_configured_sources(evidence)
 
 
-def acquire(source: EvidenceSource, item: SourceObject) -> AcquiredDocument:
-    """Acquire one discovered object, preserving the source's classified failures."""
-    return source.acquire(item)
+def build_preprocessing_pipeline(
+    *,
+    store: "CorpusStore",
+    sources: list[EvidenceSource],
+    embedder: "EvidenceEmbedder",
+    vector_store: "VectorStore",
+    embedding_model: str,
+    embedding_dimensions: int,
+    chunk_policy: "ChunkPolicy",
+    pipeline_version: str,
+    retain_published_generations: int = 3,
+    workspace_id: str | None = None,
+) -> "CorpusPipeline":
+    """Construct the Evidence preprocessing use case from core-owned infrastructure."""
+    from tht.evidence.preprocessing import build_preprocessing_pipeline as build_pipeline
+
+    return build_pipeline(
+        store=store,
+        sources=sources,
+        embedder=embedder,
+        vector_store=vector_store,
+        embedding_model=embedding_model,
+        embedding_dimensions=embedding_dimensions,
+        chunk_policy=chunk_policy,
+        pipeline_version=pipeline_version,
+        retain_published_generations=retain_published_generations,
+        workspace_id=workspace_id,
+    )
 
 
 def active_searcher(
@@ -135,9 +171,15 @@ __all__ = [
     "SourceObject",
     "acquire",
     "active_searcher",
+    "build_preprocessing_pipeline",
     "build_retrieval_entries",
+    "build_sources",
+    "canonical_provenance_uri",
     "discover",
+    "normalize_aware_datetime",
     "project_session",
     "resolve_citation",
     "validate_corpus_workspace",
+    "validate_namespaced_value",
+    "validate_safe_metadata",
 ]