refactor(evidence): remove legacy Python layout (#31)

harness/tht/adapters/evidence/__init__.py

@@ -1,7 +0,0 @@
-"""Evidence source adapter implementations."""
-
-from tht.adapters.evidence.filesystem import FilesystemEvidenceSource
-from tht.adapters.evidence.http import HttpManifestEvidenceSource
-from tht.adapters.evidence.s3 import S3EvidenceSource
-
-__all__ = ["FilesystemEvidenceSource", "HttpManifestEvidenceSource", "S3EvidenceSource"]

harness/tht/adapters/factory.py

@@ -42,11 +42,4 @@ def build_vector_store(cfg: Config, *, require_write: bool = False) -> VectorSto
             raise ConfigError(f"Adapter vector non supportato: {other}")
 
 
-def build_evidence_sources(cfg: Config):
-    """Compatibility shim for callers not yet migrated to ``tht.evidence``."""
-    from tht.evidence import build_sources
-
-    return build_sources(cfg.evidence)
-
-
-__all__ = ["build_dwh", "build_evidence_sources", "build_vector_store"]
+__all__ = ["build_dwh", "build_vector_store"]

harness/tht/cli/preprocess_cmd.py

@@ -96,8 +96,8 @@ def run_from_config(config: Path, *, dry_run: bool = False, resume: str | None =
     from tht.adapters.factory import build_vector_store
     from tht.cli.schema_cmd import _load_config_or_exit
     from tht.cli.vector_cmd import make_embedder
-    from tht.corpus.chunk import ChunkPolicy
-    from tht.corpus.store import CorpusStore
+    from tht.evidence.corpus.chunk import ChunkPolicy
+    from tht.evidence.corpus.store import CorpusStore
     from tht.evidence import build_preprocessing_pipeline, build_sources
 
     cfg = _load_config_or_exit(config)
@@ -130,8 +130,8 @@ def gc_from_config(config: Path, *, dry_run: bool = False):
     from tht.adapters.factory import build_vector_store
     from tht.cli.schema_cmd import _load_config_or_exit
     from tht.cli.vector_cmd import make_embedder
-    from tht.corpus.chunk import ChunkPolicy
-    from tht.corpus.store import CorpusStore
+    from tht.evidence.corpus.chunk import ChunkPolicy
+    from tht.evidence.corpus.store import CorpusStore
     from tht.evidence import build_preprocessing_pipeline, build_sources
 
     cfg = _load_config_or_exit(config)

harness/tht/corpus/__init__.py

@@ -1 +0,0 @@
-"""Canonical, transport-independent Evidence corpus."""

harness/tht/evidence/__init__.py

@@ -1,14 +1,4 @@
-"""Cohesive public entrypoint for Evidence domain capabilities.
-
-The implementation is introduced beside the legacy module layout so production callers can
-migrate one path at a time. These functions deliberately preserve the existing objects, ordering,
-and exceptions; they do not define a cross-domain service protocol.
-"""
-
-from __future__ import annotations
-
-from pathlib import Path
-from typing import TYPE_CHECKING
+"""Cohesive public entrypoint for Evidence domain capabilities."""
 
 from tht.evidence.acquisition import acquire, discover
 from tht.evidence.contracts import (
@@ -22,145 +12,17 @@ from tht.evidence.contracts import (
     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 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 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(
-    cfg,
-    delegate,
-    *,
-    workspace_id: str | None = None,
-) -> ActiveEvidenceSearcher:
-    """Bind vector search to the atomically ACTIVE Evidence corpus generation."""
-    from tht.search.evidence import active_searcher as legacy_active_searcher
-
-    return legacy_active_searcher(cfg, delegate, workspace_id=workspace_id)
-
-
-def validate_corpus_workspace(cfg, workspace_id: str) -> None:
-    """Validate persisted Evidence corpus ownership before runtime retrieval setup."""
-    from tht.search.evidence import validate_corpus_workspace as legacy_validate
-
-    legacy_validate(cfg, workspace_id)
-
-
-def build_retrieval_entries(results, *, excerpt_chars: int) -> list[dict]:
-    """Project ordered Evidence search hits into the existing retrieval-pack shape."""
-    return [
-        {
-            "title": result.label,
-            "status": result.status,
-            "excerpt": result.content[:excerpt_chars],
-        }
-        for result in results
-    ]
-
-
-def resolve_citation(
-    store: "CorpusStore",
-    evidence_id: str,
-    *,
-    materialized_root: Path | None = None,
-) -> str:
-    """Resolve an Evidence identifier to its immutable ACTIVE materialization."""
-    from tht.search.evidence import resolve_evidence_file
-
-    return resolve_evidence_file(
-        store,
-        evidence_id,
-        materialized_root=materialized_root,
-    )
-
-
-def _resolve_session_citation(evidence_root: Path, evidence_id: str) -> str:
-    # New deployments resolve only immutable materialized files from ACTIVE. Keep the
-    # curated-tree fallback for sessions created before a canonical corpus exists.
-    corpus_root = evidence_root.parent.parent / "corpus"
-    if corpus_root.exists():
-        from tht.corpus.store import CorpusStore
-
-        return resolve_citation(
-            CorpusStore(corpus_root),
-            evidence_id,
-            materialized_root=evidence_root.parent / ".materialized-evidence",
-        )
-    for match in evidence_root.rglob(f"{evidence_id}.md"):
-        return str(match)
-    return ""
-
-
-def project_session(
-    decisions: list["DecisionRecord"],
-    linking: "SchemaLinking",
-    evidence_root: Path,
-) -> list[dict]:
-    """Project cited and reviewed Evidence into the existing session artifact shape."""
-    entries: dict[str, dict] = {}
-    for candidate in linking.candidates:
-        for evidence_id in candidate.evidence:
-            entries.setdefault(evidence_id, {
-                "id": evidence_id,
-                "file": _resolve_session_citation(evidence_root, evidence_id),
-                "esito": "usata",
-                "decision_seq": candidate.decision_seq,
-            })
-    for decision in decisions:
-        if decision.type not in ("evidence_accepted", "evidence_rejected"):
-            continue
-        entries[decision.subject] = {
-            "id": decision.subject,
-            "file": _resolve_session_citation(evidence_root, decision.subject),
-            "esito": "accettata" if decision.type == "evidence_accepted" else "scartata",
-            "decision_seq": decision.seq,
-        }
-    return list(entries.values())
+from tht.evidence.preprocessing import EvidenceEmbedder, build_preprocessing_pipeline
+from tht.evidence.search import (
+    ActiveEvidenceSearcher,
+    CorpusWorkspaceMismatchError,
+    active_searcher,
+    build_retrieval_entries,
+    resolve_citation,
+    validate_corpus_workspace,
+)
+from tht.evidence.session import project_session
+from tht.evidence.sources import build_sources
 
 
 __all__ = [
@@ -168,7 +30,10 @@ __all__ = [
     "EvidenceSource",
     "EvidenceSourceError",
     "EvidenceSourceErrorCategory",
+    "EvidenceEmbedder",
     "SourceObject",
+    "ActiveEvidenceSearcher",
+    "CorpusWorkspaceMismatchError",
     "acquire",
     "active_searcher",
     "build_preprocessing_pipeline",

harness/tht/evidence/adapters/__init__.py

@@ -0,0 +1,7 @@
+"""Evidence-owned source adapter implementations."""
+
+from tht.evidence.adapters.filesystem import FilesystemEvidenceSource
+from tht.evidence.adapters.http import HttpManifestEvidenceSource
+from tht.evidence.adapters.s3 import S3EvidenceSource
+
+__all__ = ["FilesystemEvidenceSource", "HttpManifestEvidenceSource", "S3EvidenceSource"]

harness/tht/evidence/adapters/filesystem.py

@@ -1,4 +1,4 @@
-"""Contained, race-safe filesystem Evidence source."""
+"""Evidence-owned, race-safe filesystem source."""
 
 import hashlib
 import os

harness/tht/evidence/adapters/http.py

@@ -1,4 +1,4 @@
-"""Explicit-manifest HTTP Evidence source with SSRF-safe bounded acquisition."""
+"""Evidence-owned explicit-manifest HTTP source with SSRF-safe bounded acquisition."""
 
 import hashlib
 import ipaddress

harness/tht/evidence/adapters/s3.py

@@ -1,4 +1,4 @@
-"""Bounded S3-compatible Evidence source using the supported boto3 client."""
+"""Evidence-owned bounded S3-compatible source using the supported boto3 client."""
 
 import hashlib
 import ipaddress

harness/tht/evidence/corpus/__init__.py

@@ -0,0 +1 @@
+"""Evidence-owned canonical, transport-independent corpus."""

harness/tht/evidence/corpus/chunk.py

@@ -1,11 +1,11 @@
-"""Versioned deterministic chunking for canonical corpus documents."""
+"""Evidence-owned deterministic chunking for canonical corpus documents."""
 
 import hashlib
 import json
 import re
 from dataclasses import asdict, dataclass
 
-from tht.corpus.models import CanonicalChunk, CanonicalDocument
+from tht.evidence.corpus.models import CanonicalChunk, CanonicalDocument
 
 
 @dataclass(frozen=True, slots=True)

harness/tht/evidence/corpus/models.py

@@ -1,4 +1,4 @@
-"""Immutable records emitted by the Evidence preprocessing pipeline."""
+"""Evidence-owned immutable records emitted by preprocessing."""
 
 import hashlib
 import re

harness/tht/evidence/corpus/normalize.py

@@ -1,4 +1,4 @@
-"""Pure, deterministic conversion of acquired bytes into canonical text."""
+"""Evidence-owned deterministic conversion of acquired bytes into canonical text."""
 
 import hashlib
 import re
@@ -10,7 +10,7 @@ from pydantic import JsonValue, TypeAdapter, ValidationError
 from yaml.events import AliasEvent
 from yaml.nodes import MappingNode
 
-from tht.corpus.models import CanonicalDocument
+from tht.evidence.corpus.models import CanonicalDocument
 from tht.evidence.contracts import AcquiredDocument, canonical_provenance_uri
 
 

harness/tht/evidence/corpus/pipeline.py

@@ -1,4 +1,4 @@
-"""Incremental Evidence preprocessing with generation-isolated vector writes."""
+"""Evidence-owned incremental preprocessing with generation-isolated vector writes."""
 
 from __future__ import annotations
 
@@ -11,10 +11,10 @@ from dataclasses import asdict, dataclass, field
 from datetime import UTC
 from pathlib import Path
 
-from tht.corpus.chunk import ChunkPolicy, chunk
-from tht.corpus.models import CanonicalChunk, CanonicalDocument, CorpusManifest
-from tht.corpus.normalize import normalize
-from tht.corpus.store import CorpusStore
+from tht.evidence.corpus.chunk import ChunkPolicy, chunk
+from tht.evidence.corpus.models import CanonicalChunk, CanonicalDocument, CorpusManifest
+from tht.evidence.corpus.normalize import normalize
+from tht.evidence.corpus.store import CorpusStore
 import tht.evidence.acquisition as evidence_acquisition
 from tht.evidence.contracts import EvidenceSource, SourceObject, canonical_provenance_uri
 from tht.ports.vector import VectorStore, VectorWriteRecord

harness/tht/evidence/corpus/store.py

@@ -1,4 +1,4 @@
-"""Durable immutable corpus generations and an atomic ACTIVE pointer."""
+"""Evidence-owned immutable corpus generations and an atomic ACTIVE pointer."""
 
 from __future__ import annotations
 
@@ -15,7 +15,7 @@ from datetime import UTC, datetime
 from pathlib import Path
 from contextlib import contextmanager
 
-from tht.corpus.models import CorpusManifest
+from tht.evidence.corpus.models import CorpusManifest
 
 
 _GENERATION = re.compile(r"^gen:[0-9a-f]{32}$")

harness/tht/evidence/preprocessing.py

@@ -2,9 +2,9 @@
 
 from typing import Protocol
 
-from tht.corpus.chunk import ChunkPolicy
-from tht.corpus.pipeline import CorpusPipeline
-from tht.corpus.store import CorpusStore
+from tht.evidence.corpus.chunk import ChunkPolicy
+from tht.evidence.corpus.pipeline import CorpusPipeline
+from tht.evidence.corpus.store import CorpusStore
 from tht.evidence.contracts import EvidenceSource
 from tht.ports.vector import VectorStore
 

harness/tht/evidence/search.py

@@ -1,8 +1,8 @@
-"""Runtime Evidence lookup bound to the atomically active corpus generation."""
+"""Evidence-owned runtime lookup bound to the atomically active corpus generation."""
 
 import re
 
-from tht.corpus.store import CorpusStore
+from tht.evidence.corpus.store import CorpusStore
 
 
 class CorpusWorkspaceMismatchError(RuntimeError):
@@ -96,7 +96,7 @@ def validate_corpus_workspace(cfg, workspace_id: str) -> None:
             )
 
 
-def resolve_evidence_file(
+def resolve_citation(
     store: CorpusStore, evidence_id: str, *, materialized_root=None,
 ) -> str:
     with store.writer_lock():
@@ -114,3 +114,25 @@ def resolve_evidence_file(
                 )
                 return str(path) if path else ""
     return ""
+
+
+def build_retrieval_entries(results, *, excerpt_chars: int) -> list[dict]:
+    """Project ordered Evidence search hits into the retrieval-pack shape."""
+    return [
+        {
+            "title": result.label,
+            "status": result.status,
+            "excerpt": result.content[:excerpt_chars],
+        }
+        for result in results
+    ]
+
+
+__all__ = [
+    "ActiveEvidenceSearcher",
+    "CorpusWorkspaceMismatchError",
+    "active_searcher",
+    "build_retrieval_entries",
+    "resolve_citation",
+    "validate_corpus_workspace",
+]

harness/tht/evidence/session.py

@@ -0,0 +1,59 @@
+"""Evidence-specific projection into persisted session artifacts."""
+
+from pathlib import Path
+from typing import TYPE_CHECKING
+
+from tht.evidence.corpus.store import CorpusStore
+from tht.evidence.search import resolve_citation
+
+if TYPE_CHECKING:
+    from tht.decisions import DecisionRecord
+    from tht.session.models import SchemaLinking
+
+
+def _resolve_session_citation(evidence_root: Path, evidence_id: str) -> str:
+    # New deployments resolve only immutable materialized files from ACTIVE. Keep the
+    # curated-tree fallback for sessions created before a canonical corpus exists.
+    corpus_root = evidence_root.parent.parent / "corpus"
+    if corpus_root.exists():
+        return resolve_citation(
+            CorpusStore(corpus_root),
+            evidence_id,
+            materialized_root=evidence_root.parent / ".materialized-evidence",
+        )
+    for match in evidence_root.rglob(f"{evidence_id}.md"):
+        return str(match)
+    return ""
+
+
+def project_session(
+    decisions: list["DecisionRecord"],
+    linking: "SchemaLinking",
+    evidence_root: Path,
+) -> list[dict]:
+    """Project cited and reviewed Evidence into the session artifact shape."""
+    entries: dict[str, dict] = {}
+    for candidate in linking.candidates:
+        for evidence_id in candidate.evidence:
+            entries.setdefault(
+                evidence_id,
+                {
+                    "id": evidence_id,
+                    "file": _resolve_session_citation(evidence_root, evidence_id),
+                    "esito": "usata",
+                    "decision_seq": candidate.decision_seq,
+                },
+            )
+    for decision in decisions:
+        if decision.type not in ("evidence_accepted", "evidence_rejected"):
+            continue
+        entries[decision.subject] = {
+            "id": decision.subject,
+            "file": _resolve_session_citation(evidence_root, decision.subject),
+            "esito": "accettata" if decision.type == "evidence_accepted" else "scartata",
+            "decision_seq": decision.seq,
+        }
+    return list(entries.values())
+
+
+__all__ = ["project_session"]

harness/tht/evidence/sources.py

@@ -1,10 +1,16 @@
 """Construction of configured Evidence source adapters."""
 
-from tht.adapters.evidence import FilesystemEvidenceSource, HttpManifestEvidenceSource
-from tht.adapters.evidence.s3 import S3EvidenceSource
-from tht.config import ConfigError, EvidenceSourcesConfig
+from __future__ import annotations
+
+from typing import TYPE_CHECKING
+
+from tht.evidence.adapters import FilesystemEvidenceSource, HttpManifestEvidenceSource
+from tht.evidence.adapters.s3 import S3EvidenceSource
 from tht.evidence.contracts import EvidenceSource
 
+if TYPE_CHECKING:
+    from tht.config import EvidenceSourcesConfig
+
 
 def build_sources(evidence: EvidenceSourcesConfig | None) -> list[EvidenceSource]:
     """Build configured Evidence adapters in the existing deterministic order."""
@@ -59,6 +65,8 @@ def build_sources(evidence: EvidenceSourcesConfig | None) -> list[EvidenceSource
                     )
                 )
             case other:  # pragma: no cover - Pydantic rejects unsupported discriminators.
+                from tht.config import ConfigError
+
                 raise ConfigError(f"Adapter evidence non supportato: {other}")
     return sources
 

harness/tht/ports/evidence.py

@@ -1,31 +0,0 @@
-"""Legacy import path for Evidence contracts.
-
-New production code imports the leaf contracts from ``tht.evidence.contracts``. This one-way
-compatibility shim remains only until the old Evidence layout is removed.
-"""
-
-from tht.evidence.contracts import (
-    AcquiredDocument,
-    EvidenceSource,
-    EvidenceSourceError,
-    EvidenceSourceErrorCategory,
-    SourceObject,
-    canonical_provenance_uri,
-    normalize_aware_datetime,
-    validate_canonical_uri,
-    validate_namespaced_value,
-    validate_safe_metadata,
-)
-
-__all__ = [
-    "AcquiredDocument",
-    "EvidenceSource",
-    "EvidenceSourceError",
-    "EvidenceSourceErrorCategory",
-    "SourceObject",
-    "canonical_provenance_uri",
-    "normalize_aware_datetime",
-    "validate_canonical_uri",
-    "validate_namespaced_value",
-    "validate_safe_metadata",
-]

harness/tht/session/artifacts.py

@@ -1,17 +0,0 @@
-"""Legacy session-artifact entrypoints retained during the Evidence migration."""
-
-from pathlib import Path
-
-from tht.decisions import DecisionRecord
-from tht.session.models import SchemaLinking
-
-
-def build_evidence_entries(
-    decisions: list[DecisionRecord],
-    linking: SchemaLinking,
-    evidence_root: Path,
-) -> list[dict]:
-    """Compatibility shim for callers not yet migrated to ``tht.evidence``."""
-    from tht.evidence import project_session
-
-    return project_session(decisions, linking, evidence_root)