Codex
82e2c91f42
Publish documentation / publish (push) Successful in 1m27s
Add PostgreSQL-backed memory, editable evidence with source review and activation, and human-approved archive repairs across the harness, API, and UI. Include migrations, deployment support, regression coverage, and validation documentation. Refresh permissions from validated session roles so existing administrator logins can access newly deployed archive management features.
754 lines
31 KiB
Python
754 lines
31 KiB
Python
import hashlib
|
|
import subprocess
|
|
import threading
|
|
import unicodedata
|
|
|
|
import pytest
|
|
from pydantic import ValidationError
|
|
|
|
from tht.evidence import (
|
|
CuratedEvidence,
|
|
EvidenceManifest,
|
|
EvidencePreparationError,
|
|
EvidenceRestructurer,
|
|
RestructureCandidate,
|
|
dump_curated_markdown,
|
|
dump_manifest,
|
|
load_curated_tree,
|
|
load_manifest,
|
|
migrate_workspace_evidence,
|
|
prepare_workspace_evidence,
|
|
validate_workspace_evidence,
|
|
)
|
|
|
|
|
|
def _evidence(source_text: str, *, review_items=()):
|
|
normalized_source = unicodedata.normalize(
|
|
"NFC", source_text.replace("\r\n", "\n").replace("\r", "\n"),
|
|
)
|
|
if not normalized_source.endswith("\n"):
|
|
normalized_source += "\n"
|
|
return CuratedEvidence.model_validate({
|
|
"schema_version": 1,
|
|
"id": "evidence:fascia-pediatrica",
|
|
"title": "Fascia pediatrica",
|
|
"kind": "domain",
|
|
"purposes": ["disambiguation"],
|
|
"applies_to": {"concepts": ["fascia pediatrica"]},
|
|
"language": "it",
|
|
"provenance": {
|
|
"source_file": "source/domain/patient.md",
|
|
"source_sha256": "sha256:" + hashlib.sha256(normalized_source.encode()).hexdigest(),
|
|
"supporting_excerpts": ["I pazienti sotto i 18 anni sono pediatrici."],
|
|
},
|
|
"review_items": list(review_items),
|
|
"payload": {"rule": "La fascia pediatrica comprende i minori."},
|
|
})
|
|
|
|
|
|
def _write_workspace(root, evidence, source_text):
|
|
evidence_root = root / "evidence"
|
|
source_path = evidence_root / "source" / "domain" / "patient.md"
|
|
curated_path = evidence_root / "curated" / "domain" / "fascia-pediatrica.md"
|
|
source_path.parent.mkdir(parents=True)
|
|
curated_path.parent.mkdir(parents=True)
|
|
source_path.write_text(source_text, encoding="utf-8")
|
|
curated_path.write_text(dump_curated_markdown(evidence), encoding="utf-8")
|
|
(evidence_root / "manifest.yaml").write_text(dump_manifest(EvidenceManifest.model_validate({
|
|
"schema_version": 1,
|
|
"pipeline_version": "evidence-authoring-v1",
|
|
"sources": {
|
|
evidence.provenance.source_file: {
|
|
"sha256": evidence.provenance.source_sha256,
|
|
"units": [evidence.id],
|
|
},
|
|
},
|
|
"orphans": [],
|
|
})), encoding="utf-8")
|
|
|
|
|
|
def test_workspace_validation_allows_a_coherent_curated_unit(tmp_path):
|
|
source_text = "I pazienti sotto i 18 anni sono pediatrici."
|
|
_write_workspace(tmp_path, _evidence(source_text), source_text)
|
|
|
|
report = validate_workspace_evidence(tmp_path)
|
|
|
|
assert report.publishable is True
|
|
assert report.findings == ()
|
|
|
|
|
|
def test_workspace_validation_keeps_review_items_visible_and_blocks_publication(tmp_path):
|
|
source_text = "I pazienti sotto i 18 anni sono pediatrici."
|
|
_write_workspace(tmp_path, _evidence(source_text, review_items=[{
|
|
"code": "ambiguous_source_statement",
|
|
"message": "Il sorgente non chiarisce la data di riferimento.",
|
|
"field": "payload.rule",
|
|
}]), source_text)
|
|
|
|
report = validate_workspace_evidence(tmp_path)
|
|
|
|
assert report.publishable is False
|
|
assert [finding.code for finding in report.findings] == ["unresolved_review_item"]
|
|
|
|
|
|
def test_manifest_round_trip_is_versioned_and_deterministic(tmp_path):
|
|
manifest = EvidenceManifest.model_validate({
|
|
"schema_version": 1,
|
|
"pipeline_version": "evidence-authoring-v1",
|
|
"sources": {
|
|
"source/domain/patient.md": {
|
|
"sha256": "sha256:" + "a" * 64,
|
|
"units": ["evidence:fascia-pediatrica"],
|
|
},
|
|
},
|
|
"orphans": [],
|
|
})
|
|
path = tmp_path / "evidence" / "manifest.yaml"
|
|
path.parent.mkdir(parents=True)
|
|
path.write_text(dump_manifest(manifest), encoding="utf-8")
|
|
|
|
loaded = load_manifest(path)
|
|
|
|
assert loaded == manifest
|
|
assert path.read_text(encoding="utf-8") == dump_manifest(manifest)
|
|
|
|
|
|
def test_manifest_rejects_unknown_fields_and_incompatible_pipeline_version():
|
|
with pytest.raises(ValidationError):
|
|
EvidenceManifest.model_validate({
|
|
"schema_version": 1,
|
|
"pipeline_version": "evidence-authoring-v2",
|
|
"sources": {},
|
|
"orphans": [],
|
|
"approved": True,
|
|
})
|
|
|
|
|
|
def test_workspace_validation_requires_a_manifest(tmp_path):
|
|
source_text = "I pazienti sotto i 18 anni sono pediatrici."
|
|
_write_workspace(tmp_path, _evidence(source_text), source_text)
|
|
(tmp_path / "evidence" / "manifest.yaml").unlink()
|
|
|
|
report = validate_workspace_evidence(tmp_path)
|
|
|
|
assert report.publishable is False
|
|
assert [finding.code for finding in report.findings] == ["manifest_missing"]
|
|
|
|
|
|
def test_workspace_validation_reports_a_non_utf8_manifest_without_raising(tmp_path):
|
|
evidence_root = tmp_path / "evidence"
|
|
evidence_root.mkdir()
|
|
(evidence_root / "manifest.yaml").write_bytes(b"\xff")
|
|
|
|
report = validate_workspace_evidence(tmp_path)
|
|
|
|
assert report.publishable is False
|
|
assert [finding.code for finding in report.findings] == ["manifest_invalid"]
|
|
|
|
|
|
def test_workspace_validation_reports_invalid_curated_documents_without_raising(tmp_path):
|
|
source_text = "I pazienti sotto i 18 anni sono pediatrici."
|
|
_write_workspace(tmp_path, _evidence(source_text), source_text)
|
|
(tmp_path / "evidence" / "curated" / "domain" / "fascia-pediatrica.md").write_text(
|
|
"not canonical frontmatter", encoding="utf-8",
|
|
)
|
|
|
|
report = validate_workspace_evidence(tmp_path)
|
|
|
|
assert report.publishable is False
|
|
assert [finding.code for finding in report.findings] == ["curated_invalid"]
|
|
|
|
|
|
def test_workspace_validation_preserves_manifest_orphans_but_blocks_publication(tmp_path):
|
|
source_text = "I pazienti sotto i 18 anni sono pediatrici."
|
|
_write_workspace(tmp_path, _evidence(source_text), source_text)
|
|
manifest_path = tmp_path / "evidence" / "manifest.yaml"
|
|
manifest = load_manifest(manifest_path).model_copy(update={"orphans": ("evidence:retired",)})
|
|
manifest_path.write_text(dump_manifest(manifest), encoding="utf-8")
|
|
|
|
report = validate_workspace_evidence(tmp_path)
|
|
|
|
assert report.publishable is False
|
|
assert [finding.code for finding in report.findings] == ["orphaned_unit"]
|
|
|
|
|
|
def test_workspace_validation_requires_manifest_and_provenance_to_agree(tmp_path):
|
|
source_text = "I pazienti sotto i 18 anni sono pediatrici."
|
|
evidence = _evidence(source_text)
|
|
_write_workspace(tmp_path, evidence, source_text)
|
|
manifest_path = tmp_path / "evidence" / "manifest.yaml"
|
|
manifest = EvidenceManifest.model_validate({
|
|
"schema_version": 1,
|
|
"pipeline_version": "evidence-authoring-v1",
|
|
"sources": {
|
|
evidence.provenance.source_file: {
|
|
"sha256": "sha256:" + "b" * 64,
|
|
"units": ["evidence:another-unit"],
|
|
},
|
|
},
|
|
"orphans": [],
|
|
})
|
|
manifest_path.write_text(dump_manifest(manifest), encoding="utf-8")
|
|
|
|
report = validate_workspace_evidence(tmp_path)
|
|
|
|
assert report.publishable is False
|
|
assert {finding.code for finding in report.findings} == {
|
|
"manifest_source_hash_mismatch",
|
|
"manifest_unit_missing",
|
|
"source_hash_mismatch",
|
|
"manifest_unit_without_curated",
|
|
}
|
|
|
|
|
|
def test_workspace_validation_checks_unreferenced_manifest_sources_and_units(tmp_path):
|
|
source_text = "I pazienti sotto i 18 anni sono pediatrici."
|
|
_write_workspace(tmp_path, _evidence(source_text), source_text)
|
|
manifest_path = tmp_path / "evidence" / "manifest.yaml"
|
|
manifest = EvidenceManifest.model_validate({
|
|
"schema_version": 1,
|
|
"pipeline_version": "evidence-authoring-v1",
|
|
"sources": {
|
|
"source/domain/missing.md": {
|
|
"sha256": "sha256:" + "a" * 64,
|
|
"units": ["evidence:removed"],
|
|
},
|
|
},
|
|
"orphans": [],
|
|
})
|
|
manifest_path.write_text(dump_manifest(manifest), encoding="utf-8")
|
|
|
|
report = validate_workspace_evidence(tmp_path)
|
|
|
|
assert report.publishable is False
|
|
assert {finding.code for finding in report.findings} == {
|
|
"manifest_source_missing",
|
|
"source_missing",
|
|
"manifest_unit_without_curated",
|
|
}
|
|
|
|
|
|
def test_manifest_rejects_unsafe_source_paths_and_legacy_unit_identifiers():
|
|
with pytest.raises(ValidationError):
|
|
EvidenceManifest.model_validate({
|
|
"schema_version": 1,
|
|
"pipeline_version": "evidence-authoring-v1",
|
|
"sources": {"../patient.pdf": {"sha256": "sha256:" + "a" * 64, "units": ["domain:patient"]}},
|
|
"orphans": [],
|
|
})
|
|
|
|
|
|
def test_manifest_rejects_a_unit_declared_by_two_sources():
|
|
with pytest.raises(ValidationError, match="only one source"):
|
|
EvidenceManifest.model_validate({
|
|
"schema_version": 1,
|
|
"pipeline_version": "evidence-authoring-v1",
|
|
"sources": {
|
|
"source/domain/one.md": {
|
|
"sha256": "sha256:" + "a" * 64,
|
|
"units": ["evidence:shared"],
|
|
},
|
|
"source/domain/two.md": {
|
|
"sha256": "sha256:" + "b" * 64,
|
|
"units": ["evidence:shared"],
|
|
},
|
|
},
|
|
"orphans": [],
|
|
})
|
|
|
|
|
|
def test_workspace_validation_rejects_duplicate_curated_ids_and_changed_sources(tmp_path):
|
|
source_text = "I pazienti sotto i 18 anni sono pediatrici."
|
|
evidence = _evidence(source_text)
|
|
_write_workspace(tmp_path, evidence, source_text)
|
|
duplicate_path = tmp_path / "evidence" / "curated" / "domain" / "duplicate.md"
|
|
duplicate_path.write_text(dump_curated_markdown(evidence), encoding="utf-8")
|
|
(tmp_path / "evidence" / "source" / "domain" / "patient.md").write_text(
|
|
"Il testo sorgente è cambiato.", encoding="utf-8",
|
|
)
|
|
|
|
report = validate_workspace_evidence(tmp_path)
|
|
|
|
assert report.publishable is False
|
|
assert {finding.code for finding in report.findings} == {
|
|
"duplicate_evidence_id",
|
|
"source_hash_mismatch",
|
|
"supporting_excerpt_missing",
|
|
}
|
|
|
|
|
|
def test_workspace_validation_reports_unreadable_or_oversized_sources(tmp_path, monkeypatch):
|
|
source_text = "I pazienti sotto i 18 anni sono pediatrici."
|
|
_write_workspace(tmp_path, _evidence(source_text), source_text)
|
|
source_path = tmp_path / "evidence" / "source" / "domain" / "patient.md"
|
|
source_path.write_bytes(b"\xff")
|
|
|
|
unreadable = validate_workspace_evidence(tmp_path)
|
|
|
|
assert [finding.code for finding in unreadable.findings] == ["source_unreadable"]
|
|
|
|
source_path.write_text(source_text, encoding="utf-8")
|
|
monkeypatch.setattr("tht.evidence.authoring.MAX_AUTHORING_FILE_BYTES", 1)
|
|
|
|
oversized = validate_workspace_evidence(tmp_path)
|
|
|
|
assert [finding.code for finding in oversized.findings] == ["source_oversized"]
|
|
|
|
|
|
def test_workspace_validation_rejects_a_source_symlink(tmp_path):
|
|
source_text = "I pazienti sotto i 18 anni sono pediatrici."
|
|
_write_workspace(tmp_path, _evidence(source_text), source_text)
|
|
source_path = tmp_path / "evidence" / "source" / "domain" / "patient.md"
|
|
outside = tmp_path / "outside.md"
|
|
outside.write_text(source_text, encoding="utf-8")
|
|
source_path.unlink()
|
|
source_path.symlink_to(outside)
|
|
|
|
report = validate_workspace_evidence(tmp_path)
|
|
|
|
assert [finding.code for finding in report.findings] == ["source_unsafe"]
|
|
|
|
|
|
class _Restructurer(EvidenceRestructurer):
|
|
def __init__(self, candidates):
|
|
self.candidates = candidates
|
|
self.requests = []
|
|
|
|
def restructure(self, request):
|
|
self.requests.append(request)
|
|
return tuple(self.candidates)
|
|
|
|
|
|
def _candidate(*, title="Fascia pediatrica", existing_id=None, kind="domain", payload=None):
|
|
return RestructureCandidate.model_validate({
|
|
"schema_version": 1,
|
|
"existing_id": existing_id,
|
|
"title": title,
|
|
"kind": kind,
|
|
"purposes": ["disambiguation"],
|
|
"applies_to": {"concepts": ["fascia pediatrica"]},
|
|
"language": "it",
|
|
"supporting_excerpts": ["I pazienti sotto i 18 anni sono pediatrici."],
|
|
"review_items": [],
|
|
"payload": payload or {"rule": "La fascia pediatrica comprende i minori."},
|
|
})
|
|
|
|
|
|
def test_prepare_changed_source_uses_one_model_call_and_applies_a_valid_batch(tmp_path):
|
|
source_text = "I pazienti sotto i 18 anni sono pediatrici."
|
|
_write_workspace(tmp_path, _evidence(source_text), source_text)
|
|
source_path = tmp_path / "evidence" / "source" / "domain" / "patient.md"
|
|
source_path.write_text(source_text + " La regola è revisionata.\n", encoding="utf-8")
|
|
restructurer = _Restructurer([_candidate(existing_id="evidence:fascia-pediatrica")])
|
|
|
|
report = prepare_workspace_evidence(tmp_path, restructurer=restructurer, git_status=lambda _: ())
|
|
|
|
assert report.model_calls == 1
|
|
assert report.changed == ("source/domain/patient.md",)
|
|
assert report.created == ()
|
|
assert len(restructurer.requests) == 1
|
|
assert restructurer.requests[0].previous_units[0].id == "evidence:fascia-pediatrica"
|
|
curated_path = tmp_path / "evidence" / "curated" / "domain" / "fascia-pediatrica.md"
|
|
curated = load_curated_tree(tmp_path / "evidence" / "curated")[0]
|
|
assert curated.schema_version == 4
|
|
assert "# Fascia pediatrica\n" in curated_path.read_text(encoding="utf-8")
|
|
assert validate_workspace_evidence(tmp_path).publishable is True
|
|
|
|
|
|
def test_migrate_workspace_evidence_rewrites_v1_units_as_v4_without_a_model_call(tmp_path):
|
|
source_text = "I pazienti sotto i 18 anni sono pediatrici."
|
|
_write_workspace(tmp_path, _evidence(source_text), source_text)
|
|
|
|
report = migrate_workspace_evidence(tmp_path, git_status=lambda _: ())
|
|
|
|
curated_path = tmp_path / "evidence" / "curated" / "domain" / "fascia-pediatrica.md"
|
|
migrated = load_curated_tree(tmp_path / "evidence" / "curated")[0]
|
|
assert report.migrated == ("evidence:fascia-pediatrica",)
|
|
assert report.unchanged == ()
|
|
assert migrated.schema_version == 4
|
|
assert migrated.payload.rule == "La fascia pediatrica comprende i minori."
|
|
text = curated_path.read_text(encoding="utf-8")
|
|
assert "## Regola\n\n" in text
|
|
assert "<!-- tht:" not in text
|
|
assert "La fascia pediatrica comprende i minori." in text
|
|
assert report.findings == ()
|
|
|
|
|
|
def test_migrate_workspace_evidence_rewrites_v2_units_as_editable_v4(tmp_path):
|
|
source_text = "I pazienti sotto i 18 anni sono pediatrici."
|
|
evidence = _evidence(source_text).model_copy(update={"schema_version": 2})
|
|
_write_workspace(tmp_path, evidence, source_text)
|
|
|
|
first = migrate_workspace_evidence(tmp_path, git_status=lambda _: ())
|
|
second = migrate_workspace_evidence(tmp_path, git_status=lambda _: ())
|
|
|
|
curated_path = tmp_path / "evidence" / "curated" / "domain" / "fascia-pediatrica.md"
|
|
text = curated_path.read_text(encoding="utf-8")
|
|
migrated = load_curated_tree(tmp_path / "evidence" / "curated")[0]
|
|
assert first.migrated == ("evidence:fascia-pediatrica",)
|
|
assert first.unchanged == ()
|
|
assert second.migrated == ()
|
|
assert second.unchanged == ("evidence:fascia-pediatrica",)
|
|
assert migrated.schema_version == 4
|
|
assert text.startswith("---\nschema_version: 4")
|
|
assert not any(line.startswith("|") for line in text.splitlines())
|
|
|
|
|
|
def test_migrate_workspace_evidence_rewrites_legacy_v3_rule_presentation(tmp_path):
|
|
source_text = "I pazienti sotto i 18 anni sono pediatrici."
|
|
evidence = _evidence(source_text).model_copy(update={"schema_version": 3})
|
|
_write_workspace(tmp_path, evidence, source_text)
|
|
curated_path = tmp_path / "evidence" / "curated" / "domain" / "fascia-pediatrica.md"
|
|
legacy_text = curated_path.read_text(encoding="utf-8")
|
|
legacy_text = legacy_text.replace(
|
|
"## Regola\n\n<!-- tht:raw-rule:",
|
|
"## Regola\n\n<!-- tht:legacy-raw-rule:",
|
|
1,
|
|
)
|
|
# Recreate the exact pre-structured v3 body from the canonical value.
|
|
start = legacy_text.index("<!-- tht:field:rule -->")
|
|
end = legacy_text.index("<!-- /tht:field:rule -->", start)
|
|
legacy_rule = (
|
|
"<!-- tht:field:rule -->\n"
|
|
"## Regola\n\n"
|
|
f"{evidence.payload.rule}\n"
|
|
)
|
|
curated_path.write_text(
|
|
legacy_text[:start] + legacy_rule + legacy_text[end:],
|
|
encoding="utf-8",
|
|
)
|
|
|
|
report = migrate_workspace_evidence(tmp_path, git_status=lambda _: ())
|
|
|
|
migrated_text = curated_path.read_text(encoding="utf-8")
|
|
assert report.migrated == ("evidence:fascia-pediatrica",)
|
|
assert report.unchanged == ()
|
|
assert "## Regola\n\n" in migrated_text
|
|
assert "<!-- tht:" not in migrated_text
|
|
assert load_curated_tree(tmp_path / "evidence" / "curated")[0].payload.rule == (
|
|
evidence.payload.rule
|
|
)
|
|
|
|
|
|
def test_migrate_workspace_evidence_preserves_uncommitted_content_without_git_operations(tmp_path):
|
|
subprocess.run(["git", "init", "--quiet", str(tmp_path)], check=True)
|
|
workspace_root = tmp_path / "psd-clinical"
|
|
source_text = "I pazienti sotto i 18 anni sono pediatrici."
|
|
_write_workspace(workspace_root, _evidence(source_text), source_text)
|
|
|
|
result = migrate_workspace_evidence(workspace_root)
|
|
assert result.migrated == ("evidence:fascia-pediatrica",)
|
|
assert load_curated_tree(workspace_root / "evidence/curated")[0].payload.rule == _evidence(source_text).payload.rule
|
|
assert subprocess.run(["git", "-C", str(tmp_path), "rev-parse", "HEAD"],
|
|
capture_output=True, check=False).returncode != 0
|
|
|
|
|
|
def test_prepare_can_issue_independent_source_calls_concurrently(tmp_path):
|
|
source_text = "I pazienti sotto i 18 anni sono pediatrici."
|
|
_write_workspace(tmp_path, _evidence(source_text), source_text)
|
|
source_root = tmp_path / "evidence" / "source" / "domain"
|
|
changed_text = source_text + "\nRegola revisionata.\n"
|
|
(source_root / "patient.md").write_text(changed_text, encoding="utf-8")
|
|
(source_root / "second.md").write_text(changed_text, encoding="utf-8")
|
|
barrier = threading.Barrier(2)
|
|
|
|
class ConcurrentRestructurer:
|
|
def __init__(self):
|
|
self.requests = []
|
|
|
|
def restructure(self, request):
|
|
self.requests.append(request)
|
|
barrier.wait(timeout=2)
|
|
return (_candidate(title=f"Unit {request.source_file}"),)
|
|
|
|
restructurer = ConcurrentRestructurer()
|
|
|
|
report = prepare_workspace_evidence(
|
|
tmp_path,
|
|
restructurer=restructurer,
|
|
git_status=lambda _: (),
|
|
max_workers=2,
|
|
)
|
|
|
|
assert report.model_calls == 2
|
|
assert sorted(request.source_file for request in restructurer.requests) == [
|
|
"source/domain/patient.md",
|
|
"source/domain/second.md",
|
|
]
|
|
|
|
|
|
def test_prepare_unchanged_source_skips_model_and_does_not_write(tmp_path):
|
|
source_text = "I pazienti sotto i 18 anni sono pediatrici."
|
|
_write_workspace(tmp_path, _evidence(source_text), source_text)
|
|
curated_path = tmp_path / "evidence" / "curated" / "domain" / "fascia-pediatrica.md"
|
|
original = curated_path.read_text(encoding="utf-8")
|
|
restructurer = _Restructurer([])
|
|
|
|
report = prepare_workspace_evidence(tmp_path, restructurer=restructurer, git_status=lambda _: ())
|
|
|
|
assert report.model_calls == 0
|
|
assert report.unchanged == ("source/domain/patient.md",)
|
|
assert curated_path.read_text(encoding="utf-8") == original
|
|
assert restructurer.requests == []
|
|
|
|
|
|
def test_prepare_rejects_dirty_curated_state_before_model_call(tmp_path):
|
|
source_text = "I pazienti sotto i 18 anni sono pediatrici."
|
|
_write_workspace(tmp_path, _evidence(source_text), source_text)
|
|
restructurer = _Restructurer([])
|
|
|
|
with pytest.raises(EvidencePreparationError, match="authoring_worktree_dirty"):
|
|
prepare_workspace_evidence(
|
|
tmp_path,
|
|
restructurer=restructurer,
|
|
git_status=lambda _: (" M evidence/curated/domain/patient.md",),
|
|
)
|
|
|
|
assert restructurer.requests == []
|
|
|
|
|
|
def test_prepare_refuses_an_incompatible_pipeline_version_without_writing(tmp_path):
|
|
source_text = "I pazienti sotto i 18 anni sono pediatrici."
|
|
_write_workspace(tmp_path, _evidence(source_text), source_text)
|
|
manifest_path = tmp_path / "evidence" / "manifest.yaml"
|
|
curated_path = tmp_path / "evidence" / "curated" / "domain" / "fascia-pediatrica.md"
|
|
manifest_path.write_text(
|
|
manifest_path.read_text(encoding="utf-8").replace("evidence-authoring-v1", "evidence-authoring-v2"),
|
|
encoding="utf-8",
|
|
)
|
|
original_manifest = manifest_path.read_text(encoding="utf-8")
|
|
original_curated = curated_path.read_text(encoding="utf-8")
|
|
|
|
with pytest.raises(EvidencePreparationError, match="pipeline_upgrade_required"):
|
|
prepare_workspace_evidence(tmp_path, restructurer=_Restructurer([]), git_status=lambda _: ())
|
|
|
|
assert manifest_path.read_text(encoding="utf-8") == original_manifest
|
|
assert curated_path.read_text(encoding="utf-8") == original_curated
|
|
|
|
|
|
def test_prepare_preserves_an_id_when_reclassified_and_allocates_a_new_id_for_a_split(tmp_path):
|
|
source_text = "I pazienti sotto i 18 anni sono pediatrici."
|
|
_write_workspace(tmp_path, _evidence(source_text), source_text)
|
|
source_path = tmp_path / "evidence" / "source" / "domain" / "patient.md"
|
|
source_path.write_text(source_text + " I pazienti sopra i 18 anni sono adulti.\n", encoding="utf-8")
|
|
restructurer = _Restructurer([
|
|
_candidate(
|
|
existing_id="evidence:fascia-pediatrica",
|
|
kind="glossary",
|
|
payload={"definition": "Paziente con meno di 18 anni."},
|
|
),
|
|
_candidate(title="Fascia adulta"),
|
|
])
|
|
|
|
report = prepare_workspace_evidence(tmp_path, restructurer=restructurer, git_status=lambda _: ())
|
|
|
|
assert report.created == ("evidence:fascia-adulta",)
|
|
documents = {document.id: document for document in load_curated_tree(tmp_path / "evidence" / "curated")}
|
|
assert documents["evidence:fascia-pediatrica"].kind == "glossary"
|
|
assert documents["evidence:fascia-adulta"].kind == "domain"
|
|
|
|
|
|
def test_prepare_retains_units_from_a_removed_source_as_blocking_orphans(tmp_path):
|
|
source_text = "I pazienti sotto i 18 anni sono pediatrici."
|
|
_write_workspace(tmp_path, _evidence(source_text), source_text)
|
|
(tmp_path / "evidence" / "source" / "domain" / "patient.md").unlink()
|
|
restructurer = _Restructurer([])
|
|
|
|
report = prepare_workspace_evidence(tmp_path, restructurer=restructurer, git_status=lambda _: ())
|
|
|
|
assert report.model_calls == 0
|
|
assert report.orphaned == ("evidence:fascia-pediatrica",)
|
|
assert (tmp_path / "evidence" / "curated" / "domain" / "fascia-pediatrica.md").is_file()
|
|
assert [finding.code for finding in report.findings] == ["orphaned_unit"]
|
|
|
|
|
|
def test_prepare_preserves_ids_without_orphans_for_a_unique_source_rename(tmp_path):
|
|
source_text = "I pazienti sotto i 18 anni sono pediatrici."
|
|
_write_workspace(tmp_path, _evidence(source_text), source_text)
|
|
source_root = tmp_path / "evidence" / "source" / "domain"
|
|
(source_root / "patient.md").rename(source_root / "patient-age.md")
|
|
restructurer = _Restructurer([])
|
|
|
|
report = prepare_workspace_evidence(tmp_path, restructurer=restructurer, git_status=lambda _: ())
|
|
|
|
assert report.model_calls == 0
|
|
assert report.orphaned == ()
|
|
assert report.unchanged == ("source/domain/patient-age.md",)
|
|
assert validate_workspace_evidence(tmp_path).publishable is True
|
|
|
|
|
|
def test_prepare_rejects_an_unknown_existing_id_without_writing_the_batch(tmp_path):
|
|
source_text = "I pazienti sotto i 18 anni sono pediatrici."
|
|
_write_workspace(tmp_path, _evidence(source_text), source_text)
|
|
source_path = tmp_path / "evidence" / "source" / "domain" / "patient.md"
|
|
source_path.write_text(source_text + " Revisione.\n", encoding="utf-8")
|
|
original = (tmp_path / "evidence" / "manifest.yaml").read_text(encoding="utf-8")
|
|
restructurer = _Restructurer([_candidate(existing_id="evidence:not-supplied")])
|
|
|
|
with pytest.raises(EvidencePreparationError, match="unknown_existing_id"):
|
|
prepare_workspace_evidence(tmp_path, restructurer=restructurer, git_status=lambda _: ())
|
|
|
|
assert (tmp_path / "evidence" / "manifest.yaml").read_text(encoding="utf-8") == original
|
|
|
|
|
|
def test_prepare_marks_an_omitted_prior_unit_for_human_review(tmp_path):
|
|
source_text = "I pazienti sotto i 18 anni sono pediatrici."
|
|
_write_workspace(tmp_path, _evidence(source_text), source_text)
|
|
(tmp_path / "evidence" / "source" / "domain" / "patient.md").write_text(
|
|
"La classificazione pediatrica non è documentata.\n", encoding="utf-8",
|
|
)
|
|
restructurer = _Restructurer([])
|
|
|
|
report = prepare_workspace_evidence(tmp_path, restructurer=restructurer, git_status=lambda _: ())
|
|
|
|
assert report.model_calls == 1
|
|
assert [finding.code for finding in report.findings] == [
|
|
"supporting_excerpt_missing", "unresolved_review_item",
|
|
]
|
|
retained = load_curated_tree(tmp_path / "evidence" / "curated")[0]
|
|
assert retained.schema_version == 4
|
|
assert retained.review_items[0].code == "source_no_longer_supports_unit"
|
|
|
|
|
|
def test_resolve_retires_an_orphan_and_updates_curated_content_and_manifest_atomically(tmp_path):
|
|
from tht.evidence import resolve_workspace_evidence
|
|
|
|
source_text = "I pazienti sotto i 18 anni sono pediatrici."
|
|
_write_workspace(tmp_path, _evidence(source_text), source_text)
|
|
source_path = tmp_path / "evidence" / "source" / "domain" / "patient.md"
|
|
source_path.unlink()
|
|
prepare_workspace_evidence(tmp_path, restructurer=_Restructurer([]), git_status=lambda _: ())
|
|
|
|
report = resolve_workspace_evidence(
|
|
tmp_path,
|
|
"evidence:fascia-pediatrica",
|
|
retire=True,
|
|
git_status=lambda _: (),
|
|
)
|
|
|
|
assert report.action == "retired"
|
|
assert report.evidence_id == "evidence:fascia-pediatrica"
|
|
assert not (tmp_path / "evidence" / "curated" / "domain" / "fascia-pediatrica.md").exists()
|
|
manifest = load_manifest(tmp_path / "evidence" / "manifest.yaml")
|
|
assert manifest.orphans == ()
|
|
assert manifest.sources == {}
|
|
|
|
|
|
def test_resolve_relinks_an_unit_and_updates_its_manifest_membership(tmp_path):
|
|
from tht.evidence import resolve_workspace_evidence
|
|
|
|
source_text = "I pazienti sotto i 18 anni sono pediatrici."
|
|
_write_workspace(tmp_path, _evidence(source_text), source_text)
|
|
next_source = tmp_path / "evidence" / "source" / "domain" / "patient-age.md"
|
|
next_source.write_text(source_text, encoding="utf-8")
|
|
|
|
report = resolve_workspace_evidence(
|
|
tmp_path,
|
|
"evidence:fascia-pediatrica",
|
|
source="source/domain/patient-age.md",
|
|
git_status=lambda _: (),
|
|
)
|
|
|
|
assert report.action == "relinked"
|
|
assert report.source_file == "source/domain/patient-age.md"
|
|
document = load_curated_tree(tmp_path / "evidence" / "curated")[0]
|
|
assert document.provenance.source_file == "source/domain/patient-age.md"
|
|
manifest = load_manifest(tmp_path / "evidence" / "manifest.yaml")
|
|
assert manifest.sources["source/domain/patient.md"].units == ()
|
|
assert manifest.sources["source/domain/patient-age.md"].units == ("evidence:fascia-pediatrica",)
|
|
assert validate_workspace_evidence(tmp_path).publishable is True
|
|
|
|
|
|
def test_resolve_relink_clears_an_unsupported_review_item_when_the_new_source_supports_it(tmp_path):
|
|
from tht.evidence import resolve_workspace_evidence
|
|
|
|
source_text = "I pazienti sotto i 18 anni sono pediatrici."
|
|
_write_workspace(tmp_path, _evidence(source_text), source_text)
|
|
source_path = tmp_path / "evidence" / "source" / "domain" / "patient.md"
|
|
source_path.write_text("La classificazione pediatrica non è documentata.\n", encoding="utf-8")
|
|
prepare_workspace_evidence(tmp_path, restructurer=_Restructurer([]), git_status=lambda _: ())
|
|
replacement_source = tmp_path / "evidence" / "source" / "domain" / "patient-revised.md"
|
|
replacement_source.write_text(source_text, encoding="utf-8")
|
|
|
|
report = resolve_workspace_evidence(
|
|
tmp_path,
|
|
"evidence:fascia-pediatrica",
|
|
source="source/domain/patient-revised.md",
|
|
git_status=lambda _: (),
|
|
)
|
|
|
|
assert report.findings == ()
|
|
document = load_curated_tree(tmp_path / "evidence" / "curated")[0]
|
|
assert all(item.code != "source_no_longer_supports_unit" for item in document.review_items)
|
|
assert validate_workspace_evidence(tmp_path).publishable is True
|
|
|
|
|
|
def test_resolve_refuses_an_unsafe_relink_without_writing(tmp_path):
|
|
from tht.evidence import EvidencePreparationError, resolve_workspace_evidence
|
|
|
|
source_text = "I pazienti sotto i 18 anni sono pediatrici."
|
|
_write_workspace(tmp_path, _evidence(source_text), source_text)
|
|
curated_path = tmp_path / "evidence" / "curated" / "domain" / "fascia-pediatrica.md"
|
|
manifest_path = tmp_path / "evidence" / "manifest.yaml"
|
|
original_curated = curated_path.read_text(encoding="utf-8")
|
|
original_manifest = manifest_path.read_text(encoding="utf-8")
|
|
|
|
with pytest.raises(EvidencePreparationError, match="source_invalid"):
|
|
resolve_workspace_evidence(
|
|
tmp_path,
|
|
"evidence:fascia-pediatrica",
|
|
source="../outside.md",
|
|
git_status=lambda _: (),
|
|
)
|
|
|
|
assert curated_path.read_text(encoding="utf-8") == original_curated
|
|
assert manifest_path.read_text(encoding="utf-8") == original_manifest
|
|
|
|
|
|
def test_resolve_refuses_a_relink_through_a_symlinked_source_directory(tmp_path):
|
|
from tht.evidence import EvidencePreparationError, resolve_workspace_evidence
|
|
|
|
source_text = "I pazienti sotto i 18 anni sono pediatrici."
|
|
_write_workspace(tmp_path, _evidence(source_text), source_text)
|
|
curated_path = tmp_path / "evidence" / "curated" / "domain" / "fascia-pediatrica.md"
|
|
manifest_path = tmp_path / "evidence" / "manifest.yaml"
|
|
original_curated = curated_path.read_text(encoding="utf-8")
|
|
original_manifest = manifest_path.read_text(encoding="utf-8")
|
|
source_root = tmp_path / "evidence" / "source"
|
|
outside_source = tmp_path / "outside-source"
|
|
source_root.rename(outside_source)
|
|
source_root.symlink_to(outside_source, target_is_directory=True)
|
|
|
|
with pytest.raises(EvidencePreparationError, match="source_invalid"):
|
|
resolve_workspace_evidence(
|
|
tmp_path,
|
|
"evidence:fascia-pediatrica",
|
|
source="source/domain/patient.md",
|
|
git_status=lambda _: (),
|
|
)
|
|
|
|
assert curated_path.read_text(encoding="utf-8") == original_curated
|
|
assert manifest_path.read_text(encoding="utf-8") == original_manifest
|
|
|
|
|
|
def test_resolve_rejects_any_dirty_worktree_without_writing(tmp_path):
|
|
from tht.evidence import EvidencePreparationError, resolve_workspace_evidence
|
|
|
|
source_text = "I pazienti sotto i 18 anni sono pediatrici."
|
|
_write_workspace(tmp_path, _evidence(source_text), source_text)
|
|
curated_path = tmp_path / "evidence" / "curated" / "domain" / "fascia-pediatrica.md"
|
|
manifest_path = tmp_path / "evidence" / "manifest.yaml"
|
|
original_curated = curated_path.read_text(encoding="utf-8")
|
|
original_manifest = manifest_path.read_text(encoding="utf-8")
|
|
|
|
with pytest.raises(EvidencePreparationError, match="worktree_dirty"):
|
|
resolve_workspace_evidence(
|
|
tmp_path,
|
|
"evidence:fascia-pediatrica",
|
|
retire=True,
|
|
git_status=lambda _: (" M README.md",),
|
|
)
|
|
|
|
assert curated_path.read_text(encoding="utf-8") == original_curated
|
|
assert manifest_path.read_text(encoding="utf-8") == original_manifest
|