test: lift post_summary.py coverage 89% to 91%

tests/test_chunking.py

@@ -1,5 +1,7 @@
 from __future__ import annotations
 
+import pytest
+
 from lineageweave.chunking import (
     ConversationTurn,
     chunk_by_conversation_turn,
@@ -624,3 +626,59 @@ def test_chunk_by_source_body_maps_plain_caret_quantities() -> None:
     chunks = chunk_by_source_body("Reserve 12 m^3 and 10^{-3} M stock.")
 
     assert chunks[0].text == "Reserve 12 m³ and 10⁻³ M stock."
+
+
+@pytest.mark.parametrize(
+    ("value", "expected"),
+    [
+        ("1px", 0),  # rounds below one eight-pixel unit
+        ("16px", 2),
+        ("1em", 2),
+        ("0", 0),
+        ("-4px", 0),
+        ("10pt", 2),
+        ("garbage", 0),
+        ("", 0),
+    ],
+)
+def test_length_to_indent_units_clamps_rounds_and_rejects(value: str, expected: int) -> None:
+    from lineageweave.chunking import _length_to_indent_units
+
+    assert _length_to_indent_units(value) == expected
+
+
+@pytest.mark.parametrize(
+    ("raw", "expected"),
+    [
+        ("10px", "10px"),
+        ("10px 20px", "20px"),
+        ("10px 20px 30px", "20px"),
+        ("10px 20px 30px 40px", "40px"),
+        ("", ""),
+    ],
+)
+def test_shorthand_left_value_picks_the_box_model_slot(raw: str, expected: str) -> None:
+    from lineageweave.chunking import _shorthand_left_value
+
+    assert _shorthand_left_value(raw) == expected
+
+
+def test_chunk_by_sentence_returns_empty_for_no_sentences() -> None:
+    from lineageweave.chunking import chunk_by_sentence
+
+    assert chunk_by_sentence("   ") == []
+
+
+def test_decode_data_uri_image_accepts_png_and_rejects_malformed() -> None:
+    import base64
+
+    from lineageweave.chunking import _decode_data_uri_image
+
+    png = base64.b64encode(b"\x89PNG\r\n\x1a\n").decode("ascii")
+    mime, raw = _decode_data_uri_image(f"data:image/png;base64,{png}")
+    assert mime == "image/png"
+    assert raw == b"\x89PNG\r\n\x1a\n"
+
+    assert _decode_data_uri_image("http://example.test/image.png") is None
+    assert _decode_data_uri_image("data:image/png,notbase64") is None
+    assert _decode_data_uri_image("data:image/png;base64,%%%bad") is None

tests/test_post_summary_parse.py

@@ -225,4 +225,207 @@ def test_parse_single_part_pipe_string_is_dropped() -> None:
     content = _json_content({"roles": ["A"], "projects": ["A"]})
     roles, projects = _parse_summary_details(content)
     assert roles == ()
-    assert projects == ()
+    assert projects == ()
+
+def test_project_candidate_node_id_requires_normalized_key() -> None:
+    """A raw project label must be normalized before node construction."""
+    from lineageweave.post_summary import (
+        normalize_project_key,
+        project_candidate_node_id,
+    )
+
+    assert project_candidate_node_id(
+        "11111111-1111-1111-1111-111111111111", "hvdc-pilot"
+    ) == "11111111-1111-1111-1111-111111111111/hvdc-pilot"
+    assert normalize_project_key(" HVDC Pilot ") == "hvdc-pilot"
+    with pytest.raises(ValueError, match="already be normalized"):
+        project_candidate_node_id("11111111-1111-1111-1111-111111111111", "HVDC Pilot")
+
+
+def test_parse_project_candidate_node_id_rejects_bad_separators() -> None:
+    """A node id must contain exactly one post/key separator and be canonical."""
+    from lineageweave.post_summary import parse_project_candidate_node_id
+
+    with pytest.raises(ValueError, match="one post/key separator"):
+        parse_project_candidate_node_id("no-separator-here")
+    with pytest.raises(ValueError, match="one post/key separator"):
+        parse_project_candidate_node_id("post/key/extra")
+    with pytest.raises(ValueError, match="already be normalized|not canonical"):
+        parse_project_candidate_node_id(
+            "11111111-1111-1111-1111-111111111111/UNCANONICAL"
+        )
+    assert parse_project_candidate_node_id(
+        "11111111-1111-1111-1111-111111111111/hvdc-pilot"
+    ) == ("11111111-1111-1111-1111-111111111111", "hvdc-pilot")
+
+
+def test_major_event_action_rejects_missing_text() -> None:
+    """Action and evidence text are both required."""
+    from lineageweave.post_summary import MajorEventAction
+
+    base = dict(requester_actor_name=None, processor_actor_name=None)
+    with pytest.raises(ValueError, match="action and evidence"):
+        MajorEventAction(**base, action_text="  ", evidence_text="evidence")
+    with pytest.raises(ValueError, match="action and evidence"):
+        MajorEventAction(**base, action_text="action", evidence_text="  ")
+    assert MajorEventAction(
+        **base, action_text="action", evidence_text="evidence"
+    ).evidence_text == "evidence"
+
+
+def test_five_w1h_evidence_rejects_unknown_slot_or_missing_text() -> None:
+    """5W1H evidence requires a governed slot plus value and support text."""
+    from lineageweave.post_summary import FiveW1HEvidence
+
+    with pytest.raises(ValueError, match="unsupported 5W1H evidence slot"):
+        FiveW1HEvidence(slot_code="when-not-a-slot", value_text="v", evidence_text="e")
+    with pytest.raises(ValueError, match="requires a value and supporting text"):
+        FiveW1HEvidence(slot_code="when", value_text=" ", evidence_text="e")
+    with pytest.raises(ValueError, match="requires a value and supporting text"):
+        FiveW1HEvidence(slot_code="where", value_text="v", evidence_text="")
+
+    assert (
+        FiveW1HEvidence(slot_code="when", value_text="3월 4일", evidence_text="회의").slot_code
+        == "when"
+    )
+
+
+def test_project_mention_rejects_blank_names_or_out_of_range_confidence() -> None:
+    """Every project mention requires names and evidence plus bounded confidence."""
+    from lineageweave.post_summary import ProjectMention
+
+    with pytest.raises(ValueError, match="require names and evidence"):
+        ProjectMention(project_name="  ", canonical_name="c", evidence="e", confidence=0.5)
+    with pytest.raises(ValueError, match="require names and evidence"):
+        ProjectMention(project_name="p", canonical_name="c", evidence="  ", confidence=0.5)
+    with pytest.raises(ValueError, match="between 0 and 1"):
+        ProjectMention(project_name="p", canonical_name="c", evidence="e", confidence=2.0)
+
+
+def test_key_event_rejects_blank_text_or_explicitly_empty_project_key() -> None:
+    """Key events require text and reject an empty project key."""
+    from lineageweave.post_summary import KeyEvent
+
+    with pytest.raises(ValueError, match="require event text"):
+        KeyEvent(event_text="  ")
+    with pytest.raises(ValueError, match="must be non-empty"):
+        KeyEvent(event_text="event", project_key="   ")
+
+
+def test_hallucinated_account_name_detects_a_matching_context_hint() -> None:
+    """A context hint naming the account selects it as a hallucination guard."""
+    from lineageweave.post_summary import _hallucinated_account_name
+
+    assert (
+        _hallucinated_account_name(
+            "author_account_name=Demo Analyst [source_field=user_account.display_name]"
+        )
+        == "Demo Analyst"
+    )
+    assert _hallucinated_account_name("") is None
+    assert _hallucinated_account_name("no account hint here") is None
+
+
+def test_plain_details_accepts_three_column_role_rows() -> None:
+    """A 3-column role row defaults the actor type to person."""
+    from lineageweave.post_summary import _parse_plain_summary_details
+
+    details = _parse_plain_summary_details(
+        "ROLES:\n홍길동 | 자료 검토 | 당사\n"
+        "PROJECTS:\nNONE\n"
+        "EVIDENCE:\nNONE",
+        context_hints="author_account_name=Demo Analyst [source_field=user_account.display_name]",
+    )
+    assert details is not None
+    assert details[0][0].actor_type_code == "prov_person"
+    assert details[0][0].affiliated_organization_name == "당사"
+
+
+def test_plain_details_drops_template_echo_rows_and_hallucinated_actor() -> None:
+    """Prompt-template echoes and the logged-in account are never actors."""
+    from lineageweave.post_summary import _parse_plain_summary_details
+
+    details = _parse_plain_summary_details(
+        "ROLES:\n"
+        "actor name | responsibility | person, organization, or team | affiliation or none\n"
+        "Demo Analyst | 고객 면담 | person | Demo Corp\n"
+        "Jordi Gil | 견적 승인 | person | Northwind Labs\n"
+        "PROJECTS:\nNONE",
+        context_hints="author_account_name=Demo Analyst [source_field=user_account.display_name]; "
+        "author_affiliations=Demo Corp [source_field=account_affiliation.corporate_entity_id]",
+    )
+    assert details is not None
+    assert [role.actor_name for role in details[0]] == ["Jordi Gil"]
+
+
+def test_plain_details_skips_role_rows_with_unknown_actor_column() -> None:
+    """A row whose actor column isn't person/organization/team is dropped."""
+    from lineageweave.post_summary import _parse_plain_summary_details
+
+    details = _parse_plain_summary_details(
+        "ROLES:\n"
+        "Q&A participant | 발표 듣기 | some-description | Acme\n"
+        "PROJECTS:\nNONE"
+    )
+    assert details is not None
+    assert details[0] == ()
+
+
+def test_plain_details_project_uses_post_title_as_evidence_when_empty() -> None:
+    """A project whose evidence column is empty falls back to the post title."""
+    from lineageweave.post_summary import _parse_plain_summary_details
+
+    details = _parse_plain_summary_details(
+        "ROLES:\nNONE\n"
+        "PROJECTS:\nHVDC Pilot | hvdc-pilot | none | 0.9",
+        post_title="HVDC Pilot 견적 검토",
+    )
+    assert details is not None
+    assert details[1][0].evidence == "HVDC Pilot 견적 검토"
+
+
+def test_plain_details_drops_untitled_and_unparsable_project_rows() -> None:
+    """Projects whose empty evidence cannot borrow a title are dropped."""
+    from lineageweave.post_summary import _parse_plain_summary_details
+
+    details = _parse_plain_summary_details(
+        "ROLES:\nNONE\n"
+        "PROJECTS:\nUnknown Project | unknown-project | NONE | not-a-number\n"
+        "Mystery | mystery | NONE | 0.5",
+        post_title="",
+    )
+    assert details is not None
+    assert details[1] == ()
+
+
+def test_plain_details_five_column_actions_recognize_actors_and_fallback() -> None:
+    """5-column actions use actor columns; unrecognized rows fall back to legacy."""
+    from lineageweave.post_summary import _parse_plain_summary_details
+
+    details = _parse_plain_summary_details(
+        "ROLES:\n"
+        "홍길동 | 변경 요청 | person | 당사\n"
+        "김철수 | 도면 수정 | person | 고객사\n"
+        "PROJECTS:\nNONE\n"
+        "ACTIONS:\n"
+        "도면 변경 승인 | hvdc-pilot | 홍길동 | 김철수 | 근거 문장\n"
+        "드롭 될 행 | x | NotAnActor | NothingInteresting | bad"
+    )
+    assert details is not None
+    assert [action.action_text for action in details[2]] == [
+        "도면 변경 승인",
+        "드롭 될 행",
+    ]
+    assert details[2][0].project_key == "hvdc-pilot"
+    assert details[2][0].requester_actor_name == "홍길동"
+    assert details[2][0].processor_actor_name == "김철수"
+    assert details[2][1].project_key is None
+
+
+def test_plain_details_requires_roles_and_projects_sections() -> None:
+    """Missing ROLES/PROJECTS sections fail the whole parse."""
+    from lineageweave.post_summary import _parse_plain_summary_details
+
+    assert _parse_plain_summary_details("ROLES:\nNONE") is None
+    assert _parse_plain_summary_details("EVIDENCE:\nwhere | x | y") is None
+    assert _parse_plain_summary_details("") is None