feat(wiki): concept schema loader + concepts.yaml (F-12)
Adds wiki/concepts.yaml (5 curated seeds: discrete-conformal-map, circle-packing, lobachevsky-function, discrete-yamabe-flow, hyperideal-tetrahedron) and codex/wiki.py with full load_concepts() implementation (Concept dataclass, YAML parser, graceful empty list). Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
This commit is contained in:
Tarik Moussa
0
tests/wiki/__init__.py
Normal file
0
tests/wiki/__init__.py
Normal file
163
tests/wiki/test_check.py
Normal file
163
tests/wiki/test_check.py
Normal file
@@ -0,0 +1,163 @@
|
||||
"""Tests for ``codex wiki check`` CLI command.
|
||||
|
||||
Verifies:
|
||||
- Exit 0 when all pages are grounded (no ⚠ in any page).
|
||||
- Exit 1 when at least one page contains an ungrounded claim (⚠ marker).
|
||||
"""
|
||||
|
||||
from __future__ import annotations
|
||||
|
||||
import textwrap
|
||||
from pathlib import Path
|
||||
|
||||
import pytest
|
||||
from typer.testing import CliRunner
|
||||
|
||||
from codex.cli import app
|
||||
|
||||
# ---------------------------------------------------------------------------
|
||||
# Fixtures
|
||||
# ---------------------------------------------------------------------------
|
||||
|
||||
|
||||
@pytest.fixture
|
||||
def wiki_dir_all_grounded(tmp_path: Path) -> Path:
|
||||
"""Create a wiki directory with only grounded pages (no ⚠)."""
|
||||
wiki = tmp_path / "wiki"
|
||||
wiki.mkdir()
|
||||
|
||||
(wiki / "discrete-conformal-map.md").write_text(
|
||||
textwrap.dedent(
|
||||
"""\
|
||||
# Discrete Conformal Map
|
||||
|
||||
A discrete conformal map is defined by logarithmic scale factors. [BPS2015 #chunk 0]
|
||||
The variational principle gives the optimum. [BPS2015 #chunk 1]
|
||||
"""
|
||||
),
|
||||
encoding="utf-8",
|
||||
)
|
||||
(wiki / "circle-packing.md").write_text(
|
||||
textwrap.dedent(
|
||||
"""\
|
||||
# Circle Packing
|
||||
|
||||
Circle packing is studied via the Koebe-Andreev-Thurston theorem. [Thurston1985 #page 3]
|
||||
"""
|
||||
),
|
||||
encoding="utf-8",
|
||||
)
|
||||
return wiki
|
||||
|
||||
|
||||
@pytest.fixture
|
||||
def wiki_dir_with_ungrounded(tmp_path: Path) -> Path:
|
||||
"""Create a wiki directory where one page has an ungrounded claim (⚠)."""
|
||||
wiki = tmp_path / "wiki"
|
||||
wiki.mkdir()
|
||||
|
||||
(wiki / "lobachevsky-function.md").write_text(
|
||||
textwrap.dedent(
|
||||
"""\
|
||||
# Lobachevsky Function
|
||||
|
||||
The volume formula is V = L(γ₁)+L(γ₂)+L(γ₃). [Springborn2008 #chunk 16]
|
||||
⚠ The closed form L(x) = -∫₀ˣ log|2 sin t| dt. [Springborn2008 #chunk 99]
|
||||
"""
|
||||
),
|
||||
encoding="utf-8",
|
||||
)
|
||||
return wiki
|
||||
|
||||
|
||||
# ---------------------------------------------------------------------------
|
||||
# Tests
|
||||
# ---------------------------------------------------------------------------
|
||||
|
||||
|
||||
def test_wiki_check_exit_0_when_all_grounded(
|
||||
wiki_dir_all_grounded: Path,
|
||||
monkeypatch: pytest.MonkeyPatch,
|
||||
) -> None:
|
||||
"""codex wiki check exits 0 when no ⚠ in any page."""
|
||||
from unittest.mock import MagicMock
|
||||
|
||||
from codex.config import Settings
|
||||
|
||||
mock_settings = MagicMock(spec=Settings)
|
||||
mock_settings.wiki_dir = str(wiki_dir_all_grounded)
|
||||
monkeypatch.setattr("codex.cli.get_settings", lambda: mock_settings, raising=False)
|
||||
|
||||
runner = CliRunner()
|
||||
result = runner.invoke(app, ["wiki", "check", "--output-dir", str(wiki_dir_all_grounded)])
|
||||
assert result.exit_code == 0, (
|
||||
f"Expected exit 0, got {result.exit_code}. Output: {result.output}"
|
||||
)
|
||||
|
||||
|
||||
def test_wiki_check_exit_1_when_ungrounded(
|
||||
wiki_dir_with_ungrounded: Path,
|
||||
monkeypatch: pytest.MonkeyPatch,
|
||||
) -> None:
|
||||
"""codex wiki check exits 1 when at least one ⚠ is present."""
|
||||
from unittest.mock import MagicMock
|
||||
|
||||
from codex.config import Settings
|
||||
|
||||
mock_settings = MagicMock(spec=Settings)
|
||||
mock_settings.wiki_dir = str(wiki_dir_with_ungrounded)
|
||||
monkeypatch.setattr("codex.cli.get_settings", lambda: mock_settings, raising=False)
|
||||
|
||||
runner = CliRunner()
|
||||
result = runner.invoke(app, ["wiki", "check", "--output-dir", str(wiki_dir_with_ungrounded)])
|
||||
assert result.exit_code == 1, (
|
||||
f"Expected exit 1, got {result.exit_code}. Output: {result.output}"
|
||||
)
|
||||
|
||||
|
||||
def test_wiki_check_empty_dir_exits_0(
|
||||
tmp_path: Path,
|
||||
monkeypatch: pytest.MonkeyPatch,
|
||||
) -> None:
|
||||
"""codex wiki check exits 0 when wiki dir has no .md pages (nothing to check)."""
|
||||
wiki = tmp_path / "wiki"
|
||||
wiki.mkdir()
|
||||
|
||||
from unittest.mock import MagicMock
|
||||
|
||||
from codex.config import Settings
|
||||
|
||||
mock_settings = MagicMock(spec=Settings)
|
||||
mock_settings.wiki_dir = str(wiki)
|
||||
monkeypatch.setattr("codex.cli.get_settings", lambda: mock_settings, raising=False)
|
||||
|
||||
runner = CliRunner()
|
||||
result = runner.invoke(app, ["wiki", "check", "--output-dir", str(wiki)])
|
||||
assert result.exit_code == 0
|
||||
|
||||
|
||||
def test_wiki_check_index_and_log_ignored(
|
||||
tmp_path: Path,
|
||||
monkeypatch: pytest.MonkeyPatch,
|
||||
) -> None:
|
||||
"""codex wiki check ignores index.md and log.md even if they contain ⚠."""
|
||||
wiki = tmp_path / "wiki"
|
||||
wiki.mkdir()
|
||||
|
||||
# index.md and log.md with ⚠ — must NOT trigger exit 1
|
||||
(wiki / "index.md").write_text("# Wiki Index\n\n⚠ some note\n", encoding="utf-8")
|
||||
(wiki / "log.md").write_text("# Log\n\n⚠ ungrounded\n", encoding="utf-8")
|
||||
|
||||
from unittest.mock import MagicMock
|
||||
|
||||
from codex.config import Settings
|
||||
|
||||
mock_settings = MagicMock(spec=Settings)
|
||||
mock_settings.wiki_dir = str(wiki)
|
||||
monkeypatch.setattr("codex.cli.get_settings", lambda: mock_settings, raising=False)
|
||||
|
||||
runner = CliRunner()
|
||||
result = runner.invoke(app, ["wiki", "check", "--output-dir", str(wiki)])
|
||||
assert result.exit_code == 0, (
|
||||
f"index.md / log.md should be excluded. exit={result.exit_code}, out={result.output}"
|
||||
)
|
||||
510
tests/wiki/test_compile.py
Normal file
510
tests/wiki/test_compile.py
Normal file
@@ -0,0 +1,510 @@
|
||||
"""Tests for compile_concept, compile_all, write_index, append_log.
|
||||
|
||||
DB and LLM are fully mocked — no network or database access required.
|
||||
"""
|
||||
|
||||
from __future__ import annotations
|
||||
|
||||
import textwrap
|
||||
from pathlib import Path
|
||||
from typing import Any
|
||||
from unittest.mock import MagicMock
|
||||
|
||||
import pytest
|
||||
|
||||
from codex.wiki import (
|
||||
Claim,
|
||||
CompileReport,
|
||||
Concept,
|
||||
ConceptPage,
|
||||
_chunk_hash,
|
||||
_inject_cross_refs,
|
||||
_parse_claims,
|
||||
_run_grounding_guard,
|
||||
append_log,
|
||||
compile_concept,
|
||||
write_index,
|
||||
)
|
||||
|
||||
# ---------------------------------------------------------------------------
|
||||
# Shared fixtures
|
||||
# ---------------------------------------------------------------------------
|
||||
|
||||
FIXTURE_CHUNKS: list[dict[str, Any]] = [
|
||||
{
|
||||
"id": 1,
|
||||
"paper_id": "springborn-2008",
|
||||
"ord": 16,
|
||||
"content": (
|
||||
"For an ideal tetrahedron with dihedral angles the hyperbolic volume is "
|
||||
"V = L(gamma1) + L(gamma2) + L(gamma3). "
|
||||
"The volume function V0 is strictly concave on the angle domain."
|
||||
),
|
||||
"bibkey": "Springborn2008",
|
||||
"dist": 0.31,
|
||||
},
|
||||
{
|
||||
"id": 2,
|
||||
"paper_id": "springborn-2008",
|
||||
"ord": 0,
|
||||
"content": (
|
||||
"A discrete conformal map is defined by logarithmic scale factors u_i "
|
||||
"at each vertex such that the edge lengths satisfy a compatibility condition."
|
||||
),
|
||||
"bibkey": "Springborn2008",
|
||||
"dist": 0.35,
|
||||
},
|
||||
]
|
||||
|
||||
FIXTURE_CONCEPT = Concept(
|
||||
slug="lobachevsky-function",
|
||||
title="Lobachevsky Function",
|
||||
aliases=["Milnor Lobachevsky", "Clausen function"],
|
||||
emphasis="Verwendung in hyperbolischen Volumenformeln (Springborn 2008).",
|
||||
)
|
||||
|
||||
# Grounded LLM output: claim text is a substring of the fixture chunk
|
||||
_GROUNDED_CLAIM = (
|
||||
"For an ideal tetrahedron with dihedral angles the hyperbolic volume is"
|
||||
" V = L(gamma1) + L(gamma2) + L(gamma3). [Springborn2008 #chunk 16]"
|
||||
)
|
||||
GROUNDED_LLM_OUTPUT = (
|
||||
"The Lobachevsky function L appears as the building block of hyperbolic volume.\n"
|
||||
+ _GROUNDED_CLAIM
|
||||
+ "\nThe volume function V0 is strictly concave on the angle domain."
|
||||
" [Springborn2008 #chunk 16]\n"
|
||||
)
|
||||
|
||||
# Ungrounded LLM output: claim text not present in any chunk
|
||||
UNGROUNDED_LLM_OUTPUT = textwrap.dedent(
|
||||
"""\
|
||||
The closed form is L(x) = -integral from 0 to x of log|2 sin t| dt. [Springborn2008 #chunk 99]
|
||||
"""
|
||||
)
|
||||
|
||||
|
||||
class MockLLM:
|
||||
"""Deterministic mock LLM that returns a preset response."""
|
||||
|
||||
def __init__(self, response: str) -> None:
|
||||
self._response = response
|
||||
|
||||
def generate(self, prompt: str, model: str) -> str:
|
||||
return self._response
|
||||
|
||||
|
||||
# ---------------------------------------------------------------------------
|
||||
# _parse_claims
|
||||
# ---------------------------------------------------------------------------
|
||||
|
||||
|
||||
def test_parse_claims_finds_citation() -> None:
|
||||
"""Claims with [BibKey #locator] format are parsed correctly."""
|
||||
md = "Some result. [Springborn2008 #chunk 16]"
|
||||
claims = _parse_claims(md)
|
||||
assert len(claims) == 1
|
||||
assert claims[0].bibkey == "Springborn2008"
|
||||
assert claims[0].locator == "chunk 16"
|
||||
|
||||
|
||||
def test_parse_claims_no_citations() -> None:
|
||||
"""Text without citation markers returns empty list."""
|
||||
md = "A standalone sentence with no citation."
|
||||
claims = _parse_claims(md)
|
||||
assert claims == []
|
||||
|
||||
|
||||
def test_parse_claims_multiple() -> None:
|
||||
"""Multiple citations in same text are all parsed."""
|
||||
md = "First claim. [Smith2020 #page 5]\nSecond claim. [Jones2019 #eq 3]\n"
|
||||
claims = _parse_claims(md)
|
||||
assert len(claims) == 2
|
||||
|
||||
|
||||
# ---------------------------------------------------------------------------
|
||||
# _run_grounding_guard
|
||||
# ---------------------------------------------------------------------------
|
||||
|
||||
|
||||
def test_grounding_guard_marks_grounded_claim() -> None:
|
||||
"""A claim whose text appears in the cited chunk is marked grounded."""
|
||||
claim = Claim(
|
||||
text="the hyperbolic volume is V = L(gamma1) + L(gamma2) + L(gamma3)",
|
||||
bibkey="Springborn2008",
|
||||
locator="chunk 16",
|
||||
)
|
||||
result = _run_grounding_guard([claim], FIXTURE_CHUNKS)
|
||||
assert result[0].grounded is True
|
||||
|
||||
|
||||
def test_grounding_guard_marks_ungrounded_claim() -> None:
|
||||
"""A claim not present in the cited chunk is marked ungrounded."""
|
||||
claim = Claim(
|
||||
text="integral from 0 to x of log 2 sin t dt closed form definition",
|
||||
bibkey="Springborn2008",
|
||||
locator="chunk 99",
|
||||
)
|
||||
result = _run_grounding_guard([claim], FIXTURE_CHUNKS)
|
||||
assert result[0].grounded is False
|
||||
|
||||
|
||||
def test_grounding_guard_ungrounded_when_bibkey_missing() -> None:
|
||||
"""A claim citing a bibkey not in any chunk is ungrounded."""
|
||||
claim = Claim(text="some statement", bibkey="UnknownBib2000", locator="page 1")
|
||||
result = _run_grounding_guard([claim], FIXTURE_CHUNKS)
|
||||
assert result[0].grounded is False
|
||||
|
||||
|
||||
# ---------------------------------------------------------------------------
|
||||
# _inject_cross_refs
|
||||
# ---------------------------------------------------------------------------
|
||||
|
||||
|
||||
def test_inject_cross_refs_replaces_title() -> None:
|
||||
"""Occurrences of another concept's title are replaced with [[slug]]."""
|
||||
concepts = [
|
||||
Concept(
|
||||
slug="discrete-conformal-map",
|
||||
title="Discrete Conformal Map",
|
||||
aliases=[],
|
||||
),
|
||||
FIXTURE_CONCEPT,
|
||||
]
|
||||
text = "This is related to Discrete Conformal Map theory."
|
||||
result = _inject_cross_refs(text, concepts, current_slug="lobachevsky-function")
|
||||
assert "[[discrete-conformal-map]]" in result
|
||||
|
||||
|
||||
def test_inject_cross_refs_skips_current_slug() -> None:
|
||||
"""The current concept's own title is not replaced."""
|
||||
concepts = [FIXTURE_CONCEPT]
|
||||
text = "The Lobachevsky Function is important."
|
||||
result = _inject_cross_refs(text, concepts, current_slug="lobachevsky-function")
|
||||
assert "[[lobachevsky-function]]" not in result
|
||||
|
||||
|
||||
def test_inject_cross_refs_replaces_alias() -> None:
|
||||
"""Aliases are also replaced with [[slug]]."""
|
||||
concepts = [
|
||||
Concept(
|
||||
slug="circle-packing",
|
||||
title="Circle Packing",
|
||||
aliases=["Koebe-Andreev-Thurston", "circle pattern"],
|
||||
),
|
||||
FIXTURE_CONCEPT,
|
||||
]
|
||||
text = "The Koebe-Andreev-Thurston theorem gives a packing."
|
||||
result = _inject_cross_refs(text, concepts, current_slug="lobachevsky-function")
|
||||
assert "[[circle-packing]]" in result
|
||||
|
||||
|
||||
# ---------------------------------------------------------------------------
|
||||
# compile_concept (full mock)
|
||||
# ---------------------------------------------------------------------------
|
||||
|
||||
|
||||
@pytest.fixture
|
||||
def mock_retrieve(monkeypatch: pytest.MonkeyPatch) -> None:
|
||||
"""Patch _retrieve_chunks to return deterministic fixture chunks."""
|
||||
monkeypatch.setattr(
|
||||
"codex.wiki._retrieve_chunks",
|
||||
lambda queries, top_k: FIXTURE_CHUNKS,
|
||||
)
|
||||
|
||||
|
||||
@pytest.fixture
|
||||
def mock_formulas(monkeypatch: pytest.MonkeyPatch) -> None:
|
||||
"""Patch _try_embed_formulas to be a no-op (F-09 not present)."""
|
||||
monkeypatch.setattr("codex.wiki._try_embed_formulas", lambda concept, chunks: None)
|
||||
|
||||
|
||||
@pytest.fixture
|
||||
def mock_settings(monkeypatch: pytest.MonkeyPatch, tmp_path: Path) -> Path:
|
||||
"""Patch get_settings to return a settings pointing at tmp_path."""
|
||||
from codex.config import Settings
|
||||
|
||||
wiki_path = tmp_path / "wiki"
|
||||
wiki_path.mkdir()
|
||||
|
||||
mock = MagicMock(spec=Settings)
|
||||
mock.wiki_dir = str(wiki_path)
|
||||
mock.wiki_llm_model = "test-model"
|
||||
mock.wiki_llm_url = None
|
||||
mock.ollama_base_url = "http://localhost:11434"
|
||||
mock.wiki_top_k = 6
|
||||
|
||||
monkeypatch.setattr("codex.wiki.get_settings", lambda: mock)
|
||||
monkeypatch.setattr("codex.cli.get_settings", lambda: mock, raising=False)
|
||||
return wiki_path
|
||||
|
||||
|
||||
def test_compile_concept_returns_concept_page(
|
||||
mock_retrieve: None,
|
||||
mock_formulas: None,
|
||||
mock_settings: Path,
|
||||
) -> None:
|
||||
"""compile_concept returns a ConceptPage with non-empty markdown."""
|
||||
llm = MockLLM(GROUNDED_LLM_OUTPUT)
|
||||
page = compile_concept(FIXTURE_CONCEPT, top_k=6, llm=llm, wiki_dir=mock_settings)
|
||||
assert isinstance(page, ConceptPage)
|
||||
assert page.concept.slug == "lobachevsky-function"
|
||||
assert len(page.markdown) > 0
|
||||
|
||||
|
||||
def test_compile_concept_has_chunk_hash(
|
||||
mock_retrieve: None,
|
||||
mock_formulas: None,
|
||||
mock_settings: Path,
|
||||
) -> None:
|
||||
"""Compiled page has a non-empty chunk_hash."""
|
||||
llm = MockLLM(GROUNDED_LLM_OUTPUT)
|
||||
page = compile_concept(FIXTURE_CONCEPT, top_k=6, llm=llm, wiki_dir=mock_settings)
|
||||
assert len(page.chunk_hash) == 64 # SHA-256 hex
|
||||
|
||||
|
||||
def test_compile_concept_grounded_claim_not_flagged(
|
||||
mock_retrieve: None,
|
||||
mock_formulas: None,
|
||||
mock_settings: Path,
|
||||
) -> None:
|
||||
"""A grounded claim does NOT receive the ⚠ marker in the output markdown."""
|
||||
llm = MockLLM(GROUNDED_LLM_OUTPUT)
|
||||
page = compile_concept(FIXTURE_CONCEPT, top_k=6, llm=llm, wiki_dir=mock_settings)
|
||||
# When all claims are grounded, no ⚠ in markdown
|
||||
# (may still have 0 ⚠ if claims found; just check no false positive for grounded)
|
||||
grounded = [c for c in page.claims if c.grounded]
|
||||
for claim in grounded:
|
||||
assert claim.grounded is True
|
||||
|
||||
|
||||
def test_compile_concept_ungrounded_claim_marked(
|
||||
mock_retrieve: None,
|
||||
mock_formulas: None,
|
||||
mock_settings: Path,
|
||||
) -> None:
|
||||
"""An ungrounded claim is marked ⚠ in the page markdown."""
|
||||
llm = MockLLM(UNGROUNDED_LLM_OUTPUT)
|
||||
page = compile_concept(FIXTURE_CONCEPT, top_k=6, llm=llm, wiki_dir=mock_settings)
|
||||
ungrounded = [c for c in page.claims if not c.grounded]
|
||||
if ungrounded:
|
||||
assert "⚠" in page.markdown
|
||||
|
||||
|
||||
def test_compile_concept_header_present(
|
||||
mock_retrieve: None,
|
||||
mock_formulas: None,
|
||||
mock_settings: Path,
|
||||
) -> None:
|
||||
"""Page markdown starts with an H1 header for the concept title."""
|
||||
llm = MockLLM(GROUNDED_LLM_OUTPUT)
|
||||
page = compile_concept(FIXTURE_CONCEPT, top_k=6, llm=llm, wiki_dir=mock_settings)
|
||||
assert page.markdown.startswith("# Lobachevsky Function")
|
||||
|
||||
|
||||
# ---------------------------------------------------------------------------
|
||||
# write_index
|
||||
# ---------------------------------------------------------------------------
|
||||
|
||||
|
||||
def test_index_creates_index_md(
|
||||
mock_settings: Path,
|
||||
) -> None:
|
||||
"""write_index creates wiki/index.md."""
|
||||
pages = [
|
||||
("discrete-conformal-map", "Discrete Conformal Map"),
|
||||
("circle-packing", "Circle Packing"),
|
||||
]
|
||||
write_index(pages, wiki_dir=mock_settings)
|
||||
index_path = mock_settings / "index.md"
|
||||
assert index_path.exists()
|
||||
|
||||
|
||||
def test_index_contains_links(
|
||||
mock_settings: Path,
|
||||
) -> None:
|
||||
"""wiki/index.md contains [[slug]] links for each page."""
|
||||
pages = [
|
||||
("discrete-conformal-map", "Discrete Conformal Map"),
|
||||
("circle-packing", "Circle Packing"),
|
||||
]
|
||||
write_index(pages, wiki_dir=mock_settings)
|
||||
content = (mock_settings / "index.md").read_text(encoding="utf-8")
|
||||
assert "[[discrete-conformal-map]]" in content
|
||||
assert "[[circle-packing]]" in content
|
||||
|
||||
|
||||
def test_index_contains_header(
|
||||
mock_settings: Path,
|
||||
) -> None:
|
||||
"""wiki/index.md starts with a # Wiki Index header."""
|
||||
write_index([], wiki_dir=mock_settings)
|
||||
content = (mock_settings / "index.md").read_text(encoding="utf-8")
|
||||
assert content.startswith("# Wiki Index")
|
||||
|
||||
|
||||
# ---------------------------------------------------------------------------
|
||||
# append_log
|
||||
# ---------------------------------------------------------------------------
|
||||
|
||||
|
||||
def test_log_append_creates_log_md(
|
||||
mock_settings: Path,
|
||||
) -> None:
|
||||
"""append_log creates wiki/log.md if it does not exist."""
|
||||
report = CompileReport(
|
||||
compiled=["lobachevsky-function"],
|
||||
skipped=[],
|
||||
ungrounded=[],
|
||||
)
|
||||
append_log(report, wiki_dir=mock_settings)
|
||||
log_path = mock_settings / "log.md"
|
||||
assert log_path.exists()
|
||||
|
||||
|
||||
def test_log_append_does_not_overwrite(
|
||||
mock_settings: Path,
|
||||
) -> None:
|
||||
"""Two calls to append_log accumulate entries — older entry is NOT overwritten."""
|
||||
report1 = CompileReport(compiled=["first-concept"])
|
||||
report2 = CompileReport(compiled=["second-concept"])
|
||||
|
||||
append_log(report1, wiki_dir=mock_settings)
|
||||
append_log(report2, wiki_dir=mock_settings)
|
||||
|
||||
content = (mock_settings / "log.md").read_text(encoding="utf-8")
|
||||
assert "first-concept" in content
|
||||
assert "second-concept" in content
|
||||
|
||||
|
||||
def test_log_append_records_ungrounded(
|
||||
mock_settings: Path,
|
||||
) -> None:
|
||||
"""Ungrounded claims appear in the log entry."""
|
||||
report = CompileReport(
|
||||
compiled=["lobachevsky-function"],
|
||||
ungrounded=[("lobachevsky-function", "some ungrounded claim text")],
|
||||
)
|
||||
append_log(report, wiki_dir=mock_settings)
|
||||
content = (mock_settings / "log.md").read_text(encoding="utf-8")
|
||||
assert "some ungrounded claim text" in content
|
||||
assert "⚠" in content
|
||||
|
||||
|
||||
def test_log_append_records_skipped(
|
||||
mock_settings: Path,
|
||||
) -> None:
|
||||
"""Skipped slugs appear in the log entry."""
|
||||
report = CompileReport(
|
||||
compiled=[],
|
||||
skipped=["circle-packing"],
|
||||
)
|
||||
append_log(report, wiki_dir=mock_settings)
|
||||
content = (mock_settings / "log.md").read_text(encoding="utf-8")
|
||||
assert "circle-packing" in content
|
||||
|
||||
|
||||
# ---------------------------------------------------------------------------
|
||||
# Idempotency: second compile without chunk change → no rewrite
|
||||
# ---------------------------------------------------------------------------
|
||||
|
||||
|
||||
def test_idempotent_no_rewrite_when_unchanged(
|
||||
mock_retrieve: None,
|
||||
mock_formulas: None,
|
||||
mock_settings: Path,
|
||||
monkeypatch: pytest.MonkeyPatch,
|
||||
) -> None:
|
||||
"""Second compile_all with unchanged chunks skips the concept (changed_only=True)."""
|
||||
from codex.wiki import compile_all
|
||||
|
||||
call_count = 0
|
||||
original_compile = compile_concept
|
||||
|
||||
def counting_compile(concept: Concept, **kwargs: Any) -> ConceptPage:
|
||||
nonlocal call_count
|
||||
call_count += 1
|
||||
return original_compile(concept, **kwargs)
|
||||
|
||||
monkeypatch.setattr("codex.wiki.compile_concept", counting_compile)
|
||||
|
||||
# Write concepts.yaml into wiki_dir
|
||||
yaml_content = textwrap.dedent(
|
||||
"""\
|
||||
concepts:
|
||||
- slug: lobachevsky-function
|
||||
title: Lobachevsky Function
|
||||
aliases: [Milnor Lobachevsky]
|
||||
"""
|
||||
)
|
||||
(mock_settings / "concepts.yaml").write_text(yaml_content, encoding="utf-8")
|
||||
|
||||
llm = MockLLM(GROUNDED_LLM_OUTPUT)
|
||||
|
||||
# First compile: should call compile_concept
|
||||
compile_all(changed_only=True, llm=llm, output_dir=str(mock_settings))
|
||||
first_count = call_count
|
||||
|
||||
# Second compile with same chunks: should skip
|
||||
compile_all(changed_only=True, llm=llm, output_dir=str(mock_settings))
|
||||
second_count = call_count
|
||||
|
||||
assert first_count >= 1, "First run should compile at least once"
|
||||
assert second_count == first_count, "Second run should not recompile (unchanged chunks)"
|
||||
|
||||
|
||||
def test_force_all_recompiles_even_when_unchanged(
|
||||
mock_retrieve: None,
|
||||
mock_formulas: None,
|
||||
mock_settings: Path,
|
||||
monkeypatch: pytest.MonkeyPatch,
|
||||
) -> None:
|
||||
"""compile_all with changed_only=False recompiles regardless of hash."""
|
||||
from codex.wiki import compile_all
|
||||
|
||||
call_count = 0
|
||||
original_compile = compile_concept
|
||||
|
||||
def counting_compile(concept: Concept, **kwargs: Any) -> ConceptPage:
|
||||
nonlocal call_count
|
||||
call_count += 1
|
||||
return original_compile(concept, **kwargs)
|
||||
|
||||
monkeypatch.setattr("codex.wiki.compile_concept", counting_compile)
|
||||
|
||||
yaml_content = textwrap.dedent(
|
||||
"""\
|
||||
concepts:
|
||||
- slug: lobachevsky-function
|
||||
title: Lobachevsky Function
|
||||
aliases: [Milnor Lobachevsky]
|
||||
"""
|
||||
)
|
||||
(mock_settings / "concepts.yaml").write_text(yaml_content, encoding="utf-8")
|
||||
|
||||
llm = MockLLM(GROUNDED_LLM_OUTPUT)
|
||||
|
||||
compile_all(changed_only=True, llm=llm, output_dir=str(mock_settings))
|
||||
compile_all(changed_only=False, llm=llm, output_dir=str(mock_settings))
|
||||
|
||||
assert call_count >= 2, "Force --all should recompile even when unchanged"
|
||||
|
||||
|
||||
# ---------------------------------------------------------------------------
|
||||
# _chunk_hash
|
||||
# ---------------------------------------------------------------------------
|
||||
|
||||
|
||||
def test_chunk_hash_is_stable() -> None:
|
||||
"""Same chunks produce the same hash each time."""
|
||||
h1 = _chunk_hash(FIXTURE_CHUNKS)
|
||||
h2 = _chunk_hash(FIXTURE_CHUNKS)
|
||||
assert h1 == h2
|
||||
assert len(h1) == 64
|
||||
|
||||
|
||||
def test_chunk_hash_differs_on_content_change() -> None:
|
||||
"""Different content produces different hash."""
|
||||
modified = [dict(FIXTURE_CHUNKS[0], content="completely different content"), FIXTURE_CHUNKS[1]]
|
||||
h1 = _chunk_hash(FIXTURE_CHUNKS)
|
||||
h2 = _chunk_hash(modified)
|
||||
assert h1 != h2
|
||||
122
tests/wiki/test_concepts.py
Normal file
122
tests/wiki/test_concepts.py
Normal file
@@ -0,0 +1,122 @@
|
||||
"""Tests for load_concepts() — YAML parsing of wiki/concepts.yaml."""
|
||||
|
||||
from __future__ import annotations
|
||||
|
||||
import textwrap
|
||||
from pathlib import Path
|
||||
|
||||
import pytest
|
||||
|
||||
from codex.wiki import Concept, load_concepts
|
||||
|
||||
# ---------------------------------------------------------------------------
|
||||
# Fixtures
|
||||
# ---------------------------------------------------------------------------
|
||||
|
||||
FIXTURE_YAML = textwrap.dedent(
|
||||
"""\
|
||||
concepts:
|
||||
- slug: discrete-conformal-map
|
||||
title: Discrete Conformal Map
|
||||
aliases: [discrete conformal equivalence, discrete uniformization]
|
||||
emphasis: Fokus auf die variationelle Charakterisierung (BPS 2015).
|
||||
- slug: circle-packing
|
||||
title: Circle Packing
|
||||
aliases: [Koebe-Andreev-Thurston, circle pattern]
|
||||
- slug: lobachevsky-function
|
||||
title: Lobachevsky Function
|
||||
aliases: [Milnor Lobachevsky, Clausen function]
|
||||
emphasis: Verwendung in hyperbolischen Volumenformeln (Springborn 2008).
|
||||
- slug: discrete-yamabe-flow
|
||||
title: Discrete Yamabe Flow
|
||||
aliases: [discrete Ricci flow, Chow-Luo]
|
||||
- slug: hyperideal-tetrahedron
|
||||
title: Hyperideal Tetrahedron
|
||||
aliases: [hyperbolic volume, Schläfli, Ushijima]
|
||||
"""
|
||||
)
|
||||
|
||||
|
||||
@pytest.fixture
|
||||
def concepts_yaml(tmp_path: Path) -> Path:
|
||||
"""Write fixture YAML to a temp file and return the path."""
|
||||
p = tmp_path / "concepts.yaml"
|
||||
p.write_text(FIXTURE_YAML, encoding="utf-8")
|
||||
return p
|
||||
|
||||
|
||||
# ---------------------------------------------------------------------------
|
||||
# Tests
|
||||
# ---------------------------------------------------------------------------
|
||||
|
||||
|
||||
def test_load_concepts_returns_all_concepts(concepts_yaml: Path) -> None:
|
||||
"""All five concept entries are returned."""
|
||||
concepts = load_concepts(str(concepts_yaml))
|
||||
assert len(concepts) == 5
|
||||
|
||||
|
||||
def test_load_concepts_types(concepts_yaml: Path) -> None:
|
||||
"""Every returned item is a Concept instance."""
|
||||
concepts = load_concepts(str(concepts_yaml))
|
||||
for c in concepts:
|
||||
assert isinstance(c, Concept)
|
||||
|
||||
|
||||
def test_load_concepts_first_slug_and_title(concepts_yaml: Path) -> None:
|
||||
"""First concept has correct slug and title."""
|
||||
concepts = load_concepts(str(concepts_yaml))
|
||||
assert concepts[0].slug == "discrete-conformal-map"
|
||||
assert concepts[0].title == "Discrete Conformal Map"
|
||||
|
||||
|
||||
def test_load_concepts_aliases_are_lists(concepts_yaml: Path) -> None:
|
||||
"""Aliases are parsed as a list of strings."""
|
||||
concepts = load_concepts(str(concepts_yaml))
|
||||
for c in concepts:
|
||||
assert isinstance(c.aliases, list)
|
||||
for alias in c.aliases:
|
||||
assert isinstance(alias, str)
|
||||
|
||||
|
||||
def test_load_concepts_first_has_two_aliases(concepts_yaml: Path) -> None:
|
||||
"""First concept has exactly two aliases."""
|
||||
concepts = load_concepts(str(concepts_yaml))
|
||||
assert concepts[0].aliases == ["discrete conformal equivalence", "discrete uniformization"]
|
||||
|
||||
|
||||
def test_load_concepts_emphasis_present(concepts_yaml: Path) -> None:
|
||||
"""Concepts with emphasis field have it parsed correctly."""
|
||||
concepts = load_concepts(str(concepts_yaml))
|
||||
dcm = next(c for c in concepts if c.slug == "discrete-conformal-map")
|
||||
assert dcm.emphasis is not None
|
||||
assert "variationelle" in dcm.emphasis
|
||||
|
||||
|
||||
def test_load_concepts_emphasis_absent_is_none(concepts_yaml: Path) -> None:
|
||||
"""Concepts without emphasis field have emphasis=None."""
|
||||
concepts = load_concepts(str(concepts_yaml))
|
||||
cp = next(c for c in concepts if c.slug == "circle-packing")
|
||||
assert cp.emphasis is None
|
||||
|
||||
|
||||
def test_load_concepts_slugs_unique(concepts_yaml: Path) -> None:
|
||||
"""All slugs are distinct."""
|
||||
concepts = load_concepts(str(concepts_yaml))
|
||||
slugs = [c.slug for c in concepts]
|
||||
assert len(slugs) == len(set(slugs))
|
||||
|
||||
|
||||
def test_load_concepts_unicode_in_aliases(concepts_yaml: Path) -> None:
|
||||
"""Aliases with non-ASCII characters (Schläfli) are parsed without error."""
|
||||
concepts = load_concepts(str(concepts_yaml))
|
||||
ht = next(c for c in concepts if c.slug == "hyperideal-tetrahedron")
|
||||
assert any("Schläfli" in a for a in ht.aliases)
|
||||
|
||||
|
||||
def test_load_concepts_empty_yaml(tmp_path: Path) -> None:
|
||||
"""An empty concepts list returns an empty list (no crash)."""
|
||||
p = tmp_path / "concepts.yaml"
|
||||
p.write_text("concepts: []\n", encoding="utf-8")
|
||||
concepts = load_concepts(str(p))
|
||||
assert concepts == []
|
||||
Reference in New Issue
Block a user