feat(tsl-api-reference): index class and unit APIs

Migrate reference pages to declaration-line storage and rebuild the
13-column index.

Add qualified member lookup plus regression and end-to-end coverage.
This commit is contained in:
csh
2026-07-29 15:47:46 +08:00
parent 9edc8fc868
commit 13be5ea0aa
239 changed files with 39157 additions and 12846 deletions
+752
View File
@@ -1,5 +1,10 @@
import io
import importlib.util
import subprocess
import sys
import tempfile
import unittest
from contextlib import redirect_stdout
from pathlib import Path
@@ -10,6 +15,7 @@ SCRIPT = (
/ "scripts"
/ "lookup.py"
)
SKILL_MD = SCRIPT.parents[1] / "SKILL.md"
def load_script():
@@ -20,6 +26,752 @@ def load_script():
class LookupTest(unittest.TestCase):
def write_unified_api_fixture(self, skill_root):
data_dir = skill_root / "data"
codegen_dir = skill_root / "references" / "codegen"
data_dir.mkdir()
(codegen_dir / "project").mkdir(parents=True)
header = [
"name",
"scope",
"module",
"signature",
"page",
"anchor",
"tags",
"summary",
"kind",
"binding",
"visibility",
"owner",
"qualified_name",
]
def row(
name,
signature,
page,
anchor,
summary,
kind,
binding="",
visibility="",
owner="",
qualified_name="",
):
return [
name,
"project",
"demo",
signature,
page,
anchor,
"",
summary,
kind,
binding,
visibility,
owner,
qualified_name or name,
]
rows = [
row(
"DemoUnit",
"DemoUnit",
"project/demo_unit.md",
"demounit",
"接口。",
"unit",
),
row(
"Open",
"Open()",
"project/demo_unit.md",
"open",
"打开。",
"function",
"unit",
"public",
"DemoUnit",
"DemoUnit.Open",
),
row(
"Document",
"Document",
"project/demo_unit.md",
"document",
"文档。",
"class",
"unit",
"public",
"DemoUnit",
"DemoUnit.Document",
),
row(
"Save",
"Save(path)",
"project/demo_unit.md",
"savepath",
"按路径保存。",
"method",
"instance",
"public",
"DemoUnit.Document",
"DemoUnit.Document.Save",
),
row(
"Save",
"Save(mode)",
"project/demo_unit.md",
"savemode",
"按模式保存。",
"method",
"instance",
"protected",
"DemoUnit.Document",
"DemoUnit.Document.Save",
),
row(
"Name",
"Name",
"project/demo_unit.md",
"name",
"文档名称。",
"field",
"instance",
"public",
"DemoUnit.Document",
"DemoUnit.Document.Name",
),
row(
"Other",
"Other",
"project/demo_unit.md",
"other",
"其他对象。",
"class",
"unit",
"public",
"DemoUnit",
"DemoUnit.Other",
),
row(
"Save",
"Save()",
"project/demo_unit.md",
"save",
"保存其他对象。",
"method",
"instance",
"public",
"DemoUnit.Other",
"DemoUnit.Other.Save",
),
row(
"Widget",
"Widget",
"project/widget.md",
"widget",
"组件。",
"class",
),
row(
"Create",
"Create()",
"project/widget.md",
"create",
"创建组件。",
"method",
"class",
"public",
"Widget",
"Widget.Create",
),
row(
"Count",
"Count",
"project/widget.md",
"count",
"组件数量。",
"field",
"static",
"protected",
"Widget",
"Widget.Count",
),
]
index_text = "\n".join(
["\t".join(header), *("\t".join(values) for values in rows)]
)
tsv_path = data_dir / "function_index.tsv"
tsv_path.write_text(index_text + "\n", encoding="utf-8")
(codegen_dir / "project" / "demo_unit.md").write_text(
"# Demo / Unit\n\n"
"## `DemoUnit`\n\n"
"声明:unit\n\n"
"接口。\n\n"
"### `Open()`\n\n"
"声明:function\n\n"
"打开。\n\n"
"返回:Document\n\n"
"### `Document`\n\n"
"声明:class\n\n"
"文档。\n\n"
"#### `Save(path)`\n\n"
"声明:function\n\n"
"按路径保存。\n\n"
"可见性:`public`\n\n"
"#### `Save(mode)`\n\n"
"声明:function\n\n"
"按模式保存。\n\n"
"可见性:`protected`\n\n"
"#### `Name`\n\n"
"声明:field\n\n"
"文档名称。\n\n"
"可见性:`public`\n\n"
"类型:string\n\n"
"### `Other`\n\n"
"声明:class\n\n"
"其他对象。\n\n"
"#### `Save()`\n\n"
"声明:function\n\n"
"保存其他对象。\n\n"
"可见性:`public`\n",
encoding="utf-8",
)
(codegen_dir / "project" / "widget.md").write_text(
"# Demo / Class\n\n"
"## `Widget`\n\n"
"声明:class\n\n"
"组件。\n\n"
"### `Create()`\n\n"
"声明:class function\n\n"
"创建组件。\n\n"
"可见性:`public`\n\n"
"### `Count`\n\n"
"声明:static field\n\n"
"组件数量。\n\n"
"可见性:`protected`\n\n"
"类型:integer\n",
encoding="utf-8",
)
return tsv_path
def write_mixed_api_fixture(self, skill_root):
data_dir = skill_root / "data"
codegen_dir = skill_root / "references" / "codegen" / "project"
data_dir.mkdir()
codegen_dir.mkdir(parents=True)
header = [
"name",
"scope",
"module",
"signature",
"page",
"anchor",
"tags",
"summary",
"kind",
"binding",
"visibility",
"owner",
"qualified_name",
]
page = "project/mixed.md"
rows = [
[
"Widget",
"project",
"mixed",
"Widget",
page,
"widget",
"",
"组件。",
"class",
"",
"",
"",
"Widget",
],
[
"Open",
"project",
"mixed",
"Open()",
page,
"open",
"",
"打开。",
"method",
"instance",
"public",
"Widget",
"Widget.Open",
],
[
"Parse",
"project",
"mixed",
"Parse(text)",
page,
"parse",
"",
"解析。",
"function",
"",
"",
"",
"Parse",
],
[
"Runtime",
"project",
"mixed",
"Runtime",
page,
"runtime",
"",
"接口。",
"unit",
"",
"",
"",
"Runtime",
],
[
"Document",
"project",
"mixed",
"Document",
page,
"document",
"",
"文档。",
"class",
"unit",
"public",
"Runtime",
"Runtime.Document",
],
[
"Save",
"project",
"mixed",
"Save()",
page,
"save",
"",
"保存。",
"method",
"instance",
"public",
"Runtime.Document",
"Runtime.Document.Save",
],
[
"Other",
"project",
"mixed",
"Other",
page,
"other",
"",
"其他组件。",
"class",
"",
"",
"",
"Other",
],
]
tsv_path = data_dir / "function_index.tsv"
tsv_path.write_text(
"\n".join(
["\t".join(header), *("\t".join(row) for row in rows)]
)
+ "\n",
encoding="utf-8",
)
(codegen_dir / "mixed.md").write_text(
"# Mixed\n\n"
"## `Widget`\n\n声明:class\n\n组件。\n\n"
"### `Open()`\n\n声明:function\n\n打开。\n\n可见性:public\n\n"
"## `Parse(text)`\n\n声明:function\n\n解析。\n\n"
"| 参数 | 类型 | 说明 |\n"
"| --- | --- | --- |\n"
"| `text` | string | 文本 |\n\n"
"返回:Widget\n\n"
"### 示例\n\n"
"```markdown\n## `NotBoundary`\n```\n\n"
"## `Runtime`\n\n声明:unit\n\n接口。\n\n"
"### `Document`\n\n声明:class\n\n文档。\n\n"
"#### `Save()`\n\n声明:function\n\n保存。\n\n可见性:public\n\n"
"## `Other`\n\n声明:class\n\n其他组件。\n",
encoding="utf-8",
)
return tsv_path
def run_lookup(self, tsv_path, *args):
return subprocess.run(
[
sys.executable,
str(SCRIPT),
*args,
"--tsv",
str(tsv_path),
],
capture_output=True,
text=True,
encoding="utf-8",
check=False,
)
def test_tsv_without_tags_is_rejected(self):
with tempfile.TemporaryDirectory() as temp_dir:
tsv_path = Path(temp_dir) / "function_index.tsv"
tsv_path.write_text(
"name\tscope\tmodule\tsignature\tpage\tanchor\tsummary\n"
"foo\tbuiltin\tmisc\tfoo()\tmisc.md\tfoo\tLegacy entry\n",
encoding="utf-8",
)
result = subprocess.run(
[
sys.executable,
str(SCRIPT),
"--kw",
"Legacy",
"--tsv",
str(tsv_path),
],
capture_output=True,
text=True,
check=False,
)
self.assertEqual(1, result.returncode)
self.assertEqual("", result.stdout)
self.assertIn("missing required columns: tags", result.stderr)
self.assertNotIn("Traceback", result.stderr)
def test_missing_required_tsv_columns_are_rejected(self):
with tempfile.TemporaryDirectory() as temp_dir:
tsv_path = Path(temp_dir) / "function_index.tsv"
tsv_path.write_text("name\nfoo\n", encoding="utf-8")
result = subprocess.run(
[
sys.executable,
str(SCRIPT),
"--name",
"foo",
"--tsv",
str(tsv_path),
],
capture_output=True,
text=True,
check=False,
)
self.assertEqual(1, result.returncode)
self.assertIn("missing required columns", result.stderr)
self.assertNotIn("Traceback", result.stderr)
def test_tsv_row_width_is_validated(self):
with tempfile.TemporaryDirectory() as temp_dir:
tsv_path = Path(temp_dir) / "function_index.tsv"
tsv_path.write_text(
"name\tscope\tmodule\tsignature\tpage\tanchor\ttags\tsummary\n"
"foo\tbuiltin\n",
encoding="utf-8",
)
result = subprocess.run(
[
sys.executable,
str(SCRIPT),
"--name",
"foo",
"--tsv",
str(tsv_path),
],
capture_output=True,
text=True,
check=False,
)
self.assertEqual(1, result.returncode)
self.assertIn("line 2 has 2 columns; expected 8", result.stderr)
self.assertNotIn("Traceback", result.stderr)
def test_missing_entry_body_is_rejected(self):
with tempfile.TemporaryDirectory() as temp_dir:
skill_root = Path(temp_dir)
data_dir = skill_root / "data"
codegen_dir = skill_root / "references" / "codegen"
data_dir.mkdir()
codegen_dir.mkdir(parents=True)
(data_dir / "function_index.tsv").write_text(
"name\tscope\tmodule\tsignature\tpage\tanchor\ttags\tsummary\n"
"foo\tbuiltin\tmisc\tfoo()\tmissing.md\tfoo\t\tFallback only\n",
encoding="utf-8",
)
result = subprocess.run(
[
sys.executable,
str(SCRIPT),
"--name",
"foo",
"--tsv",
str(data_dir / "function_index.tsv"),
],
capture_output=True,
text=True,
check=False,
)
self.assertEqual(1, result.returncode)
self.assertEqual("", result.stdout)
self.assertIn("entry body not found", result.stderr)
self.assertIn("missing.md#foo", result.stderr)
def test_empty_queries_are_rejected(self):
for option in ("--name", "--kw"):
with self.subTest(option=option):
result = subprocess.run(
[sys.executable, str(SCRIPT), option, ""],
capture_output=True,
text=True,
check=False,
)
self.assertEqual(2, result.returncode)
self.assertIn("must not be empty", result.stderr)
def test_non_positive_limit_is_rejected(self):
for limit in ("0", "-1"):
with self.subTest(limit=limit):
result = subprocess.run(
[
sys.executable,
str(SCRIPT),
"--kw",
"数组",
"--limit",
limit,
],
capture_output=True,
text=True,
check=False,
)
self.assertEqual(2, result.returncode)
self.assertIn("--limit must be >= 1", result.stderr)
def test_exact_search_uses_anchor_for_duplicate_signatures(self):
module = load_script()
with tempfile.TemporaryDirectory() as temp_dir:
skill_root = Path(temp_dir)
data_dir = skill_root / "data"
codegen_dir = skill_root / "references" / "codegen"
data_dir.mkdir()
codegen_dir.mkdir(parents=True)
header = [
"name",
"scope",
"module",
"signature",
"page",
"anchor",
"tags",
"summary",
]
rows = [
[
"CAPA",
"builtin",
"language",
"CAPA()",
"objects.md",
"capa",
"",
"IMAP summary",
],
[
"CAPA",
"builtin",
"language",
"CAPA()",
"objects.md",
"capa-1",
"",
"Pop3 summary",
],
]
index_text = "\n".join(
["\t".join(header), *("\t".join(row) for row in rows)]
)
(data_dir / "function_index.tsv").write_text(
index_text + "\n", encoding="utf-8"
)
(codegen_dir / "objects.md").write_text(
"# Objects\n\n"
"## `CAPA()`\n\n"
"声明:function\n\n"
"对象:IMAP\n\n"
"## `CAPA()`\n\n"
"声明:function\n\n"
"对象:Pop3\n\n"
"",
encoding="utf-8",
)
stdout = io.StringIO()
with redirect_stdout(stdout):
status = module.main(
[
"--name",
"CAPA",
"--tsv",
str(data_dir / "function_index.tsv"),
]
)
output = stdout.getvalue()
self.assertEqual(0, status)
self.assertEqual(1, output.count("对象:IMAP"))
self.assertEqual(1, output.count("对象:Pop3"))
self.assertIn("objects.md#capa -->", output)
self.assertIn("objects.md#capa-1 -->", output)
def test_exact_qualified_member_query_returns_only_member_overloads(self):
with tempfile.TemporaryDirectory() as temp_dir:
tsv_path = self.write_unified_api_fixture(Path(temp_dir))
result = self.run_lookup(
tsv_path, "--name", "demounit.document.save"
)
self.assertEqual(0, result.returncode, result.stderr)
self.assertEqual(2, result.stdout.count("#### `Save("))
self.assertNotIn("#### `Name`", result.stdout)
self.assertNotIn("### `Other`", result.stdout)
def test_exact_unit_class_query_returns_its_h4_subtree_only(self):
with tempfile.TemporaryDirectory() as temp_dir:
tsv_path = self.write_unified_api_fixture(Path(temp_dir))
result = self.run_lookup(
tsv_path, "--name", "DemoUnit.Document"
)
self.assertEqual(0, result.returncode, result.stderr)
self.assertIn("### `Document`", result.stdout)
self.assertIn("#### `Save(path)`", result.stdout)
self.assertIn("#### `Name`", result.stdout)
self.assertNotIn("### `Other`", result.stdout)
def test_exact_unit_query_returns_the_complete_unit_subtree(self):
with tempfile.TemporaryDirectory() as temp_dir:
tsv_path = self.write_unified_api_fixture(Path(temp_dir))
result = self.run_lookup(tsv_path, "--name", "DemoUnit")
self.assertEqual(0, result.returncode, result.stderr)
self.assertIn("## `DemoUnit`", result.stdout)
self.assertIn("### `Open()`", result.stdout)
self.assertIn("### `Document`", result.stdout)
self.assertIn("#### `Save(path)`", result.stdout)
self.assertIn("### `Other`", result.stdout)
self.assertNotIn("## `Widget`", result.stdout)
def test_exact_class_query_returns_all_class_members(self):
with tempfile.TemporaryDirectory() as temp_dir:
tsv_path = self.write_unified_api_fixture(Path(temp_dir))
result = self.run_lookup(tsv_path, "--name", "widget")
self.assertEqual(0, result.returncode, result.stderr)
self.assertIn("## `Widget`", result.stdout)
self.assertIn("### `Create()`\n\n声明:class function", result.stdout)
self.assertIn("### `Count`\n\n声明:static field", result.stdout)
self.assertNotIn("## `DemoUnit`", result.stdout)
def test_exact_class_in_mixed_page_stops_at_next_h2(self):
with tempfile.TemporaryDirectory() as temp_dir:
tsv_path = self.write_mixed_api_fixture(Path(temp_dir))
result = self.run_lookup(tsv_path, "--name", "Widget")
self.assertEqual(0, result.returncode, result.stderr)
self.assertIn("### `Open()`", result.stdout)
self.assertNotIn("## `Parse(text)`", result.stdout)
def test_exact_function_in_mixed_page_keeps_h3_example_only(self):
with tempfile.TemporaryDirectory() as temp_dir:
tsv_path = self.write_mixed_api_fixture(Path(temp_dir))
result = self.run_lookup(tsv_path, "--name", "Parse")
self.assertEqual(0, result.returncode, result.stderr)
self.assertIn("### 示例", result.stdout)
self.assertIn("## `NotBoundary`", result.stdout)
self.assertNotIn("## `Runtime`", result.stdout)
def test_exact_unit_in_mixed_page_stops_before_following_class(self):
with tempfile.TemporaryDirectory() as temp_dir:
tsv_path = self.write_mixed_api_fixture(Path(temp_dir))
result = self.run_lookup(tsv_path, "--name", "Runtime")
self.assertEqual(0, result.returncode, result.stderr)
self.assertIn("#### `Save()`", result.stdout)
self.assertNotIn("## `Other`", result.stdout)
def test_unqualified_member_query_returns_all_owners(self):
with tempfile.TemporaryDirectory() as temp_dir:
tsv_path = self.write_unified_api_fixture(Path(temp_dir))
result = self.run_lookup(tsv_path, "--name", "Save")
self.assertEqual(0, result.returncode, result.stderr)
self.assertEqual(3, result.stdout.count("#### `Save("))
self.assertIn("按路径保存。", result.stdout)
self.assertIn("按模式保存。", result.stdout)
self.assertIn("保存其他对象。", result.stdout)
def test_keyword_search_includes_unified_fields_and_declaration_term(self):
with tempfile.TemporaryDirectory() as temp_dir:
tsv_path = self.write_unified_api_fixture(Path(temp_dir))
qualified = self.run_lookup(
tsv_path, "--kw", "Widget.Create"
)
visibility = self.run_lookup(
tsv_path, "--kw", "protected", "method"
)
self.assertEqual(0, qualified.returncode, qualified.stderr)
self.assertIn("Widget.Create\tclass function\tCreate()", qualified.stdout)
self.assertNotIn("static function", qualified.stdout)
self.assertEqual(0, visibility.returncode, visibility.stderr)
self.assertIn("DemoUnit.Document.Save", visibility.stdout)
def test_skill_documents_qualified_queries_and_legacy_index_compatibility(self):
text = SKILL_MD.read_text(encoding="utf-8")
self.assertIn("完全限定名称", text)
self.assertIn("qualified_name", text)
self.assertIn("owner、kind、binding、visibility", text)
self.assertIn("class function", text)
self.assertIn("旧 8 列", text)
def test_keyword_search_includes_tags_and_summary(self):
module = load_script()
rows = [