improve deps CI formatting and name resolution (#6847)

<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->

## Summary by CodeRabbit

* **Improvements**
* Enhanced test module dependency tracking in automated checks to
properly identify and report test-related dependencies.
* Improved dependency resolution logic with better module name mapping
and deduplication for clearer dependency reporting.
* Refined formatting of dependency information and todo items for
improved readability.
* Dependency review comments now automatically refresh with the latest
information when updated.

<sub>✏️ Tip: You can customize this high-level summary in your review
settings.</sub>

<!-- end of auto-generated comment: release notes by coderabbit.ai -->

.github/workflows/lib-deps-check.yaml

@@ -45,15 +45,20 @@ jobs:
           echo "Changed files:"
           echo "$changed"
 
-          # Extract unique module names (top-level only, skip test/)
+          # Extract unique module names
           modules=""
           for file in $changed; do
-            # Skip test files
             if [[ "$file" == Lib/test/* ]]; then
-              continue
+              # Test files: Lib/test/test_pydoc.py -> test_pydoc, Lib/test/test_pydoc/foo.py -> test_pydoc
+              module=$(echo "$file" | sed -E 's|^Lib/test/||; s|\.py$||; s|/.*||')
+              # Skip non-test files in test/ (e.g., support.py, __init__.py)
+              if [[ ! "$module" == test_* ]]; then
+                continue
+              fi
+            else
+              # Lib files: Lib/foo.py -> foo, Lib/foo/__init__.py -> foo
+              module=$(echo "$file" | sed -E 's|^Lib/||; s|/__init__\.py$||; s|\.py$||; s|/.*||')
             fi
-            # Extract module name: Lib/foo.py -> foo, Lib/foo/__init__.py -> foo
-            module=$(echo "$file" | sed -E 's|^Lib/||; s|/__init__\.py$||; s|\.py$||; s|/.*||')
             if [[ -n "$module" && ! " $modules " =~ " $module " ]]; then
               modules="$modules $module"
             fi
@@ -94,6 +99,7 @@ jobs:
         with:
           header: lib-deps-check
           number: ${{ github.event.pull_request.number }}
+          recreate: true
           message: |
             ## 📦 Library Dependencies
 

scripts/update_lib/deps.py

@@ -101,6 +101,32 @@ DEPENDENCIES = {
     },
 }
 
+def resolve_hard_dep_parent(name: str) -> str | None:
+    """Resolve a hard_dep name to its parent module.
+
+    If 'name' is listed as a hard_dep of another module, return that module's name.
+    E.g., 'pydoc_data' -> 'pydoc', '_pydatetime' -> 'datetime'
+
+    Args:
+        name: Module or file name (with or without .py extension)
+
+    Returns:
+        Parent module name if found, None otherwise
+    """
+    # Normalize: remove .py extension if present
+    if name.endswith(".py"):
+        name = name[:-3]
+
+    for module_name, dep_info in DEPENDENCIES.items():
+        hard_deps = dep_info.get("hard_deps", [])
+        for dep in hard_deps:
+            # Normalize dep: remove .py extension
+            dep_normalized = dep[:-3] if dep.endswith(".py") else dep
+            if dep_normalized == name:
+                return module_name
+    return None
+
+
 # Test-specific dependencies (only when auto-detection isn't enough)
 # - hard_deps: files to migrate (tightly coupled, must be migrated together)
 # - data: directories to copy without migration
@@ -254,10 +280,11 @@ def _extract_top_level_code(content: str) -> str:
 
 _FROM_TEST_IMPORT_RE = re.compile(r"^from test import (.+)", re.MULTILINE)
 _FROM_TEST_DOT_RE = re.compile(r"^from test\.(\w+)", re.MULTILINE)
+_IMPORT_TEST_DOT_RE = re.compile(r"^import test\.(\w+)", re.MULTILINE)
 
 
 def parse_test_imports(content: str) -> set[str]:
-    """Parse test file content and extract 'from test import ...' dependencies.
+    """Parse test file content and extract test package dependencies.
 
     Uses regex for speed - only matches top-level imports.
 
@@ -285,6 +312,12 @@ def parse_test_imports(content: str) -> set[str]:
         if dep not in ("support", "__init__"):
             imports.add(dep)
 
+    # Match "import test.foo" -> depends on foo
+    for match in _IMPORT_TEST_DOT_RE.finditer(content):
+        dep = match.group(1)
+        if dep not in ("support", "__init__"):
+            imports.add(dep)
+
     return imports
 
 

scripts/update_lib/show_deps.py

@@ -163,6 +163,7 @@ def format_deps(
         find_dependent_tests_tree,
         get_lib_paths,
         get_test_paths,
+        resolve_hard_dep_parent,
     )
 
     if _visited is None:
@@ -170,17 +171,34 @@ def format_deps(
 
     lines = []
 
+    # Resolve test_ prefix to module (e.g., test_pydoc -> pydoc)
+    if name.startswith("test_"):
+        module_name = name[5:]  # strip "test_"
+        lines.append(f"(redirecting {name} -> {module_name})")
+        name = module_name
+
+    # Resolve hard_dep to parent module (e.g., pydoc_data -> pydoc)
+    parent = resolve_hard_dep_parent(name)
+    if parent:
+        lines.append(f"(redirecting {name} -> {parent})")
+        name = parent
+
     # lib paths (only show existing)
     lib_paths = get_lib_paths(name, cpython_prefix)
-    for p in lib_paths:
-        if p.exists():
-            lines.append(f"[+] lib: {p}")
+    existing_lib_paths = [p for p in lib_paths if p.exists()]
+    for p in existing_lib_paths:
+        lines.append(f"[+] lib: {p}")
 
     # test paths (only show existing)
     test_paths = get_test_paths(name, cpython_prefix)
-    for p in test_paths:
-        if p.exists():
-            lines.append(f"[+] test: {p}")
+    existing_test_paths = [p for p in test_paths if p.exists()]
+    for p in existing_test_paths:
+        lines.append(f"[+] test: {p}")
+
+    # If no lib or test paths exist, module doesn't exist
+    if not existing_lib_paths and not existing_test_paths:
+        lines.append(f"(module '{name}' not found)")
+        return lines
 
     # hard_deps (from DEPENDENCIES table)
     dep_info = DEPENDENCIES.get(name, {})
@@ -188,7 +206,7 @@ def format_deps(
     if hard_deps:
         lines.append(f"packages: {hard_deps}")
 
-    lines.append("dependencies:")
+    lines.append("\ndependencies:")
     lines.extend(
         format_deps_tree(
             cpython_prefix, lib_prefix, max_depth, soft_deps={name}, _visited=_visited
@@ -227,9 +245,9 @@ def _format_dependent_tests_tree(
 
         total = count_tests(tree)
         if total == 0 and not children:
-            lines.append(f"dependent tests: (no tests depend on {module})")
+            lines.append(f"\ndependent tests: (no tests depend on {module})")
             return lines
-        lines.append(f"dependent tests: ({total} tests)")
+        lines.append(f"\ndependent tests: ({total} tests)")
 
     # Check if module is up-to-date
     synced = is_up_to_date(module.split(".")[0], cpython_prefix, lib_prefix)
@@ -258,6 +276,56 @@ def _format_dependent_tests_tree(
     return lines
 
 
+def _resolve_module_name(
+    name: str,
+    cpython_prefix: str,
+    lib_prefix: str,
+) -> list[str]:
+    """Resolve module name through redirects.
+
+    Returns a list of module names (usually 1, but test support files may expand to multiple).
+    """
+    import pathlib
+
+    from update_lib.deps import (
+        _build_test_import_graph,
+        get_lib_paths,
+        get_test_paths,
+        resolve_hard_dep_parent,
+    )
+
+    # Resolve test_ prefix
+    if name.startswith("test_"):
+        name = name[5:]
+
+    # Resolve hard_dep to parent
+    parent = resolve_hard_dep_parent(name)
+    if parent:
+        return [parent]
+
+    # Check if it's a valid module
+    lib_paths = get_lib_paths(name, cpython_prefix)
+    test_paths = get_test_paths(name, cpython_prefix)
+    if any(p.exists() for p in lib_paths) or any(p.exists() for p in test_paths):
+        return [name]
+
+    # Check for test support files (e.g., string_tests -> bytes, str, userstring)
+    test_support_path = pathlib.Path(cpython_prefix) / "Lib" / "test" / f"{name}.py"
+    if test_support_path.exists():
+        test_dir = pathlib.Path(lib_prefix) / "test"
+        if test_dir.exists():
+            import_graph, _ = _build_test_import_graph(test_dir)
+            importing_tests = []
+            for file_key, imports in import_graph.items():
+                if name in imports and file_key.startswith("test_"):
+                    importing_tests.append(file_key)
+            if importing_tests:
+                # Resolve test names to module names (test_bytes -> bytes)
+                return sorted(set(t[5:] for t in importing_tests))
+
+    return [name]
+
+
 def show_deps(
     names: list[str],
     cpython_prefix: str = "cpython",
@@ -273,10 +341,19 @@ def show_deps(
         else:
             expanded_names.append(name)
 
+    # Resolve and deduplicate names (preserving order)
+    seen: set[str] = set()
+    resolved_names: list[str] = []
+    for name in expanded_names:
+        for resolved in _resolve_module_name(name, cpython_prefix, lib_prefix):
+            if resolved not in seen:
+                seen.add(resolved)
+                resolved_names.append(resolved)
+
     # Shared visited set across all modules
     visited: set[str] = set()
 
-    for i, name in enumerate(expanded_names):
+    for i, name in enumerate(resolved_names):
         if i > 0:
             print()  # blank line between modules
         for line in format_deps(name, cpython_prefix, lib_prefix, max_depth, visited):

scripts/update_lib/show_todo.py

@@ -143,7 +143,7 @@ def format_todo_list(
 
         rev_str = f"{rev_count} dependents" if rev_count else ""
 
-        parts = [done_mark, f"[{score_str}]", name]
+        parts = ["-", done_mark, f"[{score_str}]", name]
         if rev_str:
             parts.append(f"({rev_str})")