[ci] Fix clang-tidy split mode for core file changes (#11434)

script/helpers.py

@@ -1,5 +1,6 @@
 from __future__ import annotations
 
+from collections.abc import Callable
 from functools import cache
 import json
 import os
@@ -7,6 +8,7 @@ import os.path
 from pathlib import Path
 import re
 import subprocess
+import sys
 import time
 from typing import Any
 
@@ -304,7 +306,10 @@ def get_changed_components() -> list[str] | None:
         for f in changed
     )
     if core_cpp_changed:
-        print("Core C++/header files changed - will run full clang-tidy scan")
+        print(
+            "Core C++/header files changed - will run full clang-tidy scan",
+            file=sys.stderr,
+        )
         return None
 
     # Use list-components.py to get changed components
@@ -318,7 +323,10 @@ def get_changed_components() -> list[str] | None:
         return parse_list_components_output(result.stdout)
     except subprocess.CalledProcessError:
         # If the script fails, fall back to full scan
-        print("Could not determine changed components - will run full clang-tidy scan")
+        print(
+            "Could not determine changed components - will run full clang-tidy scan",
+            file=sys.stderr,
+        )
         return None
 
 
@@ -370,14 +378,14 @@ def _filter_changed_ci(files: list[str]) -> list[str]:
             if f in changed and not f.startswith(ESPHOME_COMPONENTS_PATH)
         ]
         if not files:
-            print("No files changed")
+            print("No files changed", file=sys.stderr)
         return files
 
     # Scenario 3: Specific components changed
     # Action: Check ALL files in each changed component
     # Convert component list to set for O(1) lookups
     component_set = set(components)
-    print(f"Changed components: {', '.join(sorted(components))}")
+    print(f"Changed components: {', '.join(sorted(components))}", file=sys.stderr)
 
     # The 'files' parameter contains ALL files in the codebase that clang-tidy would check.
     # We filter this down to only files in the changed components.
@@ -648,3 +656,220 @@ def get_components_from_integration_fixtures() -> set[str]:
                     components.add(item["platform"])
 
     return components
+
+
+def filter_component_files(file_path: str) -> bool:
+    """Check if a file path is a component file.
+
+    Args:
+        file_path: Path to check
+
+    Returns:
+        True if the file is in a component directory
+    """
+    return file_path.startswith("esphome/components/") or file_path.startswith(
+        "tests/components/"
+    )
+
+
+def extract_component_names_from_files(files: list[str]) -> list[str]:
+    """Extract unique component names from a list of file paths.
+
+    Args:
+        files: List of file paths
+
+    Returns:
+        List of unique component names (preserves order)
+    """
+    return list(
+        dict.fromkeys(comp for file in files if (comp := get_component_from_path(file)))
+    )
+
+
+def add_item_to_components_graph(
+    components_graph: dict[str, list[str]], parent: str, child: str
+) -> None:
+    """Add a dependency relationship to the components graph.
+
+    Args:
+        components_graph: Graph mapping parent components to their children
+        parent: Parent component name
+        child: Child component name (dependent)
+    """
+    if not parent.startswith("__") and parent != child:
+        if parent not in components_graph:
+            components_graph[parent] = []
+        if child not in components_graph[parent]:
+            components_graph[parent].append(child)
+
+
+def resolve_auto_load(
+    auto_load: list[str] | Callable[[], list[str]] | Callable[[dict | None], list[str]],
+    config: dict | None = None,
+) -> list[str]:
+    """Resolve AUTO_LOAD to a list, handling callables with or without config parameter.
+
+    Args:
+        auto_load: The AUTO_LOAD value (list or callable)
+        config: Optional config to pass to callable AUTO_LOAD functions
+
+    Returns:
+        List of component names to auto-load
+    """
+    if not callable(auto_load):
+        return auto_load
+
+    import inspect
+
+    if inspect.signature(auto_load).parameters:
+        return auto_load(config)
+    return auto_load()
+
+
+def create_components_graph() -> dict[str, list[str]]:
+    """Create a graph of component dependencies.
+
+    Returns:
+        Dictionary mapping parent components to their children (dependencies)
+    """
+    from pathlib import Path
+
+    from esphome import const
+    from esphome.core import CORE
+    from esphome.loader import ComponentManifest, get_component, get_platform
+
+    # The root directory of the repo
+    root = Path(__file__).parent.parent
+    components_dir = root / "esphome" / "components"
+    # Fake some directory so that get_component works
+    CORE.config_path = root
+    # Various configuration to capture different outcomes used by `AUTO_LOAD` function.
+    KEY_CORE = const.KEY_CORE
+    KEY_TARGET_FRAMEWORK = const.KEY_TARGET_FRAMEWORK
+    KEY_TARGET_PLATFORM = const.KEY_TARGET_PLATFORM
+    PLATFORM_ESP32 = const.PLATFORM_ESP32
+    PLATFORM_ESP8266 = const.PLATFORM_ESP8266
+
+    TARGET_CONFIGURATIONS = [
+        {KEY_TARGET_FRAMEWORK: None, KEY_TARGET_PLATFORM: None},
+        {KEY_TARGET_FRAMEWORK: "arduino", KEY_TARGET_PLATFORM: None},
+        {KEY_TARGET_FRAMEWORK: "esp-idf", KEY_TARGET_PLATFORM: None},
+        {KEY_TARGET_FRAMEWORK: None, KEY_TARGET_PLATFORM: PLATFORM_ESP32},
+        {KEY_TARGET_FRAMEWORK: None, KEY_TARGET_PLATFORM: PLATFORM_ESP8266},
+    ]
+    CORE.data[KEY_CORE] = TARGET_CONFIGURATIONS[0]
+
+    components_graph = {}
+    platforms = []
+    components: list[tuple[ComponentManifest, str, Path]] = []
+
+    for path in components_dir.iterdir():
+        if not path.is_dir():
+            continue
+        if not (path / "__init__.py").is_file():
+            continue
+        name = path.name
+        comp = get_component(name)
+        if comp is None:
+            raise RuntimeError(
+                f"Cannot find component {name}. Make sure current path is pip installed ESPHome"
+            )
+
+        components.append((comp, name, path))
+        if comp.is_platform_component:
+            platforms.append(name)
+
+    platforms = set(platforms)
+
+    for comp, name, path in components:
+        for dependency in comp.dependencies:
+            add_item_to_components_graph(
+                components_graph, dependency.split(".")[0], name
+            )
+
+        for target_config in TARGET_CONFIGURATIONS:
+            CORE.data[KEY_CORE] = target_config
+            for item in resolve_auto_load(comp.auto_load, config=None):
+                add_item_to_components_graph(components_graph, item, name)
+        # restore config
+        CORE.data[KEY_CORE] = TARGET_CONFIGURATIONS[0]
+
+        for platform_path in path.iterdir():
+            platform_name = platform_path.stem
+            if platform_name == name or platform_name not in platforms:
+                continue
+            platform = get_platform(platform_name, name)
+            if platform is None:
+                continue
+
+            add_item_to_components_graph(components_graph, platform_name, name)
+
+            for dependency in platform.dependencies:
+                add_item_to_components_graph(
+                    components_graph, dependency.split(".")[0], name
+                )
+
+            for target_config in TARGET_CONFIGURATIONS:
+                CORE.data[KEY_CORE] = target_config
+                for item in resolve_auto_load(platform.auto_load, config={}):
+                    add_item_to_components_graph(components_graph, item, name)
+            # restore config
+            CORE.data[KEY_CORE] = TARGET_CONFIGURATIONS[0]
+
+    return components_graph
+
+
+def find_children_of_component(
+    components_graph: dict[str, list[str]], component_name: str, depth: int = 0
+) -> list[str]:
+    """Find all components that depend on the given component (recursively).
+
+    Args:
+        components_graph: Graph mapping parent components to their children
+        component_name: Component name to find children for
+        depth: Current recursion depth (max 10)
+
+    Returns:
+        List of all dependent component names (may contain duplicates removed at end)
+    """
+    if component_name not in components_graph:
+        return []
+
+    children = []
+
+    for child in components_graph[component_name]:
+        children.append(child)
+        if depth < 10:
+            children.extend(
+                find_children_of_component(components_graph, child, depth + 1)
+            )
+    # Remove duplicate values
+    return list(set(children))
+
+
+def get_components_with_dependencies(
+    files: list[str], get_dependencies: bool = False
+) -> list[str]:
+    """Get component names from files, optionally including their dependencies.
+
+    Args:
+        files: List of file paths
+        get_dependencies: If True, include all dependent components
+
+    Returns:
+        Sorted list of component names
+    """
+    components = extract_component_names_from_files(files)
+
+    if get_dependencies:
+        components_graph = create_components_graph()
+
+        all_components = components.copy()
+        for c in components:
+            all_components.extend(find_children_of_component(components_graph, c))
+        # Remove duplicate values
+        all_changed_components = list(set(all_components))
+
+        return sorted(all_changed_components)
+
+    return sorted(components)