[perf] Improve analyze file filter for header

analyzer/codechecker_analyzer/cli/analyze.py

@@ -1012,13 +1012,15 @@ def is_checker_config_valid(
 
 def get_affected_file_paths(
     file_filters: List[str],
-    compile_commands: tu_collector.CompilationDB
+    compile_commands: tu_collector.CompilationDB,
+    jobs: int
 ) -> List[str]:
     """
     Returns a list of source files for existing header file otherwise returns
     with the same file path expression.
     """
     file_paths = []  # Use list to keep the order of the file paths.
+    dependencies: tu_collector.Dependencies = collections.defaultdict(set)
     for file_filter in file_filters:
         file_paths.append(str(Path(file_filter).resolve())
                           if '*' not in file_filter else file_filter)
@@ -1027,7 +1029,7 @@ def get_affected_file_paths(
                 file_filter.endswith(header_file_extensions):
             LOG.info("Get dependent source files for '%s'...", file_filter)
             dependent_sources = tu_collector.get_dependent_sources(
-                compile_commands, file_filter)
+                compile_commands, file_filter, dependencies, jobs)
 
             LOG.info("Get dependent source files for '%s' done.", file_filter)
             LOG.debug("Dependent source files: %s",
@@ -1046,7 +1048,7 @@ def __get_skip_handlers(args, compile_commands) -> SkipListHandlers:
     skip_handlers = SkipListHandlers()
     if 'files' in args:
         source_file_paths = get_affected_file_paths(
-            args.files, compile_commands)
+            args.files, compile_commands, args.jobs)
 
         # Creates a skip file where all source files will be skipped except
         # the given source files and all the header files.

tools/tu_collector/tu_collector/tu_collector.py

@@ -28,10 +28,11 @@
 import subprocess
 import sys
 import zipfile
+from multiprocessing import Pool
 from shutil import which
 
 from pathlib import Path
-from typing import Iterable, Iterator, List, Optional, Set, Tuple, Union
+from typing import Dict, Iterable, Iterator, List, Optional, Set, Tuple, Union
 
 if sys.version_info >= (3, 8):
     from typing import TypedDict  # pylint: disable=no-name-in-module
@@ -56,6 +57,7 @@ class CompileAction(TypedDict):
 
 
 CompilationDB = List[CompileAction]
+Dependencies = Dict[str, Set[str]]
 
 
 def __random_string(length: int) -> str:
@@ -472,21 +474,36 @@ def zip_tu_files(
                          json.dumps(compilation_database, indent=2))
 
 
+def __get_dependent_headers_for_build_action(build_action: CompileAction
+                                             ) -> Tuple[str, Set[str]]:
+    """ Return the source file and dependent headers for a build action. """
+    files, _ = get_dependent_headers(
+        build_action['command'],
+        build_action['directory'])
+
+    source_file = os.path.join(build_action['directory'],
+                               build_action['file'])
+    return source_file, files
+
+
 def get_dependent_sources(
     compilation_db: CompilationDB,
-    header_path: Optional[str] = None
+    header_path: Optional[str] = None,
+    dependencies: Optional[Dependencies] = None,
+    jobs: Optional[int] = None
 ) -> Set[str]:
     """ Get dependencies for each files in each translation unit. """
-    dependencies = collections.defaultdict(set)
-    for build_action in compilation_db:
-        files, _ = get_dependent_headers(
-            build_action['command'],
-            build_action['directory'])
-
-        source_file = os.path.join(build_action['directory'],
-                                   build_action['file'])
-        for f in files:
-            dependencies[f].add(source_file)
+    if dependencies is None:
+        dependencies = collections.defaultdict(set)
+    if jobs is None:
+        jobs = 1
+    if not dependencies:
+        with Pool(jobs) as p:
+            results = list(p.map(__get_dependent_headers_for_build_action,
+                                 compilation_db))
+        for source_file, files in results:
+            for f in files:
+                dependencies[f].add(source_file)
 
     pattern = None
     if header_path: