fix(ci): skip loaded lines calculation on iterations after the first (#17790)

This PR optimizes the pure execution latency and reduces redundant disk
I/O of the import profiler script by calculating the deterministic
codebase statistics (loaded lines count) only on the first iteration.

### Problem
For each package profiled, the script executes 11 iterations for the
baseline commit and 11 iterations for the PR commit (22 iterations
total).
On every single iteration, the worker process scans and counts the lines
of every imported `.py` file (across standard library and third-party
dependencies like `grpc`, `protobuf`, `google-api-core`, etc.). This
causes redundant and expensive disk I/O.

### Solution
- **First Iteration Only**: The first worker run profiles and counts the
loaded line count as a baseline.
- **Skip Subsequent Counting**: For iterations 2-11 (`i > 0`), the
master passes `--skip-line-count` to the worker process. The worker
skips the file-reading loop entirely and returns `-1` for
`loaded_lines`.
- **Improved Logging**: Worker process `stderr` is now forwarded to the
parent master process so that warnings are visible in GHA logs.

### Benchmarks (Timing the line-counting loop itself)
- **Large Packages (`google-cloud-compute`)**:
  - Loaded Modules: 1,414 modules (1,056,448 lines of code)
  - Loop overhead: **`~310 ms`** per run.
- **Savings**: **`~6.2 seconds`** of pure file-open and line-reading
operations saved across 22 runs.
- **Small Packages (`google-auth`)**:
  - Loaded Modules: 17 modules (9,186 lines of code)
  - Loop overhead: **`~2.8 ms`** per run.
  
  Fixes #17792

---------

Co-authored-by: Anthonios Partheniou <partheniou@google.com>

Author: hebaalazzeh

.github/workflows/import-profiler.yml

@@ -28,6 +28,11 @@ jobs:
         python-version: "3.15"
         allow-prereleases: true
         cache: 'pip'
+    - name: Run import profiler unit tests
+      run: |
+        python -m pip install --upgrade pip
+        pip install pytest pytest-cov setuptools
+        pytest scripts/import_profiler/test_profiler.py --cov=profiler --cov-report=term-missing --cov-fail-under=100
     - name: Run import profiler
       env:
         BUILD_TYPE: presubmit

scripts/import_profiler/profiler.py

@@ -40,7 +40,7 @@ def get_rss_mb():
     except Exception:
         return 0.0
 
-def run_worker(target_module):
+def run_worker(target_module, skip_line_count=False):
     """Performs ONE import and returns metrics."""
     tracemalloc.start()
     importlib.invalidate_caches()
@@ -66,33 +66,36 @@ def run_worker(target_module):
     new_modules = modules_after - modules_before
 
     loaded_lines = 0
-    for m in new_modules:
-        try:
-            file_path = sys.modules[m].__file__
-            if not file_path:
-                continue
+    if not skip_line_count:
+        for m in new_modules:
+            try:
+                file_path = sys.modules[m].__file__
+                if not file_path:
+                    continue
 
-            if file_path.endswith('.pyc'):
-                try:
-                    file_path = importlib.util.source_from_cache(file_path)
-                except ValueError:
-                    # Raised if the .pyc path does not follow standard PEP 3147/488 conventions.
-                    # We pass silently because the unresolved file_path will still end in '.pyc', 
-                    # meaning the subsequent '.endswith('.py')' check will fail and safely skip 
-                    # trying to count lines in a binary file.
-                    pass
-            if file_path.endswith('.py'):
-                try:
-                    with open(file_path, 'r', encoding='utf-8', errors='ignore') as f:
-                        loaded_lines += sum(1 for _ in f)
-                except OSError as e:
-                    print(f"WARNING: Failed to read lines from {file_path}: {e}", file=sys.stderr)
-        except KeyError:
-            # Module disappeared from sys.modules during execution
-            pass
-        except AttributeError:
-            # Module has no __file__ attribute (likely a C-extension or built-in)
-            pass
+                if file_path.endswith('.pyc'):
+                    try:
+                        file_path = importlib.util.source_from_cache(file_path)
+                    except ValueError:
+                        # Raised if the .pyc path does not follow standard PEP 3147/488 conventions.
+                        # We pass silently because the unresolved file_path will still end in '.pyc', 
+                        # meaning the subsequent '.endswith('.py')' check will fail and safely skip 
+                        # trying to count lines in a binary file.
+                        pass
+                if file_path.endswith('.py'):
+                    try:
+                        with open(file_path, 'r', encoding='utf-8', errors='ignore') as f:
+                            loaded_lines += sum(1 for _ in f)
+                    except OSError as e:
+                        print(f"WARNING: Failed to read lines from {file_path}: {e}", file=sys.stderr)
+            except KeyError:
+                # Module disappeared from sys.modules during execution
+                pass
+            except AttributeError:
+                # Module has no __file__ attribute (likely a C-extension or built-in)
+                pass
+    else:
+        loaded_lines = -1
 
     # Output to stdout for the Master to capture
     metrics = {
@@ -106,6 +109,8 @@ def run_worker(target_module):
 
 def _run_worker_and_parse(cmd):
     result = subprocess.run(cmd, capture_output=True, text=True, check=True)
+    if result.stderr.strip():
+        print(result.stderr.strip(), file=sys.stderr)
     try:
         lines = result.stdout.strip().splitlines()
         data = None
@@ -186,6 +191,8 @@ def run_master(iterations, target_module, cpu=0, csv_path=None, clear_cache=True
             cmd += ["taskset", "-c", str(cpu)]
 
         cmd += python_exe + [__file__, "--worker", f"--module={target_module}"]
+        if i > 0:
+            cmd += ["--skip-line-count"]
 
         try:
             data = _run_worker_and_parse(cmd)
@@ -195,11 +202,12 @@ def run_master(iterations, target_module, cpu=0, csv_path=None, clear_cache=True
             print(f"Iteration {i+1}/{iterations} completed in {data['time_ms']:.2f} ms")
             if i > 0 and loaded_modules_val != data["loaded_modules"]:
                 print(f"WARNING: Non-deterministic import behavior! Iteration {i+1} loaded {data['loaded_modules']} modules (expected {loaded_modules_val}).", file=sys.stderr)
-            if i > 0 and loaded_lines_val != data["loaded_lines"]:
-                print(f"WARNING: Non-deterministic import behavior! Iteration {i+1} loaded {data['loaded_lines']} lines (expected {loaded_lines_val}).", file=sys.stderr)
 
             loaded_modules_val = data["loaded_modules"]
-            loaded_lines_val = data["loaded_lines"]
+            if data["loaded_lines"] == -1:
+                data["loaded_lines"] = loaded_lines_val
+            else:
+                loaded_lines_val = data["loaded_lines"]
         except FileNotFoundError as e:
             if cpu != NO_CPU_PINNING and cmd and cmd[0] == "taskset":
                 print("ERROR: 'taskset' command not found. CPU pinning is enabled but taskset is not installed. "
@@ -362,63 +370,64 @@ def run_mprofile(target_module):
     if p.exitcode != 0:
         print(f"Error generating memory snapshot, process exited with code {p.exitcode}", file=sys.stderr)
 
-if __name__ == "__main__":
+def validate_module_name(module_name):
+    """Validates that the input is a structurally valid Python module identifier to prevent arbitrary code execution."""
     import argparse
+    if not all(part.isidentifier() for part in module_name.split('.')):
+        raise argparse.ArgumentTypeError(f"'{module_name}' is not a valid Python module identifier.")
+    return module_name
 
-    def validate_module_name(module_name):
-        """Validates that the input is a structurally valid Python module identifier to prevent arbitrary code execution."""
-        if not all(part.isidentifier() for part in module_name.split('.')):
-            raise argparse.ArgumentTypeError(f"'{module_name}' is not a valid Python module identifier.")
-        return module_name
+def find_module_from_package(pkg):
+    import importlib.metadata
+    import importlib.util
 
-    def find_module_from_package(pkg):
-        import importlib.metadata
-        import importlib.util
+    # 1. Try to use importlib.metadata.files (works for standard installations from PyPI/wheels)
+    try:
+        files = importlib.metadata.files(pkg)
+        if files:
+            init_files = [str(f) for f in files if str(f).endswith('__init__.py') and '__pycache__' not in str(f) and not str(f).startswith('tests/')]
+            if init_files:
+                from pathlib import Path
+                shortest_init = min(init_files, key=lambda p: len(Path(p).parts))
+                parts = Path(shortest_init).parent.parts
+                mod = '.'.join(parts)
+                if importlib.util.find_spec(mod):
+                    return mod
+    except Exception:
+        pass
 
-        # 1. Try to use importlib.metadata.files (works for standard installations from PyPI/wheels)
-        try:
-            files = importlib.metadata.files(pkg)
-            if files:
-                init_files = [str(f) for f in files if str(f).endswith('__init__.py') and '__pycache__' not in str(f) and not str(f).startswith('tests/')]
-                if init_files:
-                    from pathlib import Path
-                    shortest_init = min(init_files, key=lambda p: len(Path(p).parts))
-                    parts = Path(shortest_init).parent.parts
-                    mod = '.'.join(parts)
-                    if importlib.util.find_spec(mod):
-                        return mod
-        except Exception:
-            pass
+    # 2. Try setuptools.find_namespace_packages() in current directory (works for editable installs in source trees)
+    try:
+        import setuptools
+        import os
+        if os.path.exists('setup.py') or os.path.exists('pyproject.toml'):
+            pkgs = setuptools.find_namespace_packages(where='.')
+            for p in sorted(pkgs, key=len):
+                if p in ("google", "google.cloud") or p.startswith("tests"):
+                    continue
+                path = p.replace('.', os.sep)
+                if os.path.isfile(os.path.join(path, '__init__.py')):
+                    if importlib.util.find_spec(p):
+                        return p
+    except Exception:
+        pass
 
-        # 2. Try setuptools.find_namespace_packages() in current directory (works for editable installs in source trees)
+    # 3. Fallback to basic string manipulation heuristics
+    candidates = [
+        pkg.replace('-', '.'),
+        '.'.join(pkg.split('-')[:-1]) + '_' + pkg.split('-')[-1] if '-' in pkg else pkg,
+        pkg.replace('-', '_')
+    ]
+    for mod in candidates:
         try:
-            import setuptools
-            import os
-            if os.path.exists('setup.py') or os.path.exists('pyproject.toml'):
-                pkgs = setuptools.find_namespace_packages(where='.')
-                for p in sorted(pkgs, key=len):
-                    if p in ("google", "google.cloud") or p.startswith("tests"):
-                        continue
-                    path = p.replace('.', os.sep)
-                    if os.path.isfile(os.path.join(path, '__init__.py')):
-                        if importlib.util.find_spec(p):
-                            return p
+            if importlib.util.find_spec(mod):
+                return mod
         except Exception:
             pass
+    return candidates[0]
 
-        # 3. Fallback to basic string manipulation heuristics
-        candidates = [
-            pkg.replace('-', '.'),
-            '.'.join(pkg.split('-')[:-1]) + '_' + pkg.split('-')[-1] if '-' in pkg else pkg,
-            pkg.replace('-', '_')
-        ]
-        for mod in candidates:
-            try:
-                if importlib.util.find_spec(mod):
-                    return mod
-            except Exception:
-                pass
-        return candidates[0]
+if __name__ == "__main__":
+    import argparse
 
     parser = argparse.ArgumentParser(description="Python SDK Import Profiler")
     group = parser.add_mutually_exclusive_group(required=True)
@@ -436,6 +445,7 @@ def find_module_from_package(pkg):
     parser.add_argument("--diff-baseline", help="Path to a baseline CSV file to compare against.")
     parser.add_argument("--diff-threshold", type=float, default=100.0, help="Fail if Median time exceeds baseline Median by this many ms.")
     parser.add_argument("--worker", action="store_true", help=argparse.SUPPRESS)
+    parser.add_argument("--skip-line-count", action="store_true", help=argparse.SUPPRESS)
 
     args = parser.parse_args()
 
@@ -444,7 +454,7 @@ def find_module_from_package(pkg):
         target_module = find_module_from_package(args.package)
 
     if args.worker:
-        run_worker(target_module)
+        run_worker(target_module, skip_line_count=args.skip_line_count)
     elif args.trace:
         if not args.keep_pycache: clean_bytecode()
         run_trace(target_module)