Used isort to sort imports. Also added to the git pre-commit hook.

[python_utils.git] / bootstrap.py

diff --git a/bootstrap.py b/bootstrap.py
index 1b4d4843332effc7683bcbd20a7f9a1432cc299e..035a38eca691ca7255a4ee4bda0eee6da0813119 100644 (file)
--- a/bootstrap.py
+++ b/bootstrap.py
@@ -1,17 +1,20 @@
 #!/usr/bin/env python3
 
 import functools
+import importlib
 import logging
 import os
-from inspect import stack
 import sys
+from inspect import stack
+from typing import List
+
+import config
+import logging_utils
+from argparse_utils import ActionNoYes
 
 # This module is commonly used by others in here and should avoid
 # taking any unnecessary dependencies back on them.
 
-from argparse_utils import ActionNoYes
-import config
-import logging_utils
 
 logger = logging.getLogger(__name__)
 
@@ -57,6 +60,12 @@ args.add_argument(
     default=False,
     help='Should we run cProfile on this code?',
 )
+args.add_argument(
+    '--trace_memory',
+    action=ActionNoYes,
+    default=False,
+    help='Should we record/report on memory utilization?',
+)
 
 original_hook = sys.excepthook
 
@@ -93,7 +102,7 @@ def handle_uncaught_exception(exc_type, exc_value, exc_tb):
                 original_hook(exc_type, exc_value, exc_tb)
 
 
-class ImportInterceptor(object):
+class ImportInterceptor(importlib.abc.MetaPathFinder):
     def __init__(self):
         import collect.trie
 
@@ -114,6 +123,11 @@ class ImportInterceptor(object):
     def should_ignore_filename(self, filename: str) -> bool:
         return 'importlib' in filename or 'six.py' in filename
 
+    def find_module(self, fullname, path):
+        raise Exception(
+            "This method has been deprecated since Python 3.4, please upgrade."
+        )
+
     def find_spec(self, loaded_module, path=None, target=None):
         s = stack()
         for x in range(3, len(s)):
@@ -141,6 +155,9 @@ class ImportInterceptor(object):
         logger.debug(msg)
         print(msg)
 
+    def invalidate_caches(self):
+        pass
+
     def find_importer(self, module: str):
         if module in self.tree_node_by_module:
             node = self.tree_node_by_module[module]
@@ -160,7 +177,7 @@ import_interceptor = None
 for arg in sys.argv:
     if arg == '--audit_import_events':
         import_interceptor = ImportInterceptor()
-        sys.meta_path = [import_interceptor] + sys.meta_path
+        sys.meta_path.insert(0, import_interceptor)
 
 
 def dump_all_objects() -> None:
@@ -221,11 +238,25 @@ def initialize(entry_point):
         else:
             config.parse(None)
 
+        if config.config['trace_memory']:
+            import tracemalloc
+
+            tracemalloc.start()
+
         # Initialize logging... and log some remembered messages from
         # config module.
         logging_utils.initialize_logging(logging.getLogger())
         config.late_logging()
 
+        # Maybe log some info about the python interpreter itself.
+        logger.debug(
+            f'Platform: {sys.platform}, maxint=0x{sys.maxsize:x}, byteorder={sys.byteorder}'
+        )
+        logger.debug(f'Python interpreter version: {sys.version}')
+        logger.debug(f'Python implementation: {sys.implementation}')
+        logger.debug(f'Python C API version: {sys.api_version}')
+        logger.debug(f'Python path: {sys.path}')
+
         # Allow programs that don't bother to override the random seed
         # to be replayed via the commandline.
         import random
@@ -265,6 +296,14 @@ def initialize(entry_point):
 
         logger.debug(f'{entry_point.__name__} (program entry point) returned {ret}.')
 
+        if config.config['trace_memory']:
+            snapshot = tracemalloc.take_snapshot()
+            top_stats = snapshot.statistics('lineno')
+            print()
+            print("--trace_memory's top 10 memory using files:")
+            for stat in top_stats[:10]:
+                print(stat)
+
         if config.config['dump_all_objects']:
             dump_all_objects()