From e49bfaa2229e2ce27b7a212ad3ebe378dc17f613 Mon Sep 17 00:00:00 2001
From: Jacob Walls <jacobtylerwalls@gmail.com>
Date: Sun, 23 Apr 2023 10:37:15 -0400
Subject: Reduce file system access in `ast_from_file()` (#2135)

get_source_file() is needed to resolve relative to absolute paths,
but is not needed before getting a cache hit. This had the potential
to issue tens of thousands of repetitive os.path.exists() calls.
---
 astroid/manager.py  | 18 +++++++++++++-----
 astroid/modutils.py |  5 ++---
 2 files changed, 15 insertions(+), 8 deletions(-)

(limited to 'astroid')

diff --git a/astroid/manager.py b/astroid/manager.py
index 1f6ef482..7f62fd42 100644
--- a/astroid/manager.py
+++ b/astroid/manager.py
@@ -101,16 +101,24 @@ class AstroidManager:
         source: bool = False,
     ) -> nodes.Module:
         """Given a module name, return the astroid object."""
-        try:
-            filepath = get_source_file(filepath, include_no_ext=True)
-            source = True
-        except NoSourceFile:
-            pass
         if modname is None:
             try:
                 modname = ".".join(modpath_from_file(filepath))
             except ImportError:
                 modname = filepath
+        if (
+            modname in self.astroid_cache
+            and self.astroid_cache[modname].file == filepath
+        ):
+            return self.astroid_cache[modname]
+        # Call get_source_file() only after a cache miss,
+        # since it calls os.path.exists().
+        try:
+            filepath = get_source_file(filepath, include_no_ext=True)
+            source = True
+        except NoSourceFile:
+            pass
+        # Second attempt on the cache after get_source_file().
         if (
             modname in self.astroid_cache
             and self.astroid_cache[modname].file == filepath
diff --git a/astroid/modutils.py b/astroid/modutils.py
index 266344a8..4e8d8f9f 100644
--- a/astroid/modutils.py
+++ b/astroid/modutils.py
@@ -488,9 +488,8 @@ def get_module_files(
 
 def get_source_file(filename: str, include_no_ext: bool = False) -> str:
     """Given a python module's file name return the matching source file
-    name (the filename will be returned identically if it's already an.
-
-    absolute path to a python source file...)
+    name (the filename will be returned identically if it's already an
+    absolute path to a python source file).
 
     :param filename: python module's file name
 
-- 
cgit v1.2.1