Don't traverse directories that are excluded (#44)

The current code traverses the entire directory tree and checks each
file against the exclude list.

This pull request checks each directory against the excluded list and
skips them.

This improves start up times when there are large directory trees
excluded. This is common in javascript/node projects that have a lot of
dependencies.

I dont actually know what I'm doing in python. Claude wrote this. It
seems to work correctly on my machine.

src/claudesync/utils.py

@@ -188,11 +188,18 @@ def get_local_files(local_path):
     files = {}
     exclude_dirs = {".git", ".svn", ".hg", ".bzr", "_darcs", "CVS", "claude_chats"}
 
-    for root, dirs, filenames in os.walk(local_path):
-        dirs[:] = [d for d in dirs if d not in exclude_dirs]
+    for root, dirs, filenames in os.walk(local_path, topdown=True):
         rel_root = os.path.relpath(root, local_path)
         rel_root = "" if rel_root == "." else rel_root
 
+        # Filter out directories before traversing
+        dirs[:] = [
+            d for d in dirs
+            if d not in exclude_dirs
+            and not (gitignore and gitignore.match_file(os.path.join(rel_root, d)))
+            and not (claudeignore and claudeignore.match_file(os.path.join(rel_root, d)))
+        ]
+
         for filename in filenames:
             rel_path = os.path.join(rel_root, filename)
             full_path = os.path.join(root, filename)