Skip invalid tree-sitter queries in repo map generation

aider/repomap.py

@@ -16,7 +16,7 @@
 from pygments.lexers import guess_lexer_for_filename
 from pygments.token import Token
 from tqdm import tqdm
-from tree_sitter import Query
+from tree_sitter import Query, QueryError
 
 from aider.dump import dump
 from aider.special import filter_important_files
@@ -298,8 +298,12 @@ def get_tags_raw(self, fname, rel_fname):
             return
         tree = parser.parse(bytes(code, "utf-8"))
 
-        # Run the tags queries
-        captures = self._run_captures(Query(language, query_scm), tree.root_node)
+        # Skip files whose installed parser and scm query are out of sync.
+        try:
+            captures = self._run_captures(Query(language, query_scm), tree.root_node)
+        except QueryError as err:
+            print(f"Skipping file {fname}: {err}")
+            return
 
         captures_by_tag = defaultdict(list)
         matches = []

tests/basic/test_repomap.py

@@ -4,8 +4,10 @@
 import time
 import unittest
 from pathlib import Path
+from unittest.mock import patch
 
 import git
+from tree_sitter import QueryError
 
 from aider.dump import dump  # noqa: F401
 from aider.io import InputOutput
@@ -273,6 +275,28 @@ def test_get_repo_map_excludes_added_files(self):
             # close the open cache files, so Windows won't error
             del repo_map
 
+    def test_get_repo_map_skips_query_compile_failures(self):
+        with IgnorantTemporaryDirectory() as temp_dir:
+            broken_file = os.path.join(temp_dir, "broken.py")
+            healthy_file = os.path.join(temp_dir, "healthy.md")
+
+            with open(broken_file, "w", encoding="utf-8") as f:
+                f.write("def broken():\n    return 1\n")
+
+            with open(healthy_file, "w", encoding="utf-8") as f:
+                f.write("# healthy\n")
+
+            io = InputOutput()
+            repo_map = RepoMap(main_model=self.GPT35, root=temp_dir, io=io)
+
+            with patch("aider.repomap.Query", side_effect=QueryError("boom")):
+                result = repo_map.get_repo_map([], [broken_file, healthy_file])
+
+            self.assertIn("broken.py", result)
+            self.assertIn("healthy.md", result)
+
+            del repo_map
+
 
 class TestRepoMapTypescript(unittest.TestCase):
     def setUp(self):