chore: add codespell pre-commit hook to catch typos

.pre-commit-config.yaml

@@ -72,3 +72,11 @@ repos:
         exclude: (?i)SKILL\.md$
         additional_dependencies:
           - mdformat-gfm
+  - repo: https://github.com/codespell-project/codespell
+    rev: v2.4.2
+    hooks:
+      - id: codespell
+        # Configuration (skip globs, ignored words) lives in the
+        # [tool.codespell] table in pyproject.toml.
+        additional_dependencies:
+          - tomli

contributing/samples/integrations/data_agent/agent.py

@@ -36,7 +36,7 @@
 CREDENTIALS_TYPE = None
 
 if CREDENTIALS_TYPE == AuthCredentialTypes.OAUTH2:
-  # Initiaze the tools to do interactive OAuth
+  # Initialize the tools to do interactive OAuth
   # The environment variables OAUTH_CLIENT_ID and OAUTH_CLIENT_SECRET
   # must be set
   credentials_config = DataAgentCredentialsConfig(

contributing/samples/integrations/oauth_calendar_agent/agent.py

@@ -42,7 +42,7 @@
 SCOPES = ["https://www.googleapis.com/auth/calendar"]
 
 calendar_toolset = CalendarToolset(
-    # you can also replace below customized `list_calendar_events` with build-in
+    # you can also replace below customized `list_calendar_events` with built-in
     # google calendar tool by adding `calendar_events_list` in the filter list
     client_id=oauth_client_id,
     client_secret=oauth_client_secret,

contributing/samples/integrations/spanner_rag_agent/README.md

@@ -296,7 +296,7 @@ There are a few options to perform similarity search:
            "from the Spanner database."
        ),
        instruction="""
-       You are a helpful assistant that answers user questions to find the most relavant information from a Spanner database.
+       You are a helpful assistant that answers user questions to find the most relevant information from a Spanner database.
        1. Always use the `similarity_search` tool to find relevant information.
        2. If no relevant information is found, say you don't know.
        3. Present all the relevant information naturally and well formatted in your response.

contributing/samples/workflows/route/agent.py

@@ -57,7 +57,7 @@ def route_on_category(category: InputCategory):
 
 def handle_other():
   yield Event(
-      message="Sorry I can only anwer questions or comment on statements."
+      message="Sorry I can only answer questions or comment on statements."
   )
 
 

pyproject.toml

@@ -296,6 +296,17 @@ line_length = 200
 single_line_exclusions = []
 known_third_party = [ "a2a", "google.adk" ]
 
+[tool.codespell]
+# Real words/identifiers that codespell misreads as typos:
+#   hel/serie/strin -> substrings in test fixtures; te -> local variable;
+#   rouge -> the ROUGE metric; unparseable -> valid spelling variant;
+#   re-use/re-used -> intentional hyphenation.
+ignore-words-list = "hel,serie,strin,te,rouge,unparseable,re-use,re-used"
+# CHANGELOG.md is generated from commit messages; lockfiles, notebooks, JSON
+# fixtures, bundled JS/source maps, and the vendored CLI browser bundle are
+# generated or data files, not prose we own.
+skip = "*CHANGELOG.md,*.lock,*.ipynb,*.json,*.js,*.map,*/cli/browser/*"
+
 [tool.mypy]
 mypy_path = [ "src" ]
 exclude = [ "contributing/samples/", "tests/" ]

src/google/adk/evaluation/rubric_based_tool_use_quality_v1.py

@@ -130,7 +130,7 @@ class RubricBasedToolUseV1Evaluator(RubricBasedEvaluator):
   """An Evaluator for rubric based assessment of the agent's usage of Tools.
 
   Example: Lets take an example of a Weather Agent that has access to two tools:
-  1: GeoCoding Tool: Coverts a city name, address or zip code into geographic
+  1: GeoCoding Tool: Converts a city name, address or zip code into geographic
   coordinates.
   2: GetWeather Tool: Gets weather for the next 10 days for the given geographic
   coordinates.

src/google/adk/runners.py

@@ -1631,7 +1631,7 @@ async def run_live(
     *   **Other Control Events:** Most control events are saved.
 
     **Events Saved to the Session:**
-    *   **Live Model Audio Events with File Data:** Both input and ouput audio
+    *   **Live Model Audio Events with File Data:** Both input and output audio
         data are aggregated into an audio file saved into artifacts. The
         reference to the file is saved as event in the `file_data` to session
         if RunConfig.save_live_model_audio_to_session is True.

src/google/adk/tools/application_integration_tool/clients/connections_client.py

@@ -247,7 +247,7 @@ def get_connector_base_spec() -> Dict[str, Any]:
                 "host": {
                     "type": "string",
                     "default": "",
-                    "description": "Host name incase of tls service directory",
+                    "description": "Host name in case of tls service directory",
                 },
                 "entity": {
                     "type": "string",

src/google/adk/workflow/_function_node.py

@@ -476,7 +476,7 @@ def model_copy(
     # If the wrapped function is a bound method of a Node, we need to clone
     # the Node and re-bind the function to the new instance.
     # This is needed if the function is referring to params like 'name' from the "self" reference.
-    # Like Workflow or LLM use that name for event node_paths or retreving session events.
+    # Like Workflow or LLM use that name for event node_paths or retrieving session events.
     func = self._func
     if inspect.ismethod(func) and isinstance(
         getattr(func, '__self__', None), BaseNode

tests/unittests/a2a/executor/test_a2a_agent_executor.py

@@ -1154,7 +1154,7 @@ async def mock_run_async(**kwargs):
     )
     self.mock_runner._new_invocation_context.return_value = Mock()
 
-    # We patch TaskResultAggregator just to avoid other errors and simplfy
+    # We patch TaskResultAggregator just to avoid other errors and simplify
     with patch(
         "google.adk.a2a.executor.a2a_agent_executor.TaskResultAggregator"
     ) as mock_agg_class:

tests/unittests/integrations/bigquery/test_bigquery_query_tool.py

@@ -2183,7 +2183,7 @@ def test_tool_call_doesnt_change_global_settings(tool_call):
     # Call the tool
     result = tool_call(settings, tool_context)
 
-    # Test successfull executeion of the tool
+    # Test successful executeion of the tool
     assert result == {"status": "SUCCESS", "rows": []}
 
     # Test settings write mode after