docs: fixing multiple typos

Merge #4221

### Link to Issue or Description of Change

N/A

**2. Or, if no issue exists, describe the change:**

Fixing various typos in .py files to improve quality: see commit diffs for details

**Problem:**

See above

**Solution:**
Fixing the typos

### Testing Plan

N/A

**Unit Tests:**

N/A

**Manual End-to-End (E2E) Tests:**

N/A

### Checklist

- [X] I have read the [CONTRIBUTING.md](https://github.com/google/adk-python/blob/main/CONTRIBUTING.md) document.
- [X] I have performed a self-review of my own code.
- [N/A] I have commented my code, particularly in hard-to-understand areas.
- [N/A ] I have added tests that prove my fix is effective or that my feature works.
- [N/A] New and existing unit tests pass locally with my changes.
- [N/A] I have manually tested my changes end-to-end.
- [N/A ] Any dependent changes have been merged and published in downstream modules.

### Additional context

N/A

COPYBARA_INTEGRATE_REVIEW=#4221 from didier-durand:fix-typos-e 1bcc8e0
PiperOrigin-RevId: 859332402

Author: didier-durand

src/google/adk/agents/remote_a2a_agent.py

@@ -134,7 +134,7 @@ def __init__(
     Args:
       name: Agent name (must be unique identifier)
       agent_card: AgentCard object, URL string, or file path string
-      description: Agent description (auto-populated from card if empty)
+      description: Agent description (autopopulated from card if empty)
       httpx_client: Optional shared HTTP client (will create own if not
         provided) [deprecated] Use a2a_client_factory instead.
       timeout: HTTP timeout in seconds

src/google/adk/cli/adk_web_server.py

@@ -395,7 +395,7 @@ def _setup_gcp_telemetry(
           # TODO - use trace_to_cloud here as well once otel_to_cloud is no
           # longer experimental.
           enable_cloud_tracing=True,
-          # TODO - reenable metrics once errors during shutdown are fixed.
+          # TODO - re-enable metrics once errors during shutdown are fixed.
           enable_cloud_metrics=False,
           enable_cloud_logging=True,
           google_auth=(credentials, project_id),

src/google/adk/cli/cli_tools_click.py

@@ -1299,7 +1299,7 @@ def cli_web(
 ):
   """Starts a FastAPI server with Web UI for agents.
 
-  AGENTS_DIR: The directory of agents, where each sub-directory is a single
+  AGENTS_DIR: The directory of agents, where each subdirectory is a single
   agent, containing at least `__init__.py` and `agent.py` files.
 
   Example:
@@ -1366,7 +1366,7 @@ async def _lifespan(app: FastAPI):
 
 @main.command("api_server")
 @feature_options()
-# The directory of agents, where each sub-directory is a single agent.
+# The directory of agents, where each subdirectory is a single agent.
 # By default, it is the current working directory
 @click.argument(
     "agents_dir",
@@ -1401,7 +1401,7 @@ def cli_api_server(
 ):
   """Starts a FastAPI server for agents.
 
-  AGENTS_DIR: The directory of agents, where each sub-directory is a single
+  AGENTS_DIR: The directory of agents, where each subdirectory is a single
   agent, containing at least `__init__.py` and `agent.py` files.
 
   Example:

src/google/adk/evaluation/eval_config.py

@@ -89,7 +89,7 @@ class EvalConfig(BaseModel):
 In the sample below, `tool_trajectory_avg_score`, `response_match_score` and
 `final_response_match_v2` are the standard eval metric names, represented as
 keys in the dictionary. The values in the dictionary are the corresponding
-criterions. For the first two metrics, we use simple threshold as the criterion,
+criteria. For the first two metrics, we use simple threshold as the criterion,
 the third one uses `LlmAsAJudgeCriterion`.
 {
   "criteria": {

src/google/adk/evaluation/rubric_based_evaluator.py

@@ -93,7 +93,7 @@ class PerInvocationResultsAggregator(abc.ABC):
   """An interface for aggregating per invocation samples.
 
   AutoRaters that are backed by an LLM are known to have certain degree of
-  unreliabilty to their responses. In order to counter that we sample the
+  unreliability to their responses. In order to counter that we sample the
   autorater more than once for a single invocation.
 
   The aggregator helps convert those multiple samples into a single result.
@@ -419,7 +419,7 @@ def aggregate_per_invocation_samples(
     """Returns a combined result by aggregating multiple samples for the same invocation.
 
     AutoRaters that are backed by an LLM are known to have certain degree of
-    unreliabilty to their responses. In order to counter that we sample the
+    unreliability to their responses. In order to counter that we sample the
     autorater more than once for a single invocation.
 
     The aggregator helps convert those multiple samples into a single result.

src/google/adk/evaluation/simulation/per_turn_user_simulator_quality_v1.py

@@ -58,7 +58,7 @@
 
 # Definition of Conversation History
 The Conversation History is the actual dialogue between the User Simulator and the Agent.
-The Conversation History may not be complete, but the exsisting dialogue should adhere to the Conversation Plan.
+The Conversation History may not be complete, but the existing dialogue should adhere to the Conversation Plan.
 The Conversation History may contain instances where the User Simulator troubleshoots an incorrect/inappropriate response from the Agent in order to enforce the Conversation Plan.
 The Conversation History is finished only when the User Simulator outputs `{stop_signal}` in its response. If this token is missing, the conversation between the User Simulator and the Agent has not finished, and more turns can be generated.
 
@@ -171,7 +171,7 @@ def _parse_llm_response(response: str) -> Label:
       response,
   )
 
-  # If there was not match for "is_valid", return NOT_FOUND
+  # If there was no match for "is_valid", return NOT_FOUND
   if is_valid_match is None:
     return Label.NOT_FOUND
 

src/google/adk/flows/llm_flows/functions.py

@@ -411,7 +411,7 @@ async def _run_with_trace():
       function_response = altered_function_response
 
     if tool.is_long_running:
-      # Allow long running function to return None to not provide function
+      # Allow long-running function to return None to not provide function
       # response.
       if not function_response:
         return None
@@ -894,7 +894,7 @@ def find_matching_function_call(
     )
     for i in range(len(events) - 2, -1, -1):
       event = events[i]
-      # looking for the system long running request euc function call
+      # looking for the system long-running request euc function call
       function_calls = event.get_function_calls()
       if not function_calls:
         continue

src/google/adk/memory/vertex_ai_rag_memory_service.py

@@ -52,7 +52,7 @@ def __init__(
           or ``{rag_corpus_id}``
         similarity_top_k: The number of contexts to retrieve.
         vector_distance_threshold: Only returns contexts with vector distance
-          smaller than the threshold..
+          smaller than the threshold.
     """
     self._vertex_rag_store = types.VertexRagStore(
         rag_resources=[

src/google/adk/models/google_llm.py

@@ -58,7 +58,7 @@
 
 
 class _ResourceExhaustedError(ClientError):
-  """Represents an resources exhausted error received from the Model."""
+  """Represents a resources exhausted error received from the Model."""
 
   def __init__(
       self,

src/google/adk/models/lite_llm.py

@@ -152,7 +152,7 @@
 def _ensure_litellm_imported() -> None:
   """Imports LiteLLM with safe defaults.
 
-  LiteLLM defaults to DEV mode, which auto-loads a local `.env` at import time.
+  LiteLLM defaults to DEV mode, which autoloads a local `.env` at import time.
   ADK should not implicitly load `.env` just because LiteLLM is installed.
 
   Users can opt into LiteLLM's default behavior by setting LITELLM_MODE=DEV.