ai-agent-book 精选快照(<2MB 代码与文档,来自 github.com/bojieli/ai-agent-book)
Build latest book artifacts / build (push) Canceled after 0s
dependency resolution / resolve (3.11) (push) Canceled after 0s
dependency resolution / resolve (3.13) (push) Canceled after 0s
deploy-pages / build (push) Canceled after 0s
deploy-pages / deploy (push) Canceled after 0s
i18n consistency check / check (push) Canceled after 0s
provider adoption tests / test (chapter2/context-compression) (push) Canceled after 0s
provider adoption tests / test (chapter2/prompt-injection) (push) Canceled after 0s
provider adoption tests / test (chapter2/system-hint) (push) Canceled after 0s
provider adoption tests / test (chapter3/log-sanitization) (push) Canceled after 0s
web-search-agent tests / test (push) Canceled after 0s
web-search-agent tests / agentbook (push) Canceled after 0s

This commit is contained in:
2026-08-20 13:12:50 +00:00
commit b119135836
10275 changed files with 3284984 additions and 0 deletions
@@ -0,0 +1,66 @@
"""
Tests for the SchemaOptimizer to ensure it correctly processes and
optimizes the schemas for agent actions without losing information.
"""
from pydantic import BaseModel
from browser_use.agent.views import AgentOutput
from browser_use.llm.schema import SchemaOptimizer
from browser_use.tools.service import Tools
class ProductInfo(BaseModel):
"""A sample structured output model with multiple fields."""
price: str
title: str
rating: float | None = None
def test_optimizer_preserves_all_fields_in_structured_done_action():
"""
Ensures the SchemaOptimizer does not drop fields from a custom structured
output model when creating the schema for the 'done' action.
This test specifically checks for a bug where fields were being lost
during the optimization process.
"""
# 1. Setup a tools with a custom output model, simulating an Agent
# being created with an `output_model_schema`.
tools = Tools(output_model=ProductInfo)
# 2. Get the dynamically created AgentOutput model, which includes all registered actions.
ActionModel = tools.registry.create_action_model()
agent_output_model = AgentOutput.type_with_custom_actions(ActionModel)
# 3. Run the schema optimizer on the agent's output model.
optimized_schema = SchemaOptimizer.create_optimized_json_schema(agent_output_model)
# 4. Find the 'done' action schema within the optimized output.
# The path is properties -> action -> items -> anyOf -> [schema with 'done'].
done_action_schema = None
actions_schemas = optimized_schema.get('properties', {}).get('action', {}).get('items', {}).get('anyOf', [])
for action_schema in actions_schemas:
if 'done' in action_schema.get('properties', {}):
done_action_schema = action_schema
break
# 5. Assert that the 'done' action schema was successfully found.
assert done_action_schema is not None, "Could not find 'done' action in the optimized schema."
# 6. Navigate to the schema for our custom data model within the 'done' action.
# The path is properties -> done -> properties -> data -> properties.
done_params_schema = done_action_schema.get('properties', {}).get('done', {})
structured_data_schema = done_params_schema.get('properties', {}).get('data', {})
final_properties = structured_data_schema.get('properties', {})
# 7. Assert that the set of fields in the optimized schema matches the original model's fields.
original_fields = set(ProductInfo.model_fields.keys())
optimized_fields = set(final_properties.keys())
assert original_fields == optimized_fields, (
f"Field mismatch between original and optimized structured 'done' action schema.\n"
f'Missing from optimized: {original_fields - optimized_fields}\n'
f'Unexpected in optimized: {optimized_fields - original_fields}'
)