* add pyinstrument and fastapi wrapper
* Truncate vertex build responses
* Truncate vertex response and transactions
* [autofix.ci] apply automated fixes
* Update test for vertex response structure to reflect new truncation logic
* Refactor profiling middleware to use environment variable for configuration
* Update test for vertex response structure to correct truncation logic
* move pyinstrument to dev dependencies
* Update serialization test to include max_items parameter for list truncation
* remove profiling from this PR, fix constants
* [autofix.ci] apply automated fixes
* Add asyncio configuration options to pyproject.toml
---------
Co-authored-by: autofix-ci[bot] <114827586+autofix-ci[bot]@users.noreply.github.com>
Co-authored-by: Gabriel Luiz Freitas Almeida <gabriel@langflow.org>
* Update mcp_sse.py
Uses Python's built-in asyncio.timeout() context manager
Properly handles timeout exceptions
Maintains the same functionality but with correct async context management
* [autofix.ci] apply automated fixes
* add asyncio
+clean up comment
* [autofix.ci] apply automated fixes
* missing arg_schema in Tool
Missing args_schema inside cause that tools are generated without input schema and are not able to be properly executed as agent know tool, but dost know what input field tool have.
Same problem looks to be in MCP STDIO.
* fix Ruff Check
Line 56:
Error: B904 Within an `except` clause, raise exceptions with `raise ... from err` or `raise ... from None` to distinguish them from errors in exception handling
TRY003 Avoid specifying long messages outside the exception class
EM102 Exception must not use an f-string literal, assign to variable first
* [autofix.ci] apply automated fixes
* remove asyncio.timeout
Remove asyncio.timeout() (not valid for Py3.10) and replace it by asyncio.wait_for()
* [autofix.ci] apply automated fixes
* Ruff (TRY300)
Move return response.tools inside an else block. This makes it clearer that tools are returned only if the connection is successful, and not if a TimeoutError occurs.
* fix: add session initialization check in MCPSseClient
Added a check to ensure the session is initialized before attempting to list tools, raising a ValueError with a descriptive message if the session is None. This improves error handling and robustness of the MCPSseClient class.
* [autofix.ci] apply automated fixes
---------
Co-authored-by: autofix-ci[bot] <114827586+autofix-ci[bot]@users.noreply.github.com>
Co-authored-by: Sebastián Estévez <estevezsebastian@gmail.com>
Co-authored-by: Gabriel Luiz Freitas Almeida <gabriel@langflow.org>
* Bump ruff version to 0.9.7
* [autofix.ci] apply automated fixes
* refactor: Update Gmail Agent starter project to Langflow 1.2.0
This commit updates the Gmail Agent starter project with new node IDs and configuration reflecting Langflow version 1.2.0. Changes include:
- Updated node identifiers for Agent, ChatInput, ChatOutput, and ComposioAPI
- Refreshed project metadata and version information
- Minor adjustments to node configurations and viewport settings
* [autofix.ci] apply automated fixes
* fix: Revert Gmail Agent name to original
* fix: Remove trailing comma in Gmail Agent JSON
* feat: Add tags to Gmail Agent starter project
---------
Co-authored-by: autofix-ci[bot] <114827586+autofix-ci[bot]@users.noreply.github.com>
Co-authored-by: Gabriel Luiz Freitas Almeida <gabriel@langflow.org>
* feat: Implement serialization functions for various data types and add a unified serialize method
* feat: Enhance serialization by adding support for primitive types, enums, and generic types
* fix: Update Pinecone integration to use VectorStore and handle import errors gracefully
* test: Add hypothesis-based tests for serialization functions across various data types
* refactor: Replace custom serialization logic with unified serialize function for consistency and maintainability
* refactor: Replace recursive serialization function with unified serialize method for improved clarity and maintainability
* refactor: Replace custom serialization logic with unified serialize function for improved consistency and clarity
* refactor: Enhance serialization logic by adding instance handling and streamlining type checks
* refactor: Remove custom dictionary serialization from ResultDataResponse for streamlined handling
* refactor: Enhance serialization in ResultDataResponse by adding max_items_length for improved handling of outputs, logs, messages, and artifacts
* refactor: Move MAX_ITEMS_LENGTH and MAX_TEXT_LENGTH constants to serialization module for better organization
* refactor: Simplify message serialization in Log model by utilizing unified serialize function
* refactor: Remove unnecessary pytest marker from TestSerializationHypothesis class
* optimize _serialize_bytes
Co-authored-by: codeflash-ai[bot] <148906541+codeflash-ai[bot]@users.noreply.github.com>
* feat: Add support for numpy integer type serialization
* feat: Enhance serialization with support for pandas and numpy types
* test: Add comprehensive serialization tests for numpy and pandas types
* fix: Update _serialize_dispatcher to return string representation for unsupported types
* fix: Update _serialize_dispatcher to return the object directly instead of its string representation
* optmize conditional
Co-authored-by: codeflash-ai[bot] <148906541+codeflash-ai[bot]@users.noreply.github.com>
* optimize length check
Co-authored-by: codeflash-ai[bot] <148906541+codeflash-ai[bot]@users.noreply.github.com>
* fix: Update string and list truncation to include ellipsis for clarity
* fix: Update _serialize_primitive to exclude string type from primitive handling
* feat: Enhance serialization to handle numpy types and introduce unserializable sentinel
* fix: Update test cases for serialization of numpy boolean values for consistency
---------
Co-authored-by: codeflash-ai[bot] <148906541+codeflash-ai[bot]@users.noreply.github.com>
