Skip to content

Commit 4724b5e

Browse files
yanmxayanmxa
committed
change to tool call item
Signed-off-by: myan <myan@redhat.com> Adds example for financial agent Indentation for mkdocs.yml feat: add voice pipeline support > Co-authored-by: rm@openai.com fix tests update tests v0.0.6 (voice support) [doc] fix invalid imports include reference to new audio span related concepts removing erroneous changes fixing whitespace fixing object path Make the optional dependency installation compatible with zsh fixed from agents.voice import SingleAgentVoiceWorkflow, VoicePipeline in quickstart.md from agents.voice import SingleAgentVoiceWorkflow, VoicePipeline, remove extra "," from the first line docs: add Langfuse to tracing documentation Remove redundant weaker tracing assertions lint Create py.typed - Ensure library is properly typehinted More fetch_normalized_spans More fetch_normalized_spans More fetch_normalized_spans More fetch_normalized_spans empty assertions More fetch_normalized_spans fix line in guardrails Read tracing API data lazily Add Weights & Biases to tracing docs Fix circular dependency in voice streamed example by renaming agents.py to my_workflow.py Fix type-checking for circular dependency in voice streamed example Fix circular dependency in voice streamed example with cleaner import pattern fix annotation numbering in context management Fix inconsistent numbering between code and explanatory annotations in the context management documentation. Fix voice pipeline code examples in quickstart docs Update issues.yml Update quickstart.md Improve tracing error messages fix typos in /examples fix typos in src/ fix typos in tests revert src/ change per request chore: ignore PyCharm .idea/ directory Uncomment .idea/ directory in .gitignore to ensure PyCharm IDE project configuration files are excluded from version control. [0/n] Only run tests on py3.9, not mypy We don't need to run mypy on 3.9 anyway. Also it causes issues with the rest of this stack. run checks Fix potential infinite tool call loop by resetting tool_choice after tool execution Rename test file for better clarity Revert test file name to match project naming style Update documentation for tool_choice auto-reset feature fix: optimize tool_choice reset logic and fix lint errors - Refactor tool_choice reset to target only problematic edge cases - Replace manual ModelSettings recreation with dataclasses.replace - Fix line length and error handling lint issues in tests docs: Update tool_choice reset documentation to match implementation fix: prevent modifying the original agent's model_settings This fixes the issue where the original agent's model_settings was being directly modified during the tool choice reset process. The original implementation caused the agent's tool_choice to unintentionally reset to "auto" for subsequent runs, which could be unexpected behavior. The fix creates new copies of the agent and model settings objects using dataclasses.replace() instead of modifying the original objects. This ensures that the tool choice reset is limited to the current run only, maintaining the expected behavior for sequential runs with the same agent. Addresses feedback from @baderalfahad about the agent instance being modified when it should maintain its original state between runs. refactor: streamline tool_choice reset logic This update moves the tool_choice reset logic to a more appropriate location within the RunImpl class, ensuring that the original agent's model_settings remains unmodified during the reset process. The logic now checks for problematic scenarios before creating a modified copy of the agent's settings, maintaining expected behavior across sequential runs. This change enhances clarity and efficiency in handling tool choices. Addresses previous feedback regarding the modification of the agent instance and improves the overall structure of the reset logic. refactor: improve comments for clarity in tool_choice reset logic Fix parallel_tool_calls when False [1/n] Add MCP types to the SDK 1. Add the MCP dep for python 3.10, since it doesn't support 3.9 and below 2. Create MCPServer, which is the agents SDK representation of an MCP server 3. Create implementations for HTTP-SSE and StdIO servers, directly copying the [MCP SDK example](https://github.com/modelcontextprotocol/python-sdk/blob/main/examples/clients/simple-chatbot/mcp_simple_chatbot/main.py) 4. Add a util to transform MCP tools into Agent SDK tools Note: I added optional caching support to the servers. That way, if you happen to know a server's tools don't change, you can just cache them. Checks pass. I added tests at the end of the stack. [1/n] Add MCP types to the SDK 1. Add the MCP dep for python 3.10, since it doesn't support 3.9 and below 2. Create MCPServer, which is the agents SDK representation of an MCP server 3. Create implementations for HTTP-SSE and StdIO servers, directly copying the [MCP SDK example](https://github.com/modelcontextprotocol/python-sdk/blob/main/examples/clients/simple-chatbot/mcp_simple_chatbot/main.py) 4. Add a util to transform MCP tools into Agent SDK tools Note: I added optional caching support to the servers. That way, if you happen to know a server's tools don't change, you can just cache them. Checks pass. I added tests at the end of the stack. [2/n] Add MCP support to Runner This enables users to **use** MCP inside the SDK. 1. You add a list of MCP servers to `Agent`, via `mcp_server=[...]` 2. When an agent runs, we look up its MCP tools and add them to the list of tools. 3. When a tool call occurs, we call the relevant MCP server. Notes: 1. There's some refactoring to make sure we send the full list of tools to the Runner/Model etc. 2. Right now, you could have a locally defined tool that conflicts with an MCP defined tool. I didn't add errors for that, will do in a followup. See unit tests. Also has an end to end example next PR. [3/n] Add an MCP stdio example Spins up a stdio server with some local files, then asks the model questions. Run the example, see it work. [4/n] Add docs for MCP Just adding docs. - Make the reset behavior on tool use configurable 1. Making it configurable on the agent. 2. Doing bookkeeping in the Runner to track this, to prevent mutating agents. 3. Not resetting the global tool choice in RunConfig. Unit tests. . fix(examples): make sure audio playback finishes Add graphviz as a dependency in pyproject.toml Add visualization functions for agents using Graphviz Add unit tests for visualization functions in test_visualizations.py Add documentation and example for agent visualization using Graphviz Linting feat: add visualization functions for agent graphs refactor: move graphviz dependency to visualization section feat: add documentation for agent visualization using Graphviz feat: add visualization module for agent graphs using Graphviz rename: test_visualization.py style: improve code formatting and readability in visualization functions style: improve string formatting style: improve string formatting in visualization functions feat: add optional dependency for visualization using Graphviz feat: enhance visualization functions with optional type hints and improved handling of agents and handoffs refactor: clean up visualization functions by removing unused nodes and improving type hints Add visualization optional dependency for graphviz Rename visualization dependency to viz in pyproject.toml Refactor visualization node label formatting in get_all_nodes function Update installation instructions for visualization dependency to use 'viz' Update visualization.md Update pyproject.toml Add graphviz as a dependency and update import statements Add visualization.md to navigation in mkdocs.yml Refactor get_all_edges function to remove unused parent parameter Add Jupyter Notebook files to .gitignore Add start and end nodes to graph visualization and update edge generation Refactor visualization functions to improve formatting and streamline edge generation Fix type ignore comment for agent check in get_all_edges function linting Remove Jupyter Notebook files from .gitignore [5/n] MCP tracing Adds tracing and tests for tracing. - Tools are added to the agents - Theres a span for the mcp tools lookup - Functions have MCP data Unit tests . adding Git MCP server example fixing lint issues added readme, fixed typo delete config Signed-off-by: myan <myan@redhat.com> change the tool into action which contain the tool_call Signed-off-by: myan <myan@redhat.com> action Signed-off-by: myan <myan@redhat.com> fix the error Signed-off-by: myan <myan@redhat.com> fix the teset Signed-off-by: myan <myan@redhat.com> remvoe Signed-off-by: myan <myan@redhat.com>
1 parent e474d0f commit 4724b5e

40 files changed

+18011
-68
lines changed

‎.github/workflows/tests.yml

Lines changed: 3 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -10,6 +10,9 @@ on:
1010
env:
1111
UV_FROZEN: "1"
1212

13+
env:
14+
UV_FROZEN: "1"
15+
1316
jobs:
1417
lint:
1518
runs-on: ubuntu-latest

‎.gitignore

Lines changed: 4 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -141,5 +141,9 @@ cython_debug/
141141
.ruff_cache/
142142

143143
# PyPI configuration file
144+
<<<<<<< HEAD
144145
.pypirc
145146
.aider*
147+
=======
148+
.pypirc
149+
>>>>>>> da5af63 (change to tool call item)

0 commit comments

Comments
 (0)