mirror of
https://github.com/deepset-ai/haystack.git
synced 2025-06-26 22:00:13 +00:00
9ae76e1653
* Starting property schema refactor * Adding more tests * More tests * Handle null type explicitly * More updates of tests to accomodate Optional properly * Fix more tests * Remove unecessary check * Some cleanup * Update test * Add reno * Fix typing * Add license header * Use docstrings of dataclasses in parameter spec generation * More tests of Haystack dataclass types * Properly handle Sequence * Fix license header * Update OpenAI tests to add more complicated tool parameter signature * Properly set required for dataclasses * Add integration test for azure that includes additionalProperties * Add more complicated integration test for HuggingFaceAPIChatGenerator * Alternate approach using pydantic like we do in from_function.py * Cleanup and fix other affected tests * Fix mypy * PR comments * PR comment * Remove test from HF API * Update reno * Update reno
7 lines
445 B
YAML
7 lines
445 B
YAML
---
|
||
enhancements:
|
||
- |
|
||
Refactored JSON Schema generation for ComponentTool parameters using Pydantic’s model_json_schema, enabling expanded type support (e.g. Union, Enum, Dict, etc.).
|
||
We also convert dataclasses to Pydantic models before calling model_json_schema to preserve docstring descriptions of the parameters in the schema.
|
||
This means dataclasses like ChatMessage, Document, etc. now have correctly defined JSON schemas.
|