Pin pydantic==2.10.6: fixes gradio_client TypeError on bool-typed additionalProperties
Browse files- requirements.txt +6 -0
requirements.txt
CHANGED
|
@@ -12,6 +12,12 @@ gradio==4.44.1
|
|
| 12 |
# image (which installs huggingface_hub>=0.30 with no upper bound).
|
| 13 |
huggingface_hub>=0.30,<1.0
|
| 14 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 15 |
# Hybrid route classifier (TF-IDF + logistic regression).
|
| 16 |
# Loose lower bounds so the resolver picks wheels compatible with the
|
| 17 |
# Python version configured in the README YAML (currently 3.10).
|
|
|
|
| 12 |
# image (which installs huggingface_hub>=0.30 with no upper bound).
|
| 13 |
huggingface_hub>=0.30,<1.0
|
| 14 |
|
| 15 |
+
# Pydantic 2.11+ emits JSON schemas with `additionalProperties: true` (boolean),
|
| 16 |
+
# which the gradio_client 1.3.0 bundled with gradio 4.44.1 cannot parse — it
|
| 17 |
+
# crashes with "TypeError: argument of type 'bool' is not iterable" the first
|
| 18 |
+
# time the API info endpoint is hit. Pinning to 2.10.6 keeps schemas as dicts.
|
| 19 |
+
pydantic==2.10.6
|
| 20 |
+
|
| 21 |
# Hybrid route classifier (TF-IDF + logistic regression).
|
| 22 |
# Loose lower bounds so the resolver picks wheels compatible with the
|
| 23 |
# Python version configured in the README YAML (currently 3.10).
|