Regenerate Python SDK for v0.4.0

[atacan]

Summary

• regenerate the Python SDK from the updated Speechall OpenAPI document
• update the OpenAPI spec paths used by Fern and regenerate.sh for the current workspace layout
• bump the package version from 0.3.0 to 0.4.0
• update generated transcription model identifiers

Regeneration Notes

./regenerate.sh now validates the OpenAPI spec at ../../Speechall-Repositories/speechall-openapi/openapi.yaml, while Fern reads ../../../Speechall-Repositories/speechall-openapi/openapi.yaml from fern/generators.yml.

Local generation required a manual equivalent Fern run because this machine's Colima setup could not mount Fern's default temp-file mounts correctly. Fern spec validation passed, then the generated Python output from fernapi/fern-python-sdk:4.30.3 was synced into src/.

Validation

• python3 -m venv .venv
• .venv/bin/python -m pip install -e ".[dev]"
• import smoke check for SpeechallApi, AsyncSpeechallApi, package metadata 0.4.0, and new model identifier
• no tests/ directory present, so pytest was skipped
• .venv/bin/python -m build
• git diff --check

Build completed successfully. Setuptools emitted a deprecation warning for the existing project.license table format; no build failure.

Beads

Tracked issue: speechall-python-sdk-5nq.