Commit Graph

6 Commits

Author SHA1 Message Date
Alexei Vedernikov d815c8b8dc
fix(llm): support per-request provider timeouts (#832)
* fix(llm): support per-request provider timeouts

* fix(llm): convert Gemini timeout to milliseconds

* fix(llm): validate Gemini HTTP options

* test(llm): type Anthropic stream context args

* test(llm): live per-request timeout coverage for all providers

Two live checks per provider: a generous timeout asserted at the SDK
call boundary, and a tight timeout that must abort well under the 600s
client default. Gemini's async transport can be aiohttp, so its tight
timeout surfaces as asyncio.TimeoutError rather than httpx.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* style(tests): drop extra blank line in anthropic backend test

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* fix(llm): validate provider_params.timeout at config load

Move the timeout coercion into src.config as coerce_provider_timeout and
run it from a field validator on ModelOverrideSettings.provider_params, so
a bad value in config.toml/env fails at startup with the exact config path
instead of surfacing per-request as a retried 500. Good values normalize
to float seconds at load. The per-request guard in src.llm.backend now
delegates to the same coercion (wrapping ValueError in ValidationException)
and continues to cover extra_params passed programmatically.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* docs: document provider_params.timeout load-time validation and gotchas

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* refactor(llm): address review nits on timeout plumbing

Apply eisene's review feedback:
- Rename PROVIDER_TIMEOUT_ERROR → PROVIDER_TIMEOUT_ERROR_TEXT
- Move request_timeout_from_extra_params from backend.py (pure
  dataclasses) to request_builder.py (request assembly)
- Add comment explaining Gemini's ms timeout conversion
- Generalize _normalize_extra_params with _strip_none_params helper

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

---------

Co-authored-by: Aakash Kattelu <aakash@plasticlabs.ai>
Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
2026-08-04 12:43:00 -04:00
Eugene Eisenstein c7c1597d2c
Fix `unified-tests.yml` secrets (#895)
* Structured outputs for dialectic

* Fix unified-tests.yml secrets overrides

* Revert "Structured outputs for dialectic"

This reverts commit a221e2c282.

* remove throwing validator, document assumption that config overrides be backward compatible

* fix typo

* fix: ruff format config file

* fix: update pyproject.toml to include exclude-newer

---------

Co-authored-by: Rajat Ahuja <rahuja445@gmail.com>
2026-07-13 10:36:34 -04:00
Eugene Eisenstein 43d962d160 rename to REPRESENTATION_BATCH_TARGET_INPUT_TOKENS 2026-07-09 10:42:52 -04:00
Eugene Eisenstein be26c859ad split config entry into 2 2026-07-08 12:24:55 -04:00
Harish Kukreja 414e31c960
feat(deriver): age-flush stalled representation batches (#826)
Co-authored-by: Vineeth Voruganti <13438633+VVoruganti@users.noreply.github.com>
2026-06-22 17:49:08 -04:00
adavyas a420264152
feat: deriver custom instructions (#609)
* feat: wire deriver custom instructions on main

* refactor: simplify custom instruction normalization

* chore: lower deriver custom instruction cap

* chore: raise deriver custom instruction budgets

* docs: update deriver input token example

* fix: hide deriver config guidance from validation

* chore: address custom instruction review nits

* docs: document deriver custom instruction cap

* fix: remove unused tests/validation and simplify enable flag for custom instructions

---------

Co-authored-by: Vineeth Voruganti <13438633+VVoruganti@users.noreply.github.com>
2026-05-11 18:05:42 -04:00