mirror of
https://github.com/MODSetter/SurfSense.git
synced 2026-05-27 19:25:15 +02:00
32f6766cb6
LiteLLM normalizes every provider's cache fields onto usage.prompt_tokens_details (cached_tokens + cache_creation_tokens). The earlier fallback to usage.cache_read_input_tokens / usage.cache_creation_input_tokens was wrong: Anthropic-shaped fields only live there via a trailing setattr loop, and the canonical field name on the wrapper is cache_creation_tokens (not _input_tokens). |
||
|---|---|---|
| .. | ||
| agents | ||
| config | ||
| connectors | ||
| etl_pipeline | ||
| indexing_pipeline | ||
| observability | ||
| prompts | ||
| retriever | ||
| routes | ||
| schemas | ||
| services | ||
| tasks | ||
| templates | ||
| utils | ||
| __init__.py | ||
| app.py | ||
| celery_app.py | ||
| db.py | ||
| exceptions.py | ||
| rate_limiter.py | ||
| users.py | ||