(fix) unable to pass input_type parameter to Voyage AI embedding mode (#7276)
All checks were successful
Read Version from pyproject.toml / read-version (push) Successful in 46s

* VoyageEmbeddingConfig

* fix voyage logic to get params

* add voyage embedding transformation

* add get_provider_embedding_config

* use BaseEmbeddingConfig

* voyage clean up

* use llm http handler for embedding transformations

* test_voyage_ai_embedding_extra_params

* add voyage async

* test_voyage_ai_embedding_extra_params

* add async for llm http handler

* update BaseLLMEmbeddingTest

* test_voyage_ai_embedding_extra_params

* fix linting

* fix get_provider_embedding_config

* fix anthropic text test

* update location of base/chat/transformation

* fix import path

* fix IBMWatsonXAIConfig
This commit is contained in:
Ishaan Jaff 2024-12-17 19:23:49 -08:00 committed by GitHub
parent c90942e550
commit 7a5dd29fe0
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
52 changed files with 535 additions and 66 deletions

View file

@ -3058,6 +3058,7 @@ async def aembedding(*args, **kwargs) -> EmbeddingResponse:
or custom_llm_provider == "together_ai"
or custom_llm_provider == "openai_like"
or custom_llm_provider == "jina_ai"
or custom_llm_provider == "voyage"
): # currently implemented aiohttp calls for just azure and openai, soon all.
# Await normally
init_response = await loop.run_in_executor(None, func_with_context)
@ -3632,10 +3633,10 @@ def embedding( # noqa: PLR0915
aembedding=aembedding,
)
elif custom_llm_provider == "voyage":
api_key = api_key or litellm.api_key or get_secret_str("VOYAGE_API_KEY")
response = openai_chat_completions.embedding(
response = base_llm_http_handler.embedding(
model=model,
input=input,
custom_llm_provider=custom_llm_provider,
api_base=api_base,
api_key=api_key,
logging_obj=logging,