Skip to content

armgit / dspy

Public Imported from GitHub
0

DSPy: The framework for programming—not prompting—language models

Python

Update ChromadbRM to openai ^1.0

DanielUH20197cd8da42144f
diff --git a/dspy/retrieve/chromadb_rm.py b/dspy/retrieve/chromadb_rm.py
index 82ab0023f..03ddd8ad9 100644
--- a/dspy/retrieve/chromadb_rm.py
+++ b/dspy/retrieve/chromadb_rm.py
@@ -11,6 +11,7 @@ from dsp.utils import dotdict
try:
import chromadb
from chromadb.config import Settings
+ from chromadb.utils import embedding_functions
except ImportError:
chromadb = None
@@ -70,17 +71,13 @@ class ChromadbRM(dspy.Retrieve):
self._init_chromadb(collection_name, persist_directory)
- # If not provided, defaults to env vars
- if openai_api_key:
- openai.api_key = openai_api_key
- if openai_api_type:
- openai.api_type = openai_api_type
- if openai_api_base:
- openai.api_base = openai_api_base
- if openai_api_version:
- openai.api_version = openai_api_version
- if openai_api_provider:
- self._openai_api_provider = openai_api_provider
+ self.openai_ef = embedding_functions.OpenAIEmbeddingFunction(
+ api_key=openai_api_key,
+ api_base=openai_api_base,
+ api_type=openai_api_type,
+ api_version=openai_api_version,
+ model_name=openai_embed_model,
+ )
super().__init__(k=k)
@@ -111,7 +108,7 @@ class ChromadbRM(dspy.Retrieve):
@backoff.on_exception(
backoff.expo,
- (openai.error.RateLimitError, openai.error.ServiceUnavailableError),
+ (openai.RateLimitError),
max_time=15,
)
def _get_embeddings(self, queries: List[str]) -> List[List[float]]:
@@ -124,24 +121,10 @@ class ChromadbRM(dspy.Retrieve):
List[List[float]]: List of embeddings corresponding to each query.
"""
- if self._openai_api_provider == "azure":
- model_args = {
- "engine": self._openai_embed_model,
- "deployment_id": self._openai_embed_model,
- "api_version": openai.api_version,
- "api_base": openai.api_base,
- }
- embedding = openai.Embedding.create(
- input=queries,
- model=self._openai_embed_model,
- **model_args,
- api_provider=self._openai_api_provider
- )
- else:
- embedding = openai.Embedding.create(
- input=queries, model=self._openai_embed_model
- )
- return [embedding["embedding"] for embedding in embedding["data"]]
+ embedding = self.openai_ef._client.create(
+ input=queries, model=self._openai_embed_model
+ )
+ return [embedding.embedding for embedding in embedding.data]
def forward(
self, query_or_queries: Union[str, List[str]], k: Optional[int] = None