bstraehle commited on
Commit
309e834
·
1 Parent(s): f6fcf7f

Update rag.py

Browse files
Files changed (1) hide show
  1. rag.py +12 -12
rag.py CHANGED
@@ -63,33 +63,33 @@ def document_splitting(config, docs):
63
 
64
  def document_storage_chroma(chunks):
65
  Chroma.from_documents(documents = chunks,
66
- embedding = OpenAIEmbeddings(disallowed_special = ()),
67
  persist_directory = CHROMA_DIR)
68
 
69
  def document_storage_mongodb(chunks):
70
  MongoDBAtlasVectorSearch.from_documents(documents = chunks,
71
- embedding = OpenAIEmbeddings(disallowed_special = ()),
72
  collection = collection,
73
  index_name = MONGODB_INDEX_NAME)
74
 
75
- def rag_batch(config):
76
- docs = document_loading()
77
-
78
- chunks = document_splitting(config, docs)
79
-
80
- document_storage_chroma(chunks)
81
- document_storage_mongodb(chunks)
82
-
83
  def document_retrieval_chroma():
84
- return Chroma(embedding_function = OpenAIEmbeddings(disallowed_special = ()),
85
  persist_directory = CHROMA_DIR)
86
 
87
  def document_retrieval_mongodb():
88
  return MongoDBAtlasVectorSearch.from_connection_string(MONGODB_ATLAS_CLUSTER_URI,
89
  MONGODB_DB_NAME + "." + MONGODB_COLLECTION_NAME,
90
- OpenAIEmbeddings(disallowed_special = ()),
91
  index_name = MONGODB_INDEX_NAME)
92
 
 
 
 
 
 
 
 
 
93
  def get_llm(config, openai_api_key):
94
  return ChatOpenAI(model_name = config["model_name"],
95
  openai_api_key = openai_api_key,
 
63
 
64
  def document_storage_chroma(chunks):
65
  Chroma.from_documents(documents = chunks,
66
+ embedding = OpenAIEmbeddings(),
67
  persist_directory = CHROMA_DIR)
68
 
69
  def document_storage_mongodb(chunks):
70
  MongoDBAtlasVectorSearch.from_documents(documents = chunks,
71
+ embedding = OpenAIEmbeddings(),
72
  collection = collection,
73
  index_name = MONGODB_INDEX_NAME)
74
 
 
 
 
 
 
 
 
 
75
  def document_retrieval_chroma():
76
+ return Chroma(embedding_function = OpenAIEmbeddings(),
77
  persist_directory = CHROMA_DIR)
78
 
79
  def document_retrieval_mongodb():
80
  return MongoDBAtlasVectorSearch.from_connection_string(MONGODB_ATLAS_CLUSTER_URI,
81
  MONGODB_DB_NAME + "." + MONGODB_COLLECTION_NAME,
82
+ OpenAIEmbeddings(),
83
  index_name = MONGODB_INDEX_NAME)
84
 
85
+ def rag_batch(config):
86
+ docs = document_loading()
87
+
88
+ chunks = document_splitting(config, docs)
89
+
90
+ document_storage_chroma(chunks)
91
+ document_storage_mongodb(chunks)
92
+
93
  def get_llm(config, openai_api_key):
94
  return ChatOpenAI(model_name = config["model_name"],
95
  openai_api_key = openai_api_key,