create an index if none exists

App/Embedding/utils/Initialize.py

@@ -7,23 +7,37 @@ import os
 
 
 
-
-async def delete_documents(task_id):
-    # get api key from app.pinecone.io
-    PINECONE_API_KEY = os.environ.get("PINECONE_API_KEY")
+def initDocument():
+        # get api key from app.pinecone.io
+    PINECONE_API_KEY = os.environ.get("PINECONE_API_KEY",'')
     # find your environment next to the api key in pinecone console
     PINECONE_ENV = os.environ.get("PINECONE_ENVIRONMENT")
 
 
     index_name = "transcript-bits"
-    model_name = "thenlper/gte-base"
+    model_name = "Alibaba-NLP/gte-base-en-v1.5"
     embeddings = HuggingFaceEmbeddings(model_name=model_name)
     spec = PodSpec()
 
+    try:
+        pc=pinecone.Pinecone(api_key=PINECONE_API_KEY, environment=PINECONE_ENV,spec=spec)
+        vector_index = pc.Index(index_name)
+    except:
+        pc.create_index(name=index_name,
+        dimension=768,
+        metric="cosine",
+        spec=spec)
+    
+
+        docsearch = Pinecone.from_existing_index(index_name, embeddings)
+
+    return docsearch
 
-    pc=pinecone.Pinecone(api_key=PINECONE_API_KEY, environment=PINECONE_ENV,spec=spec)
-    vector_index = pc.Index(index_name)
-    docsearch = Pinecone.from_existing_index(index_name, embeddings)
+
+
+
+async def delete_documents(task_id):
+    docsearch=initDocument()
 
     docsearch.delete(
         filter={
@@ -54,20 +68,7 @@ def generateChunks(chunks, task_id, n=100):
 
 
 def search(query: str, task_id: str):
-# get api key from app.pinecone.io
-    PINECONE_API_KEY = os.environ.get("PINECONE_API_KEY")
-    # find your environment next to the api key in pinecone console
-    PINECONE_ENV = os.environ.get("PINECONE_ENVIRONMENT")
-
-
-    index_name = "transcript-bits"
-    model_name = "thenlper/gte-base"
-    embeddings = HuggingFaceEmbeddings(model_name=model_name)
-
-
-    pc=pinecone.Pinecone(api_key=PINECONE_API_KEY, environment=PINECONE_ENV)
-    vector_index = pc.Index(index_name)
-    docsearch = Pinecone.from_existing_index(index_name, embeddings)
+    docsearch=initDocument()
 
     filtering_conditions = {
             "task_id": {"$eq": task_id},
@@ -81,18 +82,10 @@ def search(query: str, task_id: str):
 
 
 def encode(temp: list[Document]):
-    PINECONE_API_KEY = os.environ.get("PINECONE_API_KEY")
-    # find your environment next to the api key in pinecone console
-    PINECONE_ENV = os.environ.get("PINECONE_ENVIRONMENT")
-
+    docsearch=initDocument()
+    docsearch.add_documents(temp)
+    # return  embeddings.embed_documents(texts = [d.page_content for d in temp])
 
-    index_name = "transcript-bits"
-    model_name = "thenlper/gte-base"
-    embeddings = HuggingFaceEmbeddings(model_name=model_name)
 
 
-    pc=pinecone.Pinecone(api_key=PINECONE_API_KEY, environment=PINECONE_ENV)
-    vector_index = pc.Index(index_name)
-    docsearch = Pinecone.from_existing_index(index_name, embeddings)
-    docsearch.add_documents(temp)
-    # return  embeddings.embed_documents(texts = [d.page_content for d in temp])
+initDocument()