paste

.gitignore

@@ -1,4 +1,4 @@
 __pycache__/
-.streamlit/
-.streamlit/
-DATA_bziiit/vectorstore_op
+.streamlit
+DATA_bziiit/vectorstore_op
+.env

chat_te.py

@@ -20,12 +20,12 @@ def get_docs_from_pdf(file):
     return docs
 
 def get_doc_chunks(docs):
-    text_splitter = SemanticChunker(OpenAIEmbeddings())
+    text_splitter = SemanticChunker(OpenAIEmbeddings(model="text-embedding-3-large"))
     chunks = text_splitter.split_documents(docs)
     return chunks 
 
 def get_vectorstore_from_docs(doc_chunks):
-    embedding = OpenAIEmbeddings()
+    embedding = OpenAIEmbeddings(model="text-embedding-3-large")
     vectorstore = FAISS.from_documents(documents=doc_chunks, embedding=embedding)
     return vectorstore
 
@@ -47,7 +47,7 @@ def create_db(file):
     # docs = get_docs_from_pdf(file)
     # doc_chunks = get_doc_chunks(docs)
     # vectorstore = get_vectorstore_from_docs(doc_chunks)
-    vectorstore = FAISS.load_local("index", OpenAIEmbeddings())
+    vectorstore = FAISS.load_local(file, OpenAIEmbeddings(model="text-embedding-3-large"),allow_dangerous_deserialization= True)
     return vectorstore
 
 def get_response(chain,user_query, chat_history):
@@ -81,7 +81,7 @@ def display_chat_te():
             AIMessage(content="Salut, posez-moi vos question sur la transistion ecologique."),
         ]
     if "chain" not in st.session_state:
-        db=create_db("DATA_bziiit/op.pdf")
+        db=create_db("./DATA_bziiit/vectorstore_op")
         chain = get_conversation_chain(db)
         st.session_state.chain = chain
         

requirements.txt

@@ -37,3 +37,4 @@ kaleido
 langchain-core 
 langchain-mistralai
 firecrawl-py
+st_copy_to_clipboard