fixe knowledge base

app.py

@@ -53,7 +53,7 @@ if prompt :
         message_placeholder = st.empty()
         full_response = ""
         # Define the basic input structure for the chains
-        input_dict = {"input": prompt}
+        input_dict = {"input": prompt.lower()}
 
 
         with collect_runs() as cb:

{sup-knowledge-eng-nomic/ec6754ec-5fa6-4b04-bfb6-d2f052cd81fe → knowledge-base/5836a1af-4663-4a50-a6a7-27a140e06517}/length.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a5fb64b021f47ff585087f63e019088911fa892704ffa3e9506f3a120d807cfa
+oid sha256:f9aaf11150dfda02ab48ea570f99528198a3a2ef4bf99b8a4f717ef51f2ba790
 size 4000

{sup-knowledge-eng-nomic → knowledge-base}/chroma.sqlite3

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:112a7ee3f7fb675803ed49ffe7901311156373f8ba3142c3a3026b2f3936d633
+oid sha256:80a088dcdeeeeda4cc6de332a7e7d8315f7c052892d9c0795ab4bb52f43fa1b4
 size 7704576

utils.py

@@ -14,12 +14,12 @@ from qdrant_client import models
 load_dotenv()
 #Retriever
 def retriever(n_docs=5):
-    vector_database_path = "sup-knowledge-eng-nomic"
+    vector_database_path = "knowledge-base"
 
     embeddings_model = NomicEmbeddings(model="nomic-embed-text-v1.5", inference_mode="local")
 
 
-    vectorstore = Chroma(collection_name="sup-store-eng-nomic",
+    vectorstore = Chroma(collection_name="knowledge-base",
                         persist_directory=vector_database_path,
                     embedding_function=embeddings_model)