Update app.py

app.py

@@ -24,6 +24,7 @@ from io import StringIO
 last = 0
 CHROMA_DATA_PATH = "chroma_data/"
 EMBED_MODEL = "sentence-transformers/all-MiniLM-L6-v2" #"BAAI/bge-m3"
+LLM_NAME = "mistralai/Mistral-Nemo-Instruct-24072"
 # all-MiniLM-L6-v2
 CHUNK_SIZE = 800
 CHUNK_OVERLAP = 50
@@ -65,7 +66,7 @@ login(token=("hf_" + HF_TOKEN))
 system_propmpt = system_sr
 
 # "facebook/blenderbot-400M-distill",  facebook/blenderbot-400M-distill, stabilityai/stablelm-zephyr-3b, BAAI/bge-small-en-v1.5
-Settings.llm = HuggingFaceInferenceAPI(model_name="mistralai/Mistral-Nemo-Instruct-2407",
+Settings.llm = HuggingFaceInferenceAPI(model_name=LLM_NAME,
                               device_map="auto",
                               system_prompt = system_propmpt,
                               context_window=5100,
@@ -117,9 +118,9 @@ def rag(input_text, history, jezik, file):
 #        resetChat()
     print(history, input_text)
 ##    if (file):
-#        documents = []
-#        for f in file:
-#            documents += SimpleDirectoryReader(f).load_data()
+        documents = []
+        for f in file:
+            documents += SimpleDirectoryReader(f).load_data()
 #        f = file + "*.pdf"
 ##        pathname = os.path.dirname
 #        shutil.copyfile(file.name, path)
@@ -127,7 +128,7 @@ def rag(input_text, history, jezik, file):
 ##        print("basename=", os.path.basename(file))
 ##        print("filename=", file.name)        
 ##        documents = SimpleDirectoryReader(file).load_data()
-##        index2 = VectorStoreIndex.from_documents(documents)
+        index2 = VectorStoreIndex.from_documents(documents)
 ##        query_engine = index2.as_query_engine()
 #        return query_engine.query(input_text)
 #        return history.append({"role": "assistant", "content": query_engine.query(input_text)})