Update app.py

replaced arxivretreiver with arxivloader

app.py

@@ -11,26 +11,29 @@ from langchain_mistralai import ChatMistralAI
 from langchain_community.document_loaders import PyPDFLoader
 import requests
 from pathlib import Path
-from langchain_community.document_loaders import WebBaseLoader
-from langchain_community.retrievers import ArxivRetriever
+from langchain_community.document_loaders import WebBaseLoader, ArxivLoader
 import bs4
 from langchain_core.rate_limiters import InMemoryRateLimiter
 from urllib.parse import urljoin
 
 
 def initialize(arxivcode):  
+    loader = ArxivLoader(query=arxivcode,)
+    docs = loader.load()
+    #retriever = ArxivRetriever(
+    #    load_max_docs=2,
+    #    get_full_documents=True,
+    #)
+    #docs = retriever.invoke(str(arxivcode))
+    #for i in range(len(docs)): 
+    #    docs[i].metadata['Published'] = str(docs[i].metadata['Published'])
+
+    # LLM model
     rate_limiter = InMemoryRateLimiter(
         requests_per_second=0.1,  # <-- MistralAI free. We can only make a request once every second
         check_every_n_seconds=0.01,  # Wake up every 100 ms to check whether allowed to make a request,
         max_bucket_size=10,  # Controls the maximum burst size.
-    )
-    
-    retriever = ArxivRetriever(
-        load_max_docs=2,
-        get_ful_documents=True,
-    )
-    
-    # LLM model
+    )    
     llm = ChatMistralAI(model="mistral-large-latest", rate_limiter=rate_limiter)
     
     # Embeddings
@@ -38,10 +41,6 @@ def initialize(arxivcode):
     # embed_model = "nvidia/NV-Embed-v2"
     embeddings = HuggingFaceInstructEmbeddings(model_name=embed_model)
     # embeddings = MistralAIEmbeddings()
-    
-    docs = retriever.invoke(str(arxivcode))
-    for i in range(len(docs)): 
-        docs[i].metadata['Published'] = str(docs[i].metadata['Published'])
 
     def format_docs(docs):
         return "\n\n".join(doc.page_content for doc in docs)