Update app.py

app.py

@@ -1,65 +1,89 @@
+import os
+from typing import List
+from dataclasses import dataclass
+
 import chainlit as cl
-import arxiv
 import requests
-from typing import List
+import feedparser
+from dotenv import load_dotenv
 
+# LangChain bits (unchanged)
 from langchain.chat_models import ChatOpenAI
 from langchain.chains import ConversationalRetrievalChain
 from langchain.memory import ConversationBufferMemory
 from langchain.text_splitter import CharacterTextSplitter
 from langchain.embeddings import OpenAIEmbeddings
 from langchain.vectorstores import FAISS
-import os
-from dotenv import load_dotenv
+
 load_dotenv()
 
+ARXIV_API = "https://export.arxiv.org/api/query"
+
+# ---------- Simple paper container (drop-in replacement for arxiv.Result we used) ----------
+@dataclass
+class Paper:
+    title: str
+    summary: str
+    comment: str
+    entry_id: str
+    authors: List[str]
+
+# ---------- Direct arXiv API fetch (HTTPS + custom UA) ----------
+def fetch_arxiv_papers(query: str, max_results: int = 5) -> List[Paper]:
+    params = {
+        "search_query": query,
+        "id_list": "",
+        "sortBy": "relevance",
+        "sortOrder": "descending",
+        "start": 0,
+        "max_results": max_results,
+    }
+    headers = {
+        "User-Agent": f"arxiv-chainlit-app/1.0 (mailto:{os.getenv('CONTACT_EMAIL','noreply@example.com')})",
+        "Accept": "application/atom+xml",
+    }
+
+    resp = requests.get(ARXIV_API, params=params, headers=headers, timeout=20)
+    # Raise on non-200 so we can show a friendly error
+    resp.raise_for_status()
+
+    feed = feedparser.parse(resp.text)
+    papers: List[Paper] = []
+    for e in feed.entries:
+        title = getattr(e, "title", "").strip()
+        summary = getattr(e, "summary", "").strip()
+        comment = getattr(e, "arxiv_comment", "") if hasattr(e, "arxiv_comment") else ""
+        entry_id = getattr(e, "id", getattr(e, "link", ""))
+        authors = [a.get("name", "").strip() for a in getattr(e, "authors", [])]
+        papers.append(Paper(title=title, summary=summary, comment=comment, entry_id=entry_id, authors=authors or ["Unknown"]))
+    return papers
+
+# ---------- Your assistant, unchanged logic but using the new fetcher ----------
 class ArxivResearchAssistant:
     def __init__(self):
-        self.selected_paper = None
+        self.selected_paper: Paper | None = None
         self.qa_chain = None
-        self.papers: List[arxiv.Result] = []
+        self.papers: List[Paper] = []
         self.state = "SEARCH"
 
-        # ---- NEW: custom session with UA (no 'user_agent' kwarg) ----
-        sess = requests.Session()
-        sess.headers.update({
-            "User-Agent": f"arxiv-chainlit-app/1.0 (mailto:{os.getenv('CONTACT_EMAIL','noreply@example.com')})"
-        })
-        # If you’re behind a proxy or want requests to use env vars:
-        sess.trust_env = True
-
-        # ArXiv client (retries + small delay)
-        self.client = arxiv.Client(
-            page_size=5,
-            delay_seconds=3,
-            num_retries=3,
-            http_session=sess
-        )
-
     async def search_papers(self, query: str):
-        # Use arxiv.Search, then fetch with our client to leverage the session/retries
-        search = arxiv.Search(
-            query=query,
-            max_results=5,
-            sort_by=arxiv.SortCriterion.Relevance
-        )
         try:
-            self.papers = list(self.client.results(search))
+            self.papers = fetch_arxiv_papers(query, max_results=5)
+        except requests.HTTPError as e:
+            # Shows the real HTTP status & message (e.g., if UA missing or rate-limited)
+            await cl.Message(content=f"Error talking to arXiv (HTTP {e.response.status_code}): {e.response.text[:200]}").send()
+            return None
         except Exception as e:
-            await cl.Message(
-                content=f"Error talking to arXiv: {e}\nTry again in a moment or tweak your query."
-            ).send()
+            await cl.Message(content=f"Error talking to arXiv: {e}").send()
             return None
 
         if not self.papers:
-            await cl.Message(
-                content="No papers found. Please try another search query."
-            ).send()
+            await cl.Message(content="No papers found. Please try another search query.").send()
             return None
 
         paper_list = "\n".join([
-            f"{i+1}. {paper.title} - {paper.authors[0]}\nLink: {paper.entry_id}"
-            for i, paper in enumerate(self.papers)
+            f"{i+1}. {p.title} - {p.authors[0]}\nLink: {p.entry_id}"
+            for i, p in enumerate(self.papers)
         ])
         await cl.Message(
             content=f"Please select a paper by entering its number:\n\n{paper_list}\n\nEnter the number of the paper you want to select:"
@@ -69,9 +93,9 @@ class ArxivResearchAssistant:
 
     async def select_paper(self, selection: str):
         try:
-            selected_index = int(selection) - 1
-            if 0 <= selected_index < len(self.papers):
-                self.selected_paper = self.papers[selected_index]
+            idx = int(selection) - 1
+            if 0 <= idx < len(self.papers):
+                self.selected_paper = self.papers[idx]
             else:
                 await cl.Message(content="Invalid selection. Please try again.").send()
                 return None
@@ -79,12 +103,14 @@ class ArxivResearchAssistant:
             await cl.Message(content="Invalid input. Please enter a number.").send()
             return None
 
+        # Compose the text from the feed fields
         paper_text = (
             f"{self.selected_paper.title}\n\n"
             f"{self.selected_paper.summary}\n\n"
             f"{self.selected_paper.comment or ''}"
         )
 
+        # Split, embed, index (unchanged)
         text_splitter = CharacterTextSplitter(chunk_size=1000, chunk_overlap=100)
         chunks = text_splitter.split_text(paper_text)
 
@@ -148,6 +174,7 @@ class ArxivResearchAssistant:
         self.papers = []
         self.state = "SEARCH"
 
+# Global assistant instance
 assistant = ArxivResearchAssistant()
 
 @cl.on_chat_start