Add rate limiting and API key management

- Add Semantic Scholar API key support in environment variables
- Implement 1 req/sec rate limiting for both LLM and Semantic Scholar APIs
- Sequential processing to avoid concurrency issues
- Load environment variables using python-dotenv
- Update README with rate limiting information and security notes
- Ensure .env is in .gitignore (already present)

.env.example

@@ -3,6 +3,9 @@ OPENAI_API_KEY=your-api-key-here
 OPENAI_BASE_URL=https://api.openai.com/v1
 MODEL_NAME=gpt-4
 
+# Semantic Scholar API Configuration
+SEMANTIC_SCHOLAR_API_KEY=your-semantic-scholar-api-key-here
+
 # Alternative configurations:
 # For Azure OpenAI:
 # OPENAI_BASE_URL=https://your-resource.openai.azure.com/

agents.py

@@ -3,6 +3,7 @@
 import json
 import re
 import os
+import time
 from typing import Any, Optional, Dict, Tuple
 from openai import OpenAI
 
@@ -35,8 +36,11 @@ def extract_json_between_markers(llm_output: str) -> Optional[Dict[str, Any]]:
 
 
 def query_model(system_prompt: str, prompt: str, client: OpenAI, model: str) -> Optional[str]:
-    """Query the model with the given prompts using OpenAI-compatible API."""
+    """Query the model with the given prompts using OpenAI-compatible API with rate limiting."""
     try:
+        # Rate limiting: 1 request per second to avoid concurrency issues
+        time.sleep(1)
+
         response = client.chat.completions.create(
             model=model,
             messages=[
@@ -49,6 +53,8 @@ def query_model(system_prompt: str, prompt: str, client: OpenAI, model: str) ->
         return response.choices[0].message.content
     except Exception as e:
         print(f"Error querying model: {e}")
+        # Wait before retry
+        time.sleep(2)
         return None
 
 

app.py

@@ -7,6 +7,10 @@ from markitdown import MarkItDown
 from agents import MultiReviewerSystem
 import requests
 import time
+from dotenv import load_dotenv
+
+# Load environment variables from .env file
+load_dotenv()
 
 
 def extract_text_from_pdf(pdf_file) -> str:
@@ -23,8 +27,8 @@ def extract_text_from_pdf(pdf_file) -> str:
         return f"Error extracting text from PDF: {str(e)}"
 
 
-def search_semantic_scholar(query: str, limit: int = 5) -> List[Dict]:
-    """Search for related papers on Semantic Scholar."""
+def search_semantic_scholar(query: str, limit: int = 5, api_key: str = None) -> List[Dict]:
+    """Search for related papers on Semantic Scholar with rate limiting."""
     try:
         url = "https://api.semanticscholar.org/graph/v1/paper/search"
         params = {
@@ -33,7 +37,14 @@ def search_semantic_scholar(query: str, limit: int = 5) -> List[Dict]:
             "fields": "title,authors,year,abstract,citationCount,url,openAccessPdf"
         }
 
-        response = requests.get(url, params=params)
+        headers = {}
+        if api_key:
+            headers["x-api-key"] = api_key
+
+        # Rate limiting: 1 request per second
+        time.sleep(1)
+
+        response = requests.get(url, params=params, headers=headers, timeout=10)
         response.raise_for_status()
 
         data = response.json()
@@ -122,9 +133,9 @@ def review_paper(
     if search_related:
         progress(0.2, desc="Searching for related papers...")
         paper_title = extract_paper_title_from_text(paper_text)
-        related_papers = search_semantic_scholar(paper_title, limit=5)
+        semantic_scholar_key = os.getenv("SEMANTIC_SCHOLAR_API_KEY", "")
+        related_papers = search_semantic_scholar(paper_title, limit=5, api_key=semantic_scholar_key)
         related_papers_md = format_semantic_scholar_results(related_papers)
-        time.sleep(1)  # Rate limiting
 
     # Initialize multi-reviewer system
     progress(0.3, desc="Initializing reviewers...")