Update tools/SemanticScholar.py

tools/SemanticScholar.py

@@ -1,41 +1,74 @@
+import os
 from smolagents import Tool
-from semanticscholar import SemanticScholar
+from googleapiclient.discovery import build
 
-class AcademicPaperSearchTool(Tool):
+class GoogleAcademicSearchTool(Tool):
+    """
+    A specialized search tool that uses the Google Custom Search Engine (CSE) 
+    API to find academic papers by applying strong search filters (e.g., site:scholar.google.com).
+    This replaces the Semantic Scholar tool which was experiencing timeouts.
+    """
     name = "academic_paper_search"
-    description = "Use Websearch first for getting factual data. If needed use this tool to Search academic papers via Semantic Scholar and returns the most relevant titles and abstracts."
+    description = "Searches for academic papers by restricting the Google search to academic domains (like Google Scholar) and returns relevant titles, snippets, and source links."
+    
     inputs = {
-        "query": {"type": "string", "description": "Search query for academic papers (title, keywords, etc.)"}
+        "query": {"type": "string", "description": "Search query for academic papers (title, keywords, author, etc.)"}
     }
     output_type = "string"
 
+    def __init__(self, **kwargs):
+        super().__init__(**kwargs)
+
+        # Retrieve credentials from environment variables
+        self.api_key = os.getenv("GOOGLE_API_KEY")
+        self.cse_id = os.getenv("GOOGLE_CSE_ID")
+
+        if not self.api_key or not self.cse_id:
+            raise ValueError("GOOGLE_API_KEY or GOOGLE_CSE_ID secret not found.")
+
+        # Initialize the Google Custom Search service
+        self.service = build(
+            "customsearch", "v1", developerKey=self.api_key
+        )
+
     def forward(self, query: str) -> str:
-        sch = SemanticScholar()
-        try:
-            # Use query=query to avoid potential keyword issues with the SemanticScholar method
-            papers = sch.search_paper(query, limit=3)  # get top-3 matching papers
-        except Exception as e:
-            return f"An error occurred during search: {e}"
+        """
+        Executes a Google search query, restricting results to academic domains.
+        """
+        print(f"Executing Google Academic search for: '{query}'")
+        
+        # Modify the query to prioritize academic sources
+        # We use OR to check multiple popular domains for better coverage
+        academic_filter = "site:scholar.google.com OR site:researchgate.net OR site:pubmed.ncbi.nlm.nih.gov"
+        google_academic_query = f"{query} {academic_filter}"
 
-        if not papers:
-            return "No papers found."
+        try:
+            # Execute the search request for up to 3 results
+            res = self.service.cse().list(
+                q=google_academic_query,
+                cx=self.cse_id,
+                num=3
+            ).execute()
 
-        # Format results concisely
-        lines = []
-        for p in papers:
-            # --- THE FIX IS HERE ---
-            # Access attributes directly using dot notation (p.title, p.year, p.abstract)
-            # instead of the dictionary method p.get(...)
-            title = p.title
-            year = p.year
-            abstract = p.abstract if p.abstract else "" # Check if abstract exists
+            items = res.get('items', [])
 
-            # Only take the first sentence or two of the abstract
-            abstract_snip = abstract.split(". ")[0] if abstract else ""
+            if not items:
+                return "XX record info: No academic results found using Google search filters."
 
-            # Ensure 'year' is handled gracefully (it can sometimes be None)
-            year_str = str(year) if year else "N/A"
+            search_results = []
+            for i, item in enumerate(items):
+                title = item.get('title', 'N/A')
+                snippet = item.get('snippet', 'No snippet available.')
+                link = item.get('link', 'N/A')
+                
+                # Format similar to the previous academic tool
+                search_results.append(
+                    f"RESULT {i+1}: **{title}**\n"
+                    f"CONTENT: {snippet}\n"
+                    f"SOURCE: {link}"
+                )
 
-            lines.append(f"**{title}** ({year_str}): {abstract_snip}...")
+            return "\n\n---SEPARATOR---\n\n".join(search_results)
 
-        return "\n".join(lines)
+        except Exception as e:
+            return f"Error during Google Academic Search API call: {e}"