Upload 2 files

agent.py

@@ -23,6 +23,22 @@ from langchain_anthropic import ChatAnthropic
 from supabase.client import Client, create_client
 
 
+import re
+from langchain_community.document_loaders import WikipediaLoader
+from langchain_core.tools import tool
+
+
+from langchain_core.tools import tool
+from youtube_transcript_api import YouTubeTranscriptApi, TranscriptsDisabled, NoTranscriptFound
+
+from langchain_core.tools import tool
+from transformers import pipeline
+
+
+import sys
+
+# Before invoking your graph:
+sys.setrecursionlimit(100)  # Increase from default 25
 
 load_dotenv()
 
@@ -80,39 +96,32 @@ def modulus(a: int, b: int) -> int:
 
 @tool
 def wiki_search(query: str) -> str:
-    """Search Wikipedia for a query and return up to 5 results summarized."""
-    search_docs = WikipediaLoader(query=query, load_max_docs=5).load()
-    if not search_docs:
-        return "No Wikipedia results found."
-
-    summaries = []
-    for doc in search_docs:
-        content = doc.page_content
-        # Truncate if too long or summarize with a simple heuristic or model
-        content = content[:1000] + ("..." if len(content) > 1000 else "")
-        summaries.append(
-            f'<Document source="{doc.metadata["source"]}" page="{doc.metadata.get("page", "")}">\n{content}\n</Document>'
-        )
-    return {"wiki_results": "\n\n---\n\n".join(summaries)}
-
+    """Search Wikipedia for a query and return maximum 2 results.
+    
+    Args:
+        query: The search query."""
+    search_docs = WikipediaLoader(query=query, load_max_docs=2).load()
+    formatted_search_docs = "\n\n---\n\n".join(
+        [
+            f'<Document source="{doc.metadata["source"]}" page="{doc.metadata.get("page", "")}"/>\n{doc.page_content}\n</Document>'
+            for doc in search_docs
+        ])
+    return {"wiki_results": formatted_search_docs}
     
 
 
 @tool
 def web_search(query: str) -> str:
-    """Search Tavily for a query and return maximum 3 results."""
+    """Search Tavily for a query and return maximum 3 results.
+    
+    Args:
+        query: The search query."""
     search_docs = TavilySearch(max_results=3).invoke(query=query)
-    if not search_docs:
-        return "No web search results found."
-
     formatted_search_docs = "\n\n---\n\n".join(
         [
-            f'<Document source="{doc.metadata.get("source", "unknown")}" page="{doc.metadata.get("page", "")}"/>\n'
-            f'{doc.page_content[:1000]}'  # truncate to 1000 chars if needed
-            f'\n</Document>'
+            f'<Document source="{doc.metadata["source"]}" page="{doc.metadata.get("page", "")}"/>\n{doc.page_content}\n</Document>'
             for doc in search_docs
-        ]
-    )
+        ])
     return {"web_results": formatted_search_docs}
 
 
@@ -129,6 +138,36 @@ def arvix_search(query: str) -> str:
             for doc in search_docs
         ])
     return {"arvix_results": formatted_search_docs}
+    
+    
+@tool
+def filtered_wiki_search(query: str, start_year: int = None, end_year: int = None) -> dict:
+    """Search Wikipedia for a query and filter results by year if provided."""
+    search_docs = WikipediaLoader(query=query, load_max_docs=5).load()
+    
+    def contains_year(text, start, end):
+        years = re.findall(r'\b(19\d{2}|20\d{2})\b', text)
+        for y in years:
+            y_int = int(y)
+            if start <= y_int <= end:
+                return True
+        return False
+
+    filtered_docs = []
+    for doc in search_docs:
+        if start_year and end_year:
+            if contains_year(doc.page_content, start_year, end_year):
+                filtered_docs.append(doc)
+        else:
+            filtered_docs.append(doc)
+
+    formatted_search_docs = "\n\n---\n\n".join(
+        [
+            f'<Document source="{doc.metadata["source"]}" page="{doc.metadata.get("page", "")}"/>\n{doc.page_content}\n</Document>'
+            for doc in filtered_docs
+        ])
+    return {"wiki_results": formatted_search_docs}
+
 
 
 @tool
@@ -141,6 +180,87 @@ def wolfram_alpha_query(query: str) -> str:
         return next(res.results).text
     except StopIteration:
         return "No result found."
+        
+
+
+
+@tool
+def youtube_transcript(url: str) -> str:
+    """Fetch YouTube transcript text from a video URL."""
+    try:
+        video_id = url.split("v=")[-1].split("&")[0]
+        transcript_list = YouTubeTranscriptApi.get_transcript(video_id)
+        transcript = " ".join([segment['text'] for segment in transcript_list])
+        return transcript
+    except (TranscriptsDisabled, NoTranscriptFound):
+        return "Transcript not available for this video."
+    except Exception as e:
+        return f"Error fetching transcript: {str(e)}"
+
+
+
+
+
+translator = pipeline("translation", model="Helsinki-NLP/opus-mt-mul-en")
+
+@tool
+def translate_to_english(text: str) -> str:
+    """Translate input text in any language to English."""
+    try:
+        # HuggingFace translation expects a list of strings
+        translated = translator(text, max_length=512)
+        return translated[0]['translation_text']
+    except Exception as e:
+        return f"Translation error: {str(e)}"
+
+
+
+from langchain_core.tools import tool
+import sympy
+
+@tool
+def solve_algebraic_expression(expression: str) -> str:
+    """Solve or simplify the given algebraic expression."""
+    try:
+        expr = sympy.sympify(expression)
+        simplified = sympy.simplify(expr)
+        return str(simplified)
+    except Exception as e:
+        return f"Error solving expression: {str(e)}"
+
+
+from langchain_core.tools import tool
+
+@tool
+def run_python_code(code: str) -> str:
+    """Execute python code and return the result of variable 'result' if defined."""
+    try:
+        local_vars = {}
+        exec(code, {}, local_vars)
+        if 'result' in local_vars:
+            return str(local_vars['result'])
+        else:
+            return "Code executed successfully but no 'result' variable found."
+    except Exception as e:
+        return f"Error executing code: {str(e)}"
+        
+        
+from langchain_core.tools import tool
+import requests
+
+@tool
+def wikidata_query(sparql_query: str) -> str:
+    """Run a SPARQL query against Wikidata and return the JSON results."""
+    endpoint = "https://query.wikidata.org/sparql"
+    headers = {"Accept": "application/sparql-results+json"}
+    try:
+        response = requests.get(endpoint, params={"query": sparql_query}, headers=headers)
+        response.raise_for_status()
+        data = response.json()
+        return str(data)  # Or format as needed
+    except Exception as e:
+        return f"Error querying Wikidata: {str(e)}"
+
 
 
 
@@ -172,16 +292,23 @@ retriever_tool = create_retriever_tool(
 
 
 tools = [
+  
     multiply,
     add,
     subtract,
     divide,
     modulus,
     wiki_search,
+    filtered_wiki_search,
     web_search,
     arvix_search,
     wolfram_alpha_query,
-    retriever_tool
+    retriever_tool,
+    youtube_transcript,
+    translate_to_english,
+    solve_algebraic_expression,
+    run_python_code,
+    wikidata_query
 ]
 
 # Build graph function

requirements.txt

@@ -21,3 +21,8 @@ python-dotenv
 sentence-transformers
 openai
 anthropic
+
+youtube-transcript-api
+transformers
+sympy
+requests