Update tools.py

tools.py

@@ -1,22 +1,6 @@
 from langchain_core.tools import tool
-from youtube_transcript_api import YouTubeTranscriptApi
-from SPARQLWrapper import SPARQLWrapper, JSON
-import json
 import pandas as pd
-import whisper
 import os
-import re
-import tempfile
-from transformers import Blip2Processor, Blip2ForConditionalGeneration
-from PIL import Image
-import torch
-
-whisper_model = whisper.load_model("base")
-device = "cuda" if torch.cuda.is_available() else "cpu"
-
-# Load model and processor once
-processor = Blip2Processor.from_pretrained("Salesforce/blip2-opt-2.7b")
-model = Blip2ForConditionalGeneration.from_pretrained("Salesforce/blip2-opt-2.7b", device_map="auto").to(device)
 
 
 @tool
@@ -31,7 +15,7 @@ def add(a: int, b: int) -> int:
     return a + b
 
 @tool
-def substract(a: int, b: int) -> int:
+def subtract(a: int, b: int) -> int:
     """
     Subtracts one value from another and returns the result of the sum
 
@@ -63,122 +47,4 @@ def divide(a: int, b: int) -> int:
     """
     if b == 0:
         raise ValueError("Cannot divide by zero.")
-    return a / b
-
-def match_referenced_files(question: str, uploaded_files: list) -> list:
-    """
-    Given a question string and a list of uploaded files with metadata,
-    return the list of files that are referenced in the question.
-
-    Parameters:
-        - question: str — The user's question.
-        - uploaded_files: list of dicts with keys:
-            - name (e.g., "data.xlsx")
-            - type (e.g., "excel", "image", "audio")
-            - path (e.g., "/tmp/data.xlsx")
-
-    Returns:
-        A list of file dicts that were referenced in the question.
-    """
-    # Find all mentions of possible filenames in the question (e.g., .xlsx, .png, .mp3)
-    referenced_names = set(re.findall(r'[\w\-\s]+\.(xlsx|xls|csv|png|jpg|jpeg|mp3|wav)', question, flags=re.IGNORECASE))
-
-    matched = []
-    for file in uploaded_files:
-        base_name = os.path.basename(file["name"]).lower()
-        if any(base_name.endswith(name.lower()) for name in referenced_names):
-            matched.append(file)
-
-    return matched
-
-
-@tool
-def analyze_image(image_path: str, question: str = "What’s in this image?") -> str:
-    """
-    Analyzes an image and answers a question about it using BLIP-2.
-    
-    Args:
-        image_path (str): Path to the image file.
-        question (str): A natural language question about the image.
-    """
-    try:
-        image = Image.open(image_path).convert("RGB")
-        inputs = processor(images=image, text=question, return_tensors="pt").to(device)
-        generated_ids = model.generate(**inputs, max_new_tokens=100)
-        response = processor.decode(generated_ids[0], skip_special_tokens=True)
-        return response
-    except Exception as e:
-        return f"Error analyzing image: {str(e)}"
-
-@tool
-def read_excel_summary(file_path: str) -> str:
-    """
-    Reads an Excel file and returns basic summary statistics, column names, and row count.
-
-    Args:
-        file_path: The path to the Excel file
-    """
-    try:
-        df = pd.read_excel(file_path, engine="openpyxl")
-        info = {
-            "columns": df.columns.tolist(),
-            "num_rows": len(df),
-            "summary": df.describe(include='all').to_dict()
-        }
-        return str(info)
-    except Exception as e:
-        return f"Error reading Excel file: {str(e)}"
-
-@tool
-def wiki_search(query: str) -> str:
-    """Query semantic knowledge bases using SPARQL
-    
-    Args:
-        query: The search query for SPARQL."""
-    try:
-        endpoint = SPARQLWrapper("https://query.wikidata.org/sparql")
-        endpoint.setQuery(query)
-        endpoint.setReturnFormat(JSON)
-
-        results = endpoint.query().convert()
-        return json.dumps(results, indent=2)
-    except Exception as e:
-        return f"Error executing SPARQL query: {str(e)}"
-
-@tool
-def transcribe_audio(file_path: str) -> str:
-    """
-    Transcribes an audio file (MP3, WAV, etc.) to text using Whisper.
-    Returns the transcribed text.
-
-    Args:
-        file_path: the path to the audio file
-    """
-    try:
-        result = whisper_model.transcribe(file_path)
-        return result["text"]
-    except Exception as e:
-        return f"Error during transcription: {str(e)}"
-
-@tool
-def fetch_youtube_transcript(video_url: str) -> str:
-    """
-    Fetch transcript of a YouTube video by URL.
-    Returns plain text if transcript is available.
-
-    Args:
-        video_url: The YouTube url link to the video
-    """
-    import re
-    video_id_match = re.search(r"(?:v=|youtu\.be/)([\w-]+)", video_url)
-    if not video_id_match:
-        return "Invalid YouTube URL."
-
-    video_id = video_id_match.group(1)
-
-    try:
-        transcript_list = YouTubeTranscriptApi.get_transcript(video_id, languages=['en'])
-        transcript_text = " ".join([entry['text'] for entry in transcript_list])
-        return transcript_text[:4000]  # Limit to fit into context window
-    except Exception as e:
-        return f"Transcript not available or error: {str(e)}"
+    return a / b