Written custom base tool

agents/tools/base_tool.py

@@ -1,6 +1,51 @@
-from pydantic import BaseModel
-class BaseTool(BaseModel):
-    name: str
-    description: str
+class BaseTool:
+    """
+    Minimal base class for CrewAI or modular AI agent tools.
+
+    Inherit from this class for each custom tool to ensure a consistent interface.
+
+    Attributes:
+        config (dict or object, optional): Configuration provided on initialization.
+    """
+
+    def __init__(self, config=None):
+        """
+        Initialize the tool with optional configuration.
+
+        Args:
+            config (dict or object, optional): Configuration dictionary or object.
+        """
+        self.config = config
+
     def __call__(self, *args, **kwargs):
-        raise NotImplementedError("Tool must implement __call__.")
+        """
+        Abstract method for executing the tool's logic.
+        To be implemented in child classes.
+
+        Raises:
+            NotImplementedError: If not overridden in a subclass.
+        """
+        raise NotImplementedError(
+            f"{self.__class__.__name__} must implement the __call__ method."
+        )
+
+    @property
+    def name(self):
+        """
+        Returns the name of the tool class.
+
+        Returns:
+            str: Class name by default.
+        """
+        return self.__class__.__name__
+
+    def description(self):
+        """
+        Optionally provide a description string for the tool.
+
+        Returns:
+            str: Tool description, can be overridden by child classes.
+        """
+        return f"{self.name} (custom tool base class for CrewAI and modular agents)."
+
+    # You can add more utility methods/properties if you wish (e.g., metadata)

agents/tools/knowledge_tools.py

@@ -1,7 +1,7 @@
-# .base_tool import BaseTool
+from .base_tool import BaseTool
 from utils.knowledge_base import KnowledgeBase
 #from pydantic import BaseModel, PrivateAttr
-from crewai_tools.tool import BaseTool
+#from crewai_tools.tool import BaseTool
 
 
 
@@ -12,7 +12,7 @@ class SearchKnowledgeTool(BaseTool):
     def __init__(self, config=None):
         super().__init__()
         self.kb = KnowledgeBase(config)
-    def _run(self, query: str, k: int = 5):
+    def __call__(self, query: str, k: int = 5):
         return self.kb.search(query, k=k) if self.kb.is_initialized() else \
             [{"text": "General wisdom", "score": 1.0}]
 
@@ -21,7 +21,7 @@ class ExtractWisdomTool(BaseTool):
     description: str = "Extract most relevant wisdom for a given query."
     def __init__(self, config=None):
         super().__init__()
-    def _run(self, search_results: list, user_context: dict):
+    def __call__(self, search_results: list, user_context: dict):
         return search_results[:3]
 
 class SuggestPracticesTool(BaseTool):
@@ -29,7 +29,7 @@ class SuggestPracticesTool(BaseTool):
     description: str = "Recommend meditations or self-care practices."
     def __init__(self, config=None):
         super().__init__()
-    def _run(self, emotional_state: str, cultural_context: str = None):
+    def __call__(self, emotional_state: str, cultural_context: str = None):
         return {"name": "Mindful Breathing", "description": "Focus on your breath to calm the mind."}
 
 class KnowledgeTools:

agents/tools/llm_tools.py

@@ -1,7 +1,7 @@
-#from .base_tool import BaseTool
+from .base_tool import BaseTool
 from models.tinygpt2_model import TinyGPT2Model
 #from pydantic import BaseModel, PrivateAttr
-from crewai_tools.tool import BaseTool
+#from crewai_tools.tool import BaseTool
 
 
 
@@ -12,7 +12,7 @@ class MistralChatTool(BaseTool):
     def __init__(self, config=None):
         super().__init__()
         self.model = TinyGPT2Model()
-    def _run(self, prompt: str, context: dict = None):
+    def __call__(self, prompt: str, context: dict = None):
         msg = f"Context: {context}\nUser: {prompt}" if context else prompt
         return self.model.generate(msg)
 
@@ -23,7 +23,7 @@ class GenerateAdviceTool(BaseTool):
     def __init__(self, config=None):
         super().__init__()
         self.model = TinyGPT2Model()
-    def _run(self, user_analysis: dict, wisdom_quotes: list):
+    def __call__(self, user_analysis: dict, wisdom_quotes: list):
         prompt = f"Advice for: {user_analysis}, with wisdom: {wisdom_quotes}"
         return self.model.generate(prompt, max_length=300)
 
@@ -34,7 +34,7 @@ class SummarizeConversationTool(BaseTool):
     def __init__(self, config=None):
         super().__init__()
         self.model = TinyGPT2Model()
-    def _run(self, conversation: list):
+    def __call__(self, conversation: list):
         prompt = f"Summarize: {conversation}"
         return self.model.generate(prompt, max_length=200)
 

agents/tools/validation_tools.py

@@ -9,7 +9,8 @@ import json
 from transformers import pipeline
 import torch
 #from pydantic import BaseModel, PrivateAttr
-from crewai_tools.tool import BaseTool
+#from crewai_tools.tool import BaseTool
+from .base_tool import BaseTool
 
 
 
@@ -411,7 +412,7 @@ class ValidateResponseTool(BaseTool):
         super().__init__(**data)
         self.config = config
         # ... any required initialization ...
-    def _run(self, response: str, context: dict = None):
+    def __call__(self, response: str, context: dict = None):
         # Place your actual validation logic here, include dummy for illustration
         # For full validation logic, use your own code! 
         #     """Result of validation check"""

agents/tools/voice_tools.py

@@ -1,6 +1,6 @@
 import numpy as np
 import asyncio
-#from .base_tool import BaseTool
+from .base_tool import BaseTool
 from models.tinygpt2_model import TinyGPT2Model
 from transformers import pipeline, AutoProcessor, AutoModelForSpeechSeq2Seq
 import os
@@ -8,7 +8,7 @@ import tempfile
 import soundfile as sf
 import torch
 #from pydantic import BaseModel, PrivateAttr
-from crewai_tools.tool import BaseTool
+#from crewai_tools.tool import BaseTool
 
 
 
@@ -50,7 +50,7 @@ class TranscribeAudioTool(BaseTool):
     def __init__(self, config=None):
         super().__init__()
         self.vp = MultilingualVoiceProcessor()
-    def _run(self, audio_data: np.ndarray, language=None):
+    def  __call__(self, audio_data: np.ndarray, language=None):
         text, detected_lang = asyncio.run(self.vp.transcribe(audio_data, language))
         return {"text": text, "language": detected_lang}
 
@@ -59,7 +59,7 @@ class DetectEmotionTool(BaseTool):
     description: str = "Detect the emotional state from text."
     def __init__(self, config=None):
         super().__init__()
-    def _run(self, text: str):
+    def  __call__(self, text: str):
         model = TinyGPT2Model()
         prompt = f'Analyse emotions in: "{text}". Format: JSON with primary_emotion, intensity, feelings, concerns.'
         response = model.generate(prompt)
@@ -73,7 +73,7 @@ class GenerateReflectiveQuestionsTool(BaseTool):
     description: str = "Generate reflective questions."
     def __init__(self, config=None):
         super().__init__()
-    def _run(self, context: dict):
+    def  __call__(self, context: dict):
         emotion = context.get("primary_emotion", "neutral")
         questions_map = {
             "anxiety": ["What triggers your anxiety?", "How do you cope?"],

requirements.txt

@@ -1,5 +1,5 @@
  
-crewai-tools
+#crewai-tools
 crewai
 langchain
 langchain-community