quantalogic 0.30.1__tar.gz → 0.30.4__tar.gz

{quantalogic-0.30.1 → quantalogic-0.30.4}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: quantalogic
-Version: 0.30.1
+Version: 0.30.4
 Summary: QuantaLogic ReAct Agents
 Author: Raphaël MANSUY
 Author-email: raphael.mansuy@gmail.com
@@ -94,12 +94,10 @@ We created [QuantaLogic](https://www.quantalogic.app) because we saw a significa
 - **Memory Management**: Intelligent context handling and optimization
 - **Enterprise Ready**: Comprehensive logging, error handling, and validation system
 
-## Environment Configuration
-
-To configure the environment API key for Quantalogic using LiteLLM, set the required environment variable for your chosen provider (e.g., `OPENAI_API_KEY` for OpenAI, `ANTHROPIC_API_KEY` for Anthropic, or `DEEPSEEK_API_KEY` for DeepSeek) and any optional variables like `OPENAI_API_BASE` or `OPENROUTER_REFERRER`. Use a `.env` file or a secrets manager to securely store these keys, and load them in your code using `python-dotenv`. For advanced configurations, refer to the [LiteLLM documentation](https://docs.litellm.ai/docs/).
 
 ## 📋 Table of Contents
 
+- [Usage](#usage)
 - [Release Notes](#release-notes)
 
 - [Installation](#-installation)
@@ -114,12 +112,60 @@ To configure the environment API key for Quantalogic using LiteLLM, set the requ
 - [License](#-license)
 - [Documentation Development](#-documentation-development)
 
+## Usage
+
+**Usage:** `quantalogic [OPTIONS] COMMAND [ARGS]...`  
+**Environment Variables:** Set `OPENAI_API_KEY`, `ANTHROPIC_API_KEY`, and `DEEPSEEK_API_KEY` for API integration.
+
+**Options:**
+- `--model-name TEXT`: Specify the model (litellm format, e.g., "openrouter/deepseek/deepseek-chat")
+- `--log [info|debug|warning]`: Set logging level
+- `--mode [code|basic|interpreter|full|code-basic|search|search-full]`: Agent mode
+- `--vision-model-name TEXT`: Specify vision model (litellm format)
+- `--max-tokens-working-memory INTEGER`: Maximum tokens in working memory (default: 4000)
+- `--max-iterations INTEGER`: Maximum task iterations (default: 30)
+- `--compact-every-n-iteration INTEGER`: Compact memory every N iterations (default: 5)
+- `--no-stream`: Disable streaming output (default: enabled)
+- `--help`: Show help message
+
+**Commands:**
+- `task`: Execute a task with the QuantaLogic AI Assistant
+  - `--file PATH`: Path to task file
+  - `--model-name TEXT`: Specify model
+  - `--verbose`: Enable verbose output
+  - `--mode`: Select agent capabilities
+  - `--log`: Set logging level
+  - `--vision-model-name`: Specify vision model
+  - `--max-iterations`: Maximum task iterations
+  - `--max-tokens-working-memory`: Memory limit
+  - `--compact-every-n-iteration`: Memory optimization
+  - `--no-stream`: Disable streaming
+
 ## Release Notes
 
 See our [Release Notes](RELEASE_NOTES.MD) for detailed version history and changes.
 
 [TODO List](TODO.md)
 
+## Environment Configuration
+
+### Supported Models
+
+| Model Name | API Key Environment Variable | Description |
+|------------|------------------------------|-------------|
+| openai/gpt-4o-mini | OPENAI_API_KEY | OpenAI's compact version of GPT-4, optimized for efficiency and cost-effectiveness while maintaining strong performance. |
+| openai/gpt-4o | OPENAI_API_KEY | OpenAI's flagship model offering state-of-the-art performance across various tasks with enhanced reasoning capabilities. |
+| anthropic/claude-3.5-sonnet | ANTHROPIC_API_KEY | Claude 3.5 Sonnet model from Anthropic, balancing performance and speed with strong reasoning capabilities. |
+| deepseek/deepseek-chat | DEEPSEEK_API_KEY | DeepSeek's conversational model optimized for chat-based interactions and general-purpose tasks. |
+| deepseek/deepseek-reasoner | DEEPSEEK_API_KEY | DeepSeek's specialized model for complex reasoning tasks and problem-solving. |
+| openrouter/deepseek/deepseek-r1 | OPENROUTER_API_KEY | DeepSeek R1 model available through OpenRouter, optimized for research and development tasks. |
+| openrouter/openai/gpt-4o | OPENROUTER_API_KEY | OpenAI's GPT-4o model accessible through OpenRouter platform. |
+| openrouter/mistralai/mistral-large-2411 | OPENROUTER_API_KEY | Mistral's large model optimized for complex reasoning tasks, available through OpenRouter with enhanced multilingual capabilities. |
+| mistral/mistral-large-2407 | MISTRAL_API_KEY | Mistral's high-performance model designed for enterprise-grade applications, offering advanced reasoning and multilingual support. |
+
+To configure the environment API key for Quantalogic using LiteLLM, set the required environment variable for your chosen provider and any optional variables like `OPENAI_API_BASE` or `OPENROUTER_REFERRER`. Use a `.env` file or a secrets manager to securely store these keys, and load them in your code using `python-dotenv`. For advanced configurations, refer to the [LiteLLM documentation](https://docs.litellm.ai/docs/).
+
+
 ## 📦 Installation
 
 ### Prerequisites
@@ -155,64 +201,7 @@ pipx install quantalogic
 
 ### Basic Usage
 
-## 📖 CLI Reference
-
-The QuantaLogic CLI provides powerful command-line capabilities:
-
-```bash
-Usage: quantalogic [OPTIONS] COMMAND [ARGS]...
-  QuantaLogic AI Assistant - A powerful AI tool for various tasks.
-
-Options:
-  --version                       Show version information.
-  --model-name TEXT               Specify the model (litellm format, e.g., "openrouter/deepseek/deepseek-chat").
-  --log [info|debug|warning]      Set logging level.
-  --verbose                       Enable verbose output.
-  --mode [code|basic|interpreter|full|code-basic|search|search-full]  Agent mode.
-  --vision-model-name TEXT        Specify the vision model (litellm format, e.g., "openrouter/A/gpt-4o-mini").
-  --max-iterations INTEGER        Maximum iterations for task solving (default: 30).
-  --max-tokens-working-memory INTEGER Maximum tokens to keep in working memory (default: 4000).
-  --compact-every-n-iteration INTEGER Compact memory every N iterations (default: 5).
-  --help                          Show this message and exit.
-
-Commands:
-  task  Execute a task with the QuantaLogic AI Assistant.
-```
 
-### Commands
-task    Execute a task with the QuantaLogic AI Assistant
-
-**Usage:** `quantalogic task [OPTIONS] [TASK]`  
-**Description:** Execute a task with the QuantaLogic AI Assistant.  
-**Options:**  
-- `--file PATH`: Path to task file.  
-- `--model-name TEXT`: Specify the model (litellm format, e.g., `openrouter/deepseek/deepseek-chat`).  
-- `--verbose`: Enable verbose output.  
-- `--mode [code|basic|interpreter|full|code-basic|search|search-full]`: Agent mode.  
-- `--log [info|debug|warning]`: Set logging level.  
-- `--vision-model-name TEXT`: Specify the vision model (litellm format).  
-- `--max-iterations INTEGER`: Maximum iterations for task solving (default: 30).  
-- `--max-tokens-working-memory INTEGER`: Maximum tokens to keep in working memory (default: 4000).  
-- `--compact-every-n-iteration INTEGER`: Compact memory every N iterations (default: 5).  
-- `--no-stream`: Disable streaming output (default: enabled).  
-- `--help`: Show this message and exit.
-
-**Detailed Parameter Descriptions:**
-
-- **--model-name**: Specifies the LLM model to use (e.g., "openrouter/deepseek/deepseek-chat")
-- **--mode**: Selects agent capabilities:
-  - *code*: Coding-focused with basic capabilities
-  - *basic*: General-purpose without coding tools
-  - *interpreter*: Interactive code execution
-  - *full*: All capabilities enabled
-  - *code-basic*: Coding with basic reasoning
-  - *search*: Web search integration
-- **--log**: Controls logging verbosity (info, debug, warning)
-- **--vision-model-name**: Specifies vision model for image processing
-- **--max-iterations**: Limits task-solving attempts (default: 30)
-- **--max-tokens-working-memory**: Controls memory usage (default: None)
-- **--compact-every-n-iteration**: Memory optimization frequency (default: None)
-- **--no-stream**: Disables real-time output streaming
 
 
 
@@ -297,7 +286,6 @@ Example prompt: [04-write-a-tutorial.md](./examples/tasks/04-write-a-tutorial.md
 
 Here are some practical examples to help you get started:
 
-Here is the markdown table based on the provided directory listing:
 
 | Example | Description | File |
 |---------|-------------|------|
@@ -306,6 +294,12 @@ Here is the markdown table based on the provided directory listing:
 | Agent with Interpreter | An example of an agent that includes an interpreter. | [examples/03-agent-with-interpreter.py](examples/03-agent-with-interpreter.py) |
 | Agent Summary Task | An example of an agent performing a summary task. | [examples/04-agent-summary-task.py](examples/04-agent-summary-task.py) |
 | Code Example | A general code example. | [examples/05-code.py](examples/05-code.py) |
+| Code Screen Example | An example demonstrating code execution with screen output. | [examples/06-code-screen.py](examples/06-code-screen.py) |
+| Write Tutorial | An example of generating a tutorial using the agent. | [examples/07-write-tutorial.py](examples/07-write-tutorial.py) |
+| PRD Writer | An example of generating a Product Requirements Document (PRD). | [examples/08-prd-writer.py](examples/08-prd-writer.py) |
+| SQL Query | An example of executing SQL queries using the agent. | [examples/09-sql-query.py](examples/09-sql-query.py) |
+| Finance Agent | An example of a finance-focused agent. | [examples/10-finance-agent.py](examples/10-finance-agent.py) |
+| Textual Agent Interface | An example of a textual user interface for the agent. | [examples/11-textual-agent-interface.py](examples/11-textual-agent-interface.py) |
 
 
 ## 🔨 Key Components

{quantalogic-0.30.1 → quantalogic-0.30.4}/README.md

@@ -36,12 +36,10 @@ We created [QuantaLogic](https://www.quantalogic.app) because we saw a significa
 - **Memory Management**: Intelligent context handling and optimization
 - **Enterprise Ready**: Comprehensive logging, error handling, and validation system
 
-## Environment Configuration
-
-To configure the environment API key for Quantalogic using LiteLLM, set the required environment variable for your chosen provider (e.g., `OPENAI_API_KEY` for OpenAI, `ANTHROPIC_API_KEY` for Anthropic, or `DEEPSEEK_API_KEY` for DeepSeek) and any optional variables like `OPENAI_API_BASE` or `OPENROUTER_REFERRER`. Use a `.env` file or a secrets manager to securely store these keys, and load them in your code using `python-dotenv`. For advanced configurations, refer to the [LiteLLM documentation](https://docs.litellm.ai/docs/).
 
 ## 📋 Table of Contents
 
+- [Usage](#usage)
 - [Release Notes](#release-notes)
 
 - [Installation](#-installation)
@@ -56,12 +54,60 @@ To configure the environment API key for Quantalogic using LiteLLM, set the requ
 - [License](#-license)
 - [Documentation Development](#-documentation-development)
 
+## Usage
+
+**Usage:** `quantalogic [OPTIONS] COMMAND [ARGS]...`  
+**Environment Variables:** Set `OPENAI_API_KEY`, `ANTHROPIC_API_KEY`, and `DEEPSEEK_API_KEY` for API integration.
+
+**Options:**
+- `--model-name TEXT`: Specify the model (litellm format, e.g., "openrouter/deepseek/deepseek-chat")
+- `--log [info|debug|warning]`: Set logging level
+- `--mode [code|basic|interpreter|full|code-basic|search|search-full]`: Agent mode
+- `--vision-model-name TEXT`: Specify vision model (litellm format)
+- `--max-tokens-working-memory INTEGER`: Maximum tokens in working memory (default: 4000)
+- `--max-iterations INTEGER`: Maximum task iterations (default: 30)
+- `--compact-every-n-iteration INTEGER`: Compact memory every N iterations (default: 5)
+- `--no-stream`: Disable streaming output (default: enabled)
+- `--help`: Show help message
+
+**Commands:**
+- `task`: Execute a task with the QuantaLogic AI Assistant
+  - `--file PATH`: Path to task file
+  - `--model-name TEXT`: Specify model
+  - `--verbose`: Enable verbose output
+  - `--mode`: Select agent capabilities
+  - `--log`: Set logging level
+  - `--vision-model-name`: Specify vision model
+  - `--max-iterations`: Maximum task iterations
+  - `--max-tokens-working-memory`: Memory limit
+  - `--compact-every-n-iteration`: Memory optimization
+  - `--no-stream`: Disable streaming
+
 ## Release Notes
 
 See our [Release Notes](RELEASE_NOTES.MD) for detailed version history and changes.
 
 [TODO List](TODO.md)
 
+## Environment Configuration
+
+### Supported Models
+
+| Model Name | API Key Environment Variable | Description |
+|------------|------------------------------|-------------|
+| openai/gpt-4o-mini | OPENAI_API_KEY | OpenAI's compact version of GPT-4, optimized for efficiency and cost-effectiveness while maintaining strong performance. |
+| openai/gpt-4o | OPENAI_API_KEY | OpenAI's flagship model offering state-of-the-art performance across various tasks with enhanced reasoning capabilities. |
+| anthropic/claude-3.5-sonnet | ANTHROPIC_API_KEY | Claude 3.5 Sonnet model from Anthropic, balancing performance and speed with strong reasoning capabilities. |
+| deepseek/deepseek-chat | DEEPSEEK_API_KEY | DeepSeek's conversational model optimized for chat-based interactions and general-purpose tasks. |
+| deepseek/deepseek-reasoner | DEEPSEEK_API_KEY | DeepSeek's specialized model for complex reasoning tasks and problem-solving. |
+| openrouter/deepseek/deepseek-r1 | OPENROUTER_API_KEY | DeepSeek R1 model available through OpenRouter, optimized for research and development tasks. |
+| openrouter/openai/gpt-4o | OPENROUTER_API_KEY | OpenAI's GPT-4o model accessible through OpenRouter platform. |
+| openrouter/mistralai/mistral-large-2411 | OPENROUTER_API_KEY | Mistral's large model optimized for complex reasoning tasks, available through OpenRouter with enhanced multilingual capabilities. |
+| mistral/mistral-large-2407 | MISTRAL_API_KEY | Mistral's high-performance model designed for enterprise-grade applications, offering advanced reasoning and multilingual support. |
+
+To configure the environment API key for Quantalogic using LiteLLM, set the required environment variable for your chosen provider and any optional variables like `OPENAI_API_BASE` or `OPENROUTER_REFERRER`. Use a `.env` file or a secrets manager to securely store these keys, and load them in your code using `python-dotenv`. For advanced configurations, refer to the [LiteLLM documentation](https://docs.litellm.ai/docs/).
+
+
 ## 📦 Installation
 
 ### Prerequisites
@@ -97,64 +143,7 @@ pipx install quantalogic
 
 ### Basic Usage
 
-## 📖 CLI Reference
-
-The QuantaLogic CLI provides powerful command-line capabilities:
-
-```bash
-Usage: quantalogic [OPTIONS] COMMAND [ARGS]...
-  QuantaLogic AI Assistant - A powerful AI tool for various tasks.
-
-Options:
-  --version                       Show version information.
-  --model-name TEXT               Specify the model (litellm format, e.g., "openrouter/deepseek/deepseek-chat").
-  --log [info|debug|warning]      Set logging level.
-  --verbose                       Enable verbose output.
-  --mode [code|basic|interpreter|full|code-basic|search|search-full]  Agent mode.
-  --vision-model-name TEXT        Specify the vision model (litellm format, e.g., "openrouter/A/gpt-4o-mini").
-  --max-iterations INTEGER        Maximum iterations for task solving (default: 30).
-  --max-tokens-working-memory INTEGER Maximum tokens to keep in working memory (default: 4000).
-  --compact-every-n-iteration INTEGER Compact memory every N iterations (default: 5).
-  --help                          Show this message and exit.
-
-Commands:
-  task  Execute a task with the QuantaLogic AI Assistant.
-```
 
-### Commands
-task    Execute a task with the QuantaLogic AI Assistant
-
-**Usage:** `quantalogic task [OPTIONS] [TASK]`  
-**Description:** Execute a task with the QuantaLogic AI Assistant.  
-**Options:**  
-- `--file PATH`: Path to task file.  
-- `--model-name TEXT`: Specify the model (litellm format, e.g., `openrouter/deepseek/deepseek-chat`).  
-- `--verbose`: Enable verbose output.  
-- `--mode [code|basic|interpreter|full|code-basic|search|search-full]`: Agent mode.  
-- `--log [info|debug|warning]`: Set logging level.  
-- `--vision-model-name TEXT`: Specify the vision model (litellm format).  
-- `--max-iterations INTEGER`: Maximum iterations for task solving (default: 30).  
-- `--max-tokens-working-memory INTEGER`: Maximum tokens to keep in working memory (default: 4000).  
-- `--compact-every-n-iteration INTEGER`: Compact memory every N iterations (default: 5).  
-- `--no-stream`: Disable streaming output (default: enabled).  
-- `--help`: Show this message and exit.
-
-**Detailed Parameter Descriptions:**
-
-- **--model-name**: Specifies the LLM model to use (e.g., "openrouter/deepseek/deepseek-chat")
-- **--mode**: Selects agent capabilities:
-  - *code*: Coding-focused with basic capabilities
-  - *basic*: General-purpose without coding tools
-  - *interpreter*: Interactive code execution
-  - *full*: All capabilities enabled
-  - *code-basic*: Coding with basic reasoning
-  - *search*: Web search integration
-- **--log**: Controls logging verbosity (info, debug, warning)
-- **--vision-model-name**: Specifies vision model for image processing
-- **--max-iterations**: Limits task-solving attempts (default: 30)
-- **--max-tokens-working-memory**: Controls memory usage (default: None)
-- **--compact-every-n-iteration**: Memory optimization frequency (default: None)
-- **--no-stream**: Disables real-time output streaming
 
 
 
@@ -239,7 +228,6 @@ Example prompt: [04-write-a-tutorial.md](./examples/tasks/04-write-a-tutorial.md
 
 Here are some practical examples to help you get started:
 
-Here is the markdown table based on the provided directory listing:
 
 | Example | Description | File |
 |---------|-------------|------|
@@ -248,6 +236,12 @@ Here is the markdown table based on the provided directory listing:
 | Agent with Interpreter | An example of an agent that includes an interpreter. | [examples/03-agent-with-interpreter.py](examples/03-agent-with-interpreter.py) |
 | Agent Summary Task | An example of an agent performing a summary task. | [examples/04-agent-summary-task.py](examples/04-agent-summary-task.py) |
 | Code Example | A general code example. | [examples/05-code.py](examples/05-code.py) |
+| Code Screen Example | An example demonstrating code execution with screen output. | [examples/06-code-screen.py](examples/06-code-screen.py) |
+| Write Tutorial | An example of generating a tutorial using the agent. | [examples/07-write-tutorial.py](examples/07-write-tutorial.py) |
+| PRD Writer | An example of generating a Product Requirements Document (PRD). | [examples/08-prd-writer.py](examples/08-prd-writer.py) |
+| SQL Query | An example of executing SQL queries using the agent. | [examples/09-sql-query.py](examples/09-sql-query.py) |
+| Finance Agent | An example of a finance-focused agent. | [examples/10-finance-agent.py](examples/10-finance-agent.py) |
+| Textual Agent Interface | An example of a textual user interface for the agent. | [examples/11-textual-agent-interface.py](examples/11-textual-agent-interface.py) |
 
 
 ## 🔨 Key Components

{quantalogic-0.30.1 → quantalogic-0.30.4}/pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "quantalogic"
-version = "0.30.1"
+version = "0.30.4"
 description = "QuantaLogic ReAct Agents"
 authors = ["Raphaël MANSUY <raphael.mansuy@gmail.com>"]
 readme = "README.md"

{quantalogic-0.30.1 → quantalogic-0.30.4}/quantalogic/agent.py

@@ -467,27 +467,6 @@ class Agent(BaseModel):
             answer=None,
         )
 
-    def _format_observation_response(self, response: str, variable_name: str, iteration: int) -> str:
-        """Format the observation response with the given response, variable name, and iteration."""
-        response_display = response
-        if len(response) > MAX_RESPONSE_LENGTH:
-            response_display = response[:MAX_RESPONSE_LENGTH]
-            response_display += (
-                f"... content was truncated. Full content available by interpolation in variable {variable_name}"
-            )
-
-        formatted_response = (
-            "\n"
-            f"--- Observations for iteration {iteration} / max {self.max_iterations} ---\n"
-            "\n"
-            f"\n --- Tool execution result stored in variable ${variable_name}$ --- \n"
-            "\n"
-            f"<{variable_name}>\n{response_display}\n</{variable_name}>\n" + "\n"
-            "\n"
-            "--- Tools --- \n"
-        )
-        return formatted_response
-
     def _format_observation_response(self, response: str, variable_name: str, iteration: int) -> str:
         """Format the observation response with the given response, variable name, and iteration."""
         response_display = response
@@ -552,11 +531,11 @@ class Agent(BaseModel):
             )
 
             question_validation: str = (
-                "Do you permit the execution of this tool?"
-                f"Tool: {tool_name}"
-                f"Arguments: {arguments_with_values}"
+                "Do you permit the execution of this tool?\n"
+                f"Tool: {tool_name}\n"
+                f"Arguments: {arguments_with_values}\n"
                 "Yes or No"
-            ).join("\n")
+            )
             permission_granted = self.ask_for_user_validation(question_validation)
 
             self._emit_event(
@@ -579,21 +558,29 @@ class Agent(BaseModel):
             arguments_with_values_interpolated = {
                 key: self._interpolate_variables(value) for key, value in arguments_with_values.items()
             }
+            
+            # Convert arguments to correct types
+            try:
+                converted_args = self.tools._convert_kwargs_types(tool_name, **arguments_with_values_interpolated)
+            except ValueError as e:
+                logger.error(f"Type conversion failed: {str(e)}")
+                return "", f"Error: Type conversion failed for tool '{tool_name}': {str(e)}"
+
             # test if tool need variables in context 
             if tool.need_variables:
                 # Inject variables into the tool if needed
-                arguments_with_values_interpolated["variables"] = self.variable_store
+                converted_args["variables"] = self.variable_store
             if tool.need_caller_context_memory:
                 # Inject caller context into the tool if needed
-                arguments_with_values_interpolated["caller_context_memory"] = self.memory.memory
+                converted_args["caller_context_memory"] = self.memory.memory
 
             # Add injectable variables
             injectable_properties = tool.get_injectable_properties_in_execution()
             for key, value in injectable_properties.items():
-                arguments_with_values_interpolated[key] = value
+                converted_args[key] = value
 
             # Call tool execute with named arguments
-            response = tool.execute(**arguments_with_values_interpolated)
+            response = tool.execute(**converted_args)
             executed_tool = tool.name
         except Exception as e:
             response = f"Error executing tool: {tool_name}: {str(e)}\n"
@@ -702,7 +689,9 @@ class Agent(BaseModel):
         # Remove the last assistant / user message
         user_message = memory_copy.pop()
         assistant_message = memory_copy.pop()
-        summary = self.model.generate_with_history(messages_history=memory_copy, prompt=prompt_summary)
+        summary = self.model.generate_with_history(
+            messages_history=memory_copy, prompt=prompt_summary
+        )
         # Remove user message
         memory_copy.pop()
         # Replace by summary

{quantalogic-0.30.1 → quantalogic-0.30.4}/quantalogic/get_model_info.py

@@ -1,6 +1,8 @@
 model_info = {
     "deepseek-reasoner": {"max_output_tokens": 8 * 1024, "max_input_tokens": 1024 * 128},
     "openrouter/deepseek/deepseek-r1": {"max_output_tokens": 8 * 1024, "max_input_tokens": 1024 * 128},
+    "openrouter/mistralai/mistral-large-2411": {"max_output_tokens": 128 * 1024, "max_input_tokens": 1024 * 128},
+    "mistralai/mistral-large-2411": {"max_output_tokens": 128 * 1024, "max_input_tokens": 1024 * 128},
 }
 
 

{quantalogic-0.30.1 → quantalogic-0.30.4}/quantalogic/tool_manager.py

@@ -66,3 +66,45 @@ class ToolManager(BaseModel):
             markdown += "\n"
             index += 1
         return markdown
+
+    def _convert_kwargs_types(self, tool_name: str, **kwargs) -> dict:
+        """Convert kwargs values to their expected types based on tool definition.
+        
+        Args:
+            tool_name: Name of the tool to get argument types from
+            **kwargs: Input arguments to convert
+            
+        Returns:
+            Dictionary of converted arguments
+            
+        Raises:
+            ValueError: If type conversion fails for a required argument
+        """
+        tool = self.tools[tool_name]
+        converted = {}
+        
+        for arg in tool.arguments:
+            if arg.name in kwargs:
+                try:
+                    if arg.arg_type == "int":
+                        converted[arg.name] = int(kwargs[arg.name])
+                    elif arg.arg_type == "float":
+                        converted[arg.name] = float(kwargs[arg.name])
+                    elif arg.arg_type == "boolean":
+                        val = str(kwargs[arg.name]).lower()
+                        converted[arg.name] = val in ("true", "1", "yes", "y")
+                    else:  # string
+                        converted[arg.name] = str(kwargs[arg.name])
+                except (ValueError, TypeError) as e:
+                    if arg.required:
+                        raise ValueError(
+                            f"Failed to convert required argument '{arg.name}' to {arg.arg_type}: {str(e)}"
+                        )
+                    logger.warning(
+                        f"Failed to convert optional argument '{arg.name}' to {arg.arg_type}: {str(e)}"
+                    )
+                    converted[arg.name] = kwargs[arg.name]  # keep original value
+            elif arg.required and arg.default is None:
+                raise ValueError(f"Missing required argument: {arg.name}")
+                
+        return converted

{quantalogic-0.30.1 → quantalogic-0.30.4}/quantalogic/tools/search_definition_names.py

@@ -103,7 +103,13 @@ class SearchDefinitionNames(Tool):
             raise ValueError(f"Unsupported language: {language_name}")
 
     def execute(
-        self, directory_path: str, language_name: str, file_pattern: str = "*", output_format: str = "text", page: int = 1, page_size: int = 10
+        self,
+        directory_path: str,
+        language_name: str,
+        file_pattern: str = "*",
+        output_format: str = "text",
+        page: int = 1,
+        page_size: int = 10,
     ) -> Union[str, Dict]:
         """Searches for definition names in a directory using Tree-sitter.
 
@@ -119,6 +125,9 @@ class SearchDefinitionNames(Tool):
             Union[str, Dict]: The search results in the specified format.
         """
         try:
+            page = int(page)
+            page_size = int(page_size)
+
             # Validate pagination parameters
             if page < 1:
                 raise ValueError("Page number must be a positive integer.")
@@ -175,7 +184,7 @@ class SearchDefinitionNames(Tool):
                 output_format,
                 page,
                 page_size,
-                len(results)
+                len(results),
             )
 
         except Exception as e: