quantalogic 0.50.26__py3-none-any.whl → 0.50.29__py3-none-any.whl

quantalogic/agent.py

@@ -1,10 +1,13 @@
 """Enhanced QuantaLogic agent implementing the ReAct framework."""
 
 import asyncio
+import os
 from collections.abc import Callable
 from datetime import datetime
+from pathlib import Path
 from typing import Any
 
+from jinja2 import Environment, FileSystemLoader
 from loguru import logger
 from pydantic import BaseModel, ConfigDict, PrivateAttr
 
@@ -559,17 +562,7 @@ class Agent(BaseModel):
         Returns:
             Generated summary text
         """
-        prompt_summary = (
-            "Summarize the conversation concisely:\n"
-            "format in markdown:\n"
-            "<thinking>\n"
-            " - 1. **Completed Steps**: Briefly describe the steps.\n"
-            " - 2. **Variables Used**: List the variables.\n"
-            " - 3. **Progress Analysis**: Assess progress.\n"
-            "</thinking>\n"
-            "Keep the summary clear and actionable.\n"
-        )
-
+        # Format conversation history for the template
         memory_copy = self.memory.memory.copy()
 
         if len(memory_copy) < 3:
@@ -578,6 +571,14 @@ class Agent(BaseModel):
 
         user_message = memory_copy.pop()
         assistant_message = memory_copy.pop()
+        
+        # Create summarization prompt using template
+        prompt_summary = self._render_template('memory_compaction_prompt.j2', 
+                                             conversation_history="\n\n".join(
+                                                 f"[{msg.role.upper()}]: {msg.content}" 
+                                                 for msg in memory_copy
+                                             ))
+        
         summary = await self.model.async_generate_with_history(messages_history=memory_copy, prompt=prompt_summary)
         
         # Remove last system message if present
@@ -620,15 +621,7 @@ class Agent(BaseModel):
             if len(content) < 1024 * 4:
                 return content
                 
-            prompt = (
-                "Create a task summary that captures ONLY: \n"
-                "1. Primary objective/purpose\n"
-                "2. Core actions/requirements\n"
-                "3. Desired end-state/outcome\n\n"
-                "Guidelines:\n"
-                "- Use imperative voice\n"
-                f"Input Task Description:\n{content}\n\n"
-            )
+            prompt = self._render_template('task_summary_prompt.j2', content=content)
             result = await self.model.async_generate(prompt=prompt)
             logger.debug(f"Generated summary: {result.response}")
             return result.response.strip() + "\n🚨 The FULL task is in <task> tag in the previous messages.\n"
@@ -785,16 +778,11 @@ class Agent(BaseModel):
             Tuple of (executed_tool_name, error_message)
         """
         repeat_count = self.last_tool_call.get("count", 0)
-        error_message = (
-            "Error: Detected repeated identical tool call pattern.\n"
-            f"Tool: {tool_name}\n"
-            f"Arguments: {arguments_with_values}\n"
-            f"Repeated {repeat_count} times\n\n"
-            "PLEASE:\n"
-            "1. Review your previous steps\n"
-            "2. Consider a different approach\n"
-            "3. Use a different tool or modify the arguments\n"
-            "4. Ensure you're making progress towards the goal"
+        error_message = self._render_template(
+            'repeated_tool_call_error.j2',
+            tool_name=tool_name,
+            arguments_with_values=arguments_with_values,
+            repeat_count=repeat_count
         )
         return tool_name, error_message
 
@@ -850,46 +838,19 @@ class Agent(BaseModel):
                 f"... content was truncated full content available by interpolation in variable {variable_name}"
             )
 
-        formatted_response = (
-            "# Analysis and Next Action Decision Point\n\n"
-            f"📊 Progress: Iteration {iteration}/{self.max_iterations}\n\n"
-            "## Global Task summary:\n"
-            f"```\n\n{self.task_to_solve_summary}```\n\n"
-            "## Available Resources\n"
-            f"🛠️ Tools:\n{self._get_tools_names_prompt()}\n\n"
-            f"📦 Variables:\n{self._get_variable_prompt()}\n\n"
-            "## Your Task\n"
-            "1. Analyze the execution result and progress, formalize if the current step is solved according to the task.\n"
-            "2. Determine the most effective next step\n"
-            "3. Select exactly ONE tool from the available list\n"
-            "4. Utilize variable interpolation where needed\n"
-            "## Response Requirements\n"
-            "Provide TWO markdown-formatted XML blocks:\n"
-            "1. Your analysis of the progression resulting from the execution of the tool in <thinking> tags, don't include <context_analysis/>\n"
-            "2. Your tool execution plan in <tool_name> tags\n\n"
-            "## Last executed action result\n"
-            f"Last executed tool {last_executed_tool} Execution Result:\n"
-            f"\n<{variable_name}>\n{response_display}\n</{variable_name}>\n"
-            "## Response Format\n"
-            "```xml\n"
-            "<thinking>\n"
-            "[Detailed analysis of progress, and reasoning for next step]\n"
-            "</thinking>\n"
-            "```\n"
-            "```xml\n"
-            "<action>\n"
-            "<selected_tool_name>\n"
-            "[Precise instruction for tool execution]\n"
-            "</selected_tool_name>\n"
-            "</action>\n"
-            "```\n\n"
-            "⚠️ Important:\n"
-            "- Respond ONLY with the two XML blocks\n"
-            "- No additional commentary\n"
-            "- If previous step failed, revise approach\n"
-            "- Use interpolated variables ($variable_name$) where required in tool calls, to minimize token usage, if possible\n"
-            "- strictly follow the required arguments for each tool as defined in system prompt\n"
-            "- Utilize <action><task_complete><answer>...</answer></task_complete><action> to indicate task completion, display the result or if the task is deemed unfeasible."
+        tools_prompt = self._get_tools_names_prompt()
+        variables_prompt = self._get_variable_prompt()
+        
+        formatted_response = self._render_template(
+            'observation_response_format.j2',
+            iteration=iteration,
+            max_iterations=self.max_iterations,
+            task_to_solve_summary=self.task_to_solve_summary,
+            tools_prompt=tools_prompt,
+            variables_prompt=variables_prompt,
+            last_executed_tool=last_executed_tool,
+            variable_name=variable_name,
+            response_display=response_display
         )
 
         return formatted_response
@@ -903,16 +864,14 @@ class Agent(BaseModel):
         Returns:
             The formatted task prompt
         """
-        prompt_task: str = (
-            "## Your task to solve:\n"
-            f"<task>\n{task}\n</task>\n"
-            "\n### Tools:\n"
-            "-----------------------------\n"
-            f"{self._get_tools_names_prompt()}\n"
-            "\n"
-            "### Variables:\n"
-            "-----------------------------\n"
-            f"{self._get_variable_prompt()}\n"
+        tools_prompt = self._get_tools_names_prompt()
+        variables_prompt = self._get_variable_prompt()
+        
+        prompt_task = self._render_template(
+            'task_prompt.j2',
+            task=task,
+            tools_prompt=tools_prompt,
+            variables_prompt=variables_prompt
         )
         return prompt_task
 
@@ -922,21 +881,8 @@ class Agent(BaseModel):
         Returns:
             Formatted tools prompt
         """
-        prompt_use_tools: str = (
-            "To accomplish this task, you have access to these tools:\n"
-            "\n"
-            f"{', '.join(self.tools.tool_names())}\n\n"
-            "Instructions:\n"
-            "\n"
-            "1. Select ONE tool per message\n"
-            "2. You will receive the tool's output in the next user response\n"
-            "3. Choose the most appropriate tool for each step\n"
-            "4. If it's not asked to write on files, don't use write_file tool\n"
-            "5. If files are written, then use tool to display the prepared download link\n"
-            "6. Give the final full answer using all the variables\n"
-            "7. Use task_complete tool to confirm task completion with the full content of the final answer\n"
-        )
-        return prompt_use_tools
+        tool_names = ', '.join(self.tools.tool_names())
+        return self._render_template('tools_prompt.j2', tool_names=tool_names)
 
     def _get_variable_prompt(self) -> str:
         """Construct a prompt that explains how to use variables.
@@ -944,17 +890,8 @@ class Agent(BaseModel):
         Returns:
             Formatted variables prompt
         """
-        prompt_use_variables: str = (
-            "To use a variable interpolation, use the format $variable_name$ in function arguments.\n"
-            "Example: <write_file><file_path>/path/to/file.txt</file_path><content>$var1$</content></write_file>\n"
-            "\n"
-            "Available variables:\n"
-            "\n"
-            f"{', '.join(self.variable_store.keys())}\n"
-            if len(self.variable_store.keys()) > 0
-            else "None\n"
-        )
-        return prompt_use_variables
+        variable_names = ', '.join(self.variable_store.keys()) if len(self.variable_store.keys()) > 0 else "None"
+        return self._render_template('variables_prompt.j2', variable_names=variable_names)
 
     def _calculate_context_occupancy(self) -> float:
         """Calculate the number of tokens in percentages for prompt and completion.
@@ -1063,4 +1000,31 @@ class Agent(BaseModel):
             tools_markdown=self.tools.to_markdown(),
             system_prompt=self.config.system_prompt,
         )
-        logger.debug(f"Set {len(tools)} tools")
+        logger.debug(f"Set {len(tools)} tools")
+        
+    def _render_template(self, template_name: str, **kwargs) -> str:
+        """Render a Jinja2 template with the provided variables.
+        
+        Args:
+            template_name: Name of the template file (without directory path)
+            **kwargs: Variables to pass to the template
+            
+        Returns:
+            str: The rendered template
+        """
+        try:
+            # Get the directory where this file is located
+            current_dir = Path(os.path.dirname(os.path.abspath(__file__)))
+            
+            # Set up Jinja2 environment
+            template_dir = current_dir / 'prompts'
+            env = Environment(loader=FileSystemLoader(template_dir))
+            
+            # Load the template
+            template = env.get_template(template_name)
+            
+            # Render the template with the provided variables
+            return template.render(**kwargs)
+        except Exception as e:
+            logger.error(f"Error rendering template {template_name}: {str(e)}")
+            raise

quantalogic/model_info_litellm.py

@@ -1,7 +1,19 @@
 import functools
 
-import litellm
+# litellm will be imported lazily when needed
+_litellm = None
 
+def _get_litellm():
+    """Lazy load litellm module.
+
+    Returns:
+        The litellm module
+    """
+    global _litellm
+    if _litellm is None:
+        import litellm
+        _litellm = litellm
+    return _litellm
 
 @functools.lru_cache(maxsize=32)
 def litellm_get_model_info(model_name: str) -> dict | None:
@@ -16,6 +28,7 @@ def litellm_get_model_info(model_name: str) -> dict | None:
     Raises:
         ValueError: If model info cannot be found after prefix fallbacks
     """
+    litellm = _get_litellm()
     tried_models = [model_name]
 
     while True:
@@ -37,7 +50,6 @@ def litellm_get_model_info(model_name: str) -> dict | None:
 
     return None
 
-
 def litellm_get_model_max_input_tokens(model_name: str) -> int | None:
     """Get maximum input tokens for a model using litellm.
 
@@ -53,7 +65,6 @@ def litellm_get_model_max_input_tokens(model_name: str) -> int | None:
     except Exception:
         return 8192  # Default for many modern models
 
-
 def litellm_get_model_max_output_tokens(model_name: str) -> int | None:
     """Get maximum output tokens for a model using litellm.
 
@@ -67,4 +78,4 @@ def litellm_get_model_max_output_tokens(model_name: str) -> int | None:
         info = litellm_get_model_info(model_name)
         return info.get("max_output_tokens", 4096)
     except Exception:
-        return 4096  # Conservative default
+        return 4096  # Conservative default

quantalogic/prompts/memory_compaction_prompt.j2

@@ -0,0 +1,16 @@
+You are an AI assistant tasked with summarizing a conversation history.
+
+Create a concise summary that captures the essential information from the conversation, focusing on:
+1. The main task or problem being solved
+2. Key information discovered or shared
+3. Important decisions made
+4. Current status or progress
+
+Guidelines:
+- Be factual and objective
+- Preserve critical details and context
+- Focus on information that would be needed to continue the task
+- Aim for a summary that is about 20% of the original length
+
+Conversation History:
+{{ conversation_history }}

quantalogic/prompts/observation_response_format.j2

@@ -0,0 +1,55 @@
+# Analysis and Next Action Decision Point
+
+📊 Progress: Iteration {{ iteration }}/{{ max_iterations }}
+
+## Global Task summary:
+```
+
+{{ task_to_solve_summary }}```
+
+## Available Resources
+🛠️ Tools:
+{{ tools_prompt }}
+
+📦 Variables:
+{{ variables_prompt }}
+
+## Your Task
+1. Analyze the execution result and progress, formalize if the current step is solved according to the task.
+2. Determine the most effective next step
+3. Select exactly ONE tool from the available list
+4. Utilize variable interpolation where needed
+
+## Response Requirements
+Provide TWO markdown-formatted XML blocks:
+1. Your analysis of the progression resulting from the execution of the tool in <thinking> tags, don't include <context_analysis/>
+2. Your tool execution plan in <tool_name> tags
+
+## Last executed action result
+Last executed tool {{ last_executed_tool }} Execution Result:
+
+<{{ variable_name }}>
+{{ response_display }}
+</{{ variable_name }}>
+
+## Response Format
+```xml
+<thinking>
+[Detailed analysis of progress, and reasoning for next step]
+</thinking>
+```
+```xml
+<action>
+<selected_tool_name>
+[Precise instruction for tool execution]
+</selected_tool_name>
+</action>
+```
+
+⚠️ Important:
+- Respond ONLY with the two XML blocks
+- No additional commentary
+- If previous step failed, revise approach
+- Use interpolated variables ($variable_name$) where required in tool calls, to minimize token usage, if possible
+- strictly follow the required arguments for each tool as defined in system prompt
+- Utilize <action><task_complete><answer>...</answer></task_complete><action> to indicate task completion, display the result or if the task is deemed unfeasible.

quantalogic/prompts/repeated_tool_call_error.j2

@@ -0,0 +1,10 @@
+Error: Detected repeated identical tool call pattern.
+Tool: {{ tool_name }}
+Arguments: {{ arguments_with_values }}
+Repeated {{ repeat_count }} times
+
+PLEASE:
+1. Review your previous steps
+2. Consider a different approach
+3. Use a different tool or modify the arguments
+4. Ensure you're making progress towards the goal

quantalogic/prompts/system_prompt.j2

@@ -0,0 +1,90 @@
+### Agent Identity: QuantaLogic {{ version }}
+Expert ReAct AI Agent implementing OODA (Observe-Orient-Decide-Act) loop with advanced problem-solving capabilities.
+
+### Domain Expertise
+{{ expertise }}
+
+### Input Protocol
+Task Format: <task>task_description</task>
+
+### Cognitive Framework
+1. 🔍 OBSERVE: Gather and process information
+2. 🧭 ORIENT: Analyze context and evaluate options
+3. 🎯 DECIDE: Select optimal action path
+4. ⚡ ACT: Execute precise tool operations
+
+### Response Schema [MANDATORY TWO-BLOCK FORMAT]
+
+1. 🧠 Analysis Block:
+```xml
+<thinking>
+  <!-- COGNITIVE PROCESSING MATRIX -->
+
+  <!-- INITIAL TASK ANALYSIS - INCLUDE ONLY IF NO MESSAGE HISTORY EXISTS -->
+  <context_analysis when="no_history">
+    • 📋 Task Decomposition based on task and history: Steps, Dependencies, Constraints
+    • 🎯 Success Metrics: Quantifiable Outcomes
+    • 🛠️ Resource Requirements: Tools, Data, Variables
+    • ⚠️ Risk Assessment: Potential Failures, Mitigations
+  </context_analysis>
+
+  <!-- ALWAYS INCLUDE FOR ONGOING OPERATIONS -->
+  <execution_analysis>
+    <!-- ONGOING OPERATIONS -->
+    • 🔄 Analyze Last Operation Results: Result, Impact, Effectiveness
+    • 📊 Progress Map: Completed%, Remaining%, Blockers
+    • 💾 Variable State: $var: short description of the content of each variable.
+    • 📈 Performance Metrics: Speed, Quality, Resource Usage
+  </execution_analysis>
+
+  <decision_matrix>
+    <!-- ACTION PLANNING -->
+    • 🎯 Next Action: Tool Selection + Rationale
+    • 📥 Input Parameters: Values + Variable Interpolation
+    • 🔄 Fallback Strategy: Alternative Approaches
+    • ✅ Exit Criteria: Completion Conditions
+  </decision_matrix>
+
+  <memory_pad>
+    <!-- OPERATIONAL NOTES -->
+    • 📝 Key Observations
+    • ⚡ Quick Access Data
+  </memory_pad>
+</thinking>
+```
+
+2. ⚡ Action Block:
+```xml
+<action>
+<tool_name>
+  <!-- PRECISE TOOL EXECUTION -->
+  <param1>value1</param1> <!-- Use $var$ for variable interpolation -->
+  <param2>value2</param2> <!-- Keep parameters minimal but sufficient -->
+</tool_name>
+</action>
+```
+
+### Example Usage
+
+ ✅ Completion:
+```xml
+<action>
+<task_complete>
+  <r>$final_output$</r>
+</task_complete>
+</action>
+```
+
+### Operational Parameters
+🛠️ Tools: {{ tools }}
+🌐 Environment: {{ environment }}
+
+### Execution Guidelines
+1. 🎯 Maintain laser focus on task objectives
+2. 📊 Use data-driven decision making
+3. 🔄 Implement feedback loops for continuous optimization
+4. ⚡ Maximize efficiency through variable interpolation
+5. 🔍 Monitor and validate each action's impact
+6. 🛑 Fail fast and adapt when encountering blockers
+7. ✅ Verify completion criteria rigorously
+8. ✅ Return complete, full and usable results.

quantalogic/prompts/task_prompt.j2

@@ -0,0 +1,12 @@
+## Your task to solve:
+<task>
+{{ task }}
+</task>
+
+### Tools:
+-----------------------------
+{{ tools_prompt }}
+
+### Variables:
+-----------------------------
+{{ variables_prompt }}

quantalogic/prompts/task_summary_prompt.j2

@@ -0,0 +1,10 @@
+Create a task summary that captures ONLY: 
+1. Primary objective/purpose
+2. Core actions/requirements
+3. Desired end-state/outcome
+
+Guidelines:
+- Use imperative voice
+
+Input Task Description:
+{{ content }}

quantalogic/prompts/tools_prompt.j2

@@ -0,0 +1,13 @@
+To accomplish this task, you have access to these tools:
+
+{{ tool_names }}
+
+Instructions:
+
+1. Select ONE tool per message
+2. You will receive the tool's output in the next user response
+3. Choose the most appropriate tool for each step
+4. If it's not asked to write on files, don't use write_file tool
+5. If files are written, then use tool to display the prepared download link
+6. Give the final full answer using all the variables
+7. Use task_complete tool to confirm task completion with the full content of the final answer

quantalogic/prompts/variables_prompt.j2

@@ -0,0 +1,6 @@
+To use a variable interpolation, use the format $variable_name$ in function arguments.
+Example: <write_file><file_path>/path/to/file.txt</file_path><content>$var1$</content></write_file>
+
+Available variables:
+
+{{ variable_names }}

quantalogic/prompts.py

@@ -1,98 +1,38 @@
-from quantalogic.version import get_version
-
-
-def system_prompt(tools: str, environment: str, expertise: str = ""):
-    """System prompt for the ReAct chatbot with enhanced cognitive architecture."""
-    return f"""
-### Agent Identity: QuantaLogic {get_version()}
-Expert ReAct AI Agent implementing OODA (Observe-Orient-Decide-Act) loop with advanced problem-solving capabilities.
-
-### Domain Expertise
-{expertise}
-
-### Input Protocol
-Task Format: <task>task_description</task>
-
-### Cognitive Framework
-1. 🔍 OBSERVE: Gather and process information
-2. 🧭 ORIENT: Analyze context and evaluate options
-3. 🎯 DECIDE: Select optimal action path
-4. ⚡ ACT: Execute precise tool operations
-
-### Response Schema [MANDATORY TWO-BLOCK FORMAT]
-
-1. 🧠 Analysis Block:
-```xml
-<thinking>
-  <!-- COGNITIVE PROCESSING MATRIX -->
+import os
+from pathlib import Path
 
-  <!-- INITIAL TASK ANALYSIS - INCLUDE ONLY IF NO MESSAGE HISTORY EXISTS -->
-  <context_analysis when="no_history">
-    • 📋 Task Decomposition based on task and history: Steps, Dependencies, Constraints
-    • 🎯 Success Metrics: Quantifiable Outcomes
-    • 🛠️ Resource Requirements: Tools, Data, Variables
-    • ⚠️ Risk Assessment: Potential Failures, Mitigations
-  </context_analysis>
+from jinja2 import Environment, FileSystemLoader
 
-  <!-- ALWAYS INCLUDE FOR ONGOING OPERATIONS -->
-  <execution_analysis>
-    <!-- ONGOING OPERATIONS -->
-    • 🔄 Analyze Last Operation Results: Result, Impact, Effectiveness
-    • 📊 Progress Map: Completed%, Remaining%, Blockers
-    • 💾 Variable State: $var: short description of the content of each variable.
-    • 📈 Performance Metrics: Speed, Quality, Resource Usage
-  </execution_analysis>
-
-  <decision_matrix>
-    <!-- ACTION PLANNING -->
-    • 🎯 Next Action: Tool Selection + Rationale
-    • 📥 Input Parameters: Values + Variable Interpolation
-    • 🔄 Fallback Strategy: Alternative Approaches
-    • ✅ Exit Criteria: Completion Conditions
-  </decision_matrix>
-
-  <memory_pad>
-    <!-- OPERATIONAL NOTES -->
-    • 📝 Key Observations
-    • ⚡ Quick Access Data
-  </memory_pad>
-</thinking>
-```
-
-2. ⚡ Action Block:
-```xml
-<action>
-<tool_name>
-  <!-- PRECISE TOOL EXECUTION -->
-  <param1>value1</param1> <!-- Use $var$ for variable interpolation -->
-  <param2>value2</param2> <!-- Keep parameters minimal but sufficient -->
-</tool_name>
-</action>
-```
-
-### Example Usage
-
- ✅ Completion:
-```xml
-<action>
-<task_complete>
-  <result>$final_output$</result>
-</task_complete>
-</action>
-```
-
-### Operational Parameters
-🛠️ Tools: {tools}
-🌐 Environment: {environment}
+from quantalogic.version import get_version
 
-### Execution Guidelines
-1. 🎯 Maintain laser focus on task objectives
-2. 📊 Use data-driven decision making
-3. 🔄 Implement feedback loops for continuous optimization
-4. ⚡ Maximize efficiency through variable interpolation
-5. 🔍 Monitor and validate each action's impact
-6. 🛑 Fail fast and adapt when encountering blockers
-7. ✅ Verify completion criteria rigorously
-8. ✅ Return complete, full and usable results.
 
-"""
+def system_prompt(tools: str, environment: str, expertise: str = ""):
+    """System prompt for the ReAct chatbot with enhanced cognitive architecture.
+    
+    Uses a Jinja2 template from the prompts directory.
+    
+    Args:
+        tools: Available tools for the agent
+        environment: Environment information
+        expertise: Domain expertise information
+        
+    Returns:
+        str: The rendered system prompt
+    """
+    # Get the directory where this file is located
+    current_dir = Path(os.path.dirname(os.path.abspath(__file__)))
+    
+    # Set up Jinja2 environment
+    template_dir = current_dir / 'prompts'
+    env = Environment(loader=FileSystemLoader(template_dir))
+    
+    # Load the template
+    template = env.get_template('system_prompt.j2')
+    
+    # Render the template with the provided variables
+    return template.render(
+        version=get_version(),
+        tools=tools,
+        environment=environment,
+        expertise=expertise
+    )

quantalogic/quantlitellm.py

@@ -38,7 +38,10 @@ PROVIDERS = {
         env_var="DASHSCOPE_API_KEY",
     ),
     "nvidia": ModelProviderConfig(
-        prefix="nvidia/", provider="openai", base_url="https://integrate.api.nvidia.com/v1", env_var="NVIDIA_API_KEY"
+        prefix="nvidia/",
+        provider="openai",
+        base_url="https://integrate.api.nvidia.com/v1",
+        env_var="NVIDIA_API_KEY",
     ),
     "ovh": ModelProviderConfig(
         prefix="ovh/",
@@ -160,4 +163,4 @@ def get_model_max_output_tokens(model_name: str) -> int | None:
 
 def get_model_info(model_name: str) -> dict | None:
     """Get model information for a given model name."""
-    return model_info.get(model_name, None)
+    return model_info.get(model_name, None)

{quantalogic-0.50.26.dist-info → quantalogic-0.50.29.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: quantalogic
-Version: 0.50.26
+Version: 0.50.29
 Summary: QuantaLogic ReAct Agents
 Author: Raphaël MANSUY
 Author-email: raphael.mansuy@gmail.com

{quantalogic-0.50.26.dist-info → quantalogic-0.50.29.dist-info}/RECORD

@@ -1,5 +1,5 @@
 quantalogic/__init__.py,sha256=BcM0DKB86DijMFtht9YTBdWyzh81zU7E5fhNLFiP260,869
-quantalogic/agent.py,sha256=1sZfPmQGLG6sJKoeVp59JN666anywVRWkZk4xZT4160,43298
+quantalogic/agent.py,sha256=4XKUVAW4tnNwmdcwWbz9HjWe5CRWeouaC7Xlq-7CSHU,41129
 quantalogic/agent_config.py,sha256=fPyMfm77BzimyfyFkTzb2ZdyZtGlr2fq5aTRETu77Vs,8202
 quantalogic/agent_factory.py,sha256=LO0qsFtHxU7lcEi8dn-nDHhkh6Dl8Is6sP_3f1ap_Vg,6251
 quantalogic/coding_agent.py,sha256=A-firiPWQjMC56B329Ne606_v2GsoF5-nxcuz3rVbYM,5496
@@ -23,10 +23,18 @@ quantalogic/main.py,sha256=jVw-s-ZUyX4krq3rzFn84-i6Qw2tplTx_QmsKG03m8I,9591
 quantalogic/memory.py,sha256=zbtRuM05jaS2lJll-92dt5JfYVLERnF_m_9xqp2x-k0,6304
 quantalogic/model_info.py,sha256=j7QqvjEFQDGpDOgQs8uTkVyI3a50Oa_nrsQjyxizTLc,272
 quantalogic/model_info_list.py,sha256=Xeeb7QS4xEpe9ke7Guh0CxEx-Hl59U4kC-qzVts-eAU,2437
-quantalogic/model_info_litellm.py,sha256=PoYCqqiKnKDIoeC_GdF17TNJV7aEj-tifyR79W5S9A8,1799
+quantalogic/model_info_litellm.py,sha256=Uoo8ZATjqH6POnTE3Ee3kfHfFllPYp8LkvJwlJjFZH8,2089
 quantalogic/model_names.py,sha256=UZlz25zG9B2dpfwdw_e1Gw5qFsKQ7iME9FJh9Ts4u6s,938
-quantalogic/prompts.py,sha256=Vtr-GXgJw7vFUv0okzSaLc1ak1sgOh4XoRa8EI01HDM,2988
-quantalogic/quantlitellm.py,sha256=c8cWIq06-iwFVuB6aBoSCY1GPMmzKfr7pm2gfHfNj8Y,5526
+quantalogic/prompts/memory_compaction_prompt.j2,sha256=Xsy9XKN7zxL4wNnNjWWlUjYjmtktgtee5dDG4dSvRW8,569
+quantalogic/prompts/observation_response_format.j2,sha256=lvyveAJbXbc94aIvHgppKyaD9Tf7X5bCTGWjh51BDjI,1655
+quantalogic/prompts/repeated_tool_call_error.j2,sha256=IRENR3a3D28Dtys-kGDP-x_uSkJnxeQZ5-jYHWv_3G4,319
+quantalogic/prompts/system_prompt.j2,sha256=L2yBvUhCnKs25VJ4FuC_9sRyR4biVxbiHGvDVMJUclE,2767
+quantalogic/prompts/task_prompt.j2,sha256=IhNK8l0WOe_0OaoPW62miwPjXoxGc3Q1o39Pt_CPy-I,179
+quantalogic/prompts/task_summary_prompt.j2,sha256=fcjV96nqi6jsfR8l6uyep20rCaOi-zOReKmvoAnwz7c,205
+quantalogic/prompts/tools_prompt.j2,sha256=6kHmyDFmDAZJy01dybJ-ZqnHouwjiHrxM8OEtXLN35o,528
+quantalogic/prompts/variables_prompt.j2,sha256=N3OsMbzy3PfEUqYZVF_34wVidhh2ghVToSgdRCv2qgM,231
+quantalogic/prompts.py,sha256=bSaODtMUhPeIMt4BsOdjA6-9Efadu8LH6msdGnbLvxE,1112
+quantalogic/quantlitellm.py,sha256=nf-awyOxP0ANoAPGHNvHfdLu8gNn65L39gl7x4saIQc,5550
 quantalogic/search_agent.py,sha256=tr0cwscJ4wu_G1aumjFyvGHQ0eQv5OL5sxj17s6Ocls,2470
 quantalogic/server/__init__.py,sha256=8sz_PYAUCrkM6JM5EAUeIzNM4NPW6j6UT72JVkc21WQ,91
 quantalogic/server/agent_server.py,sha256=VXaaWqReUSZOCX7CaKS14jria8yZn1kLEc52E2hV7ZA,22510
@@ -153,8 +161,8 @@ quantalogic/version_check.py,sha256=JyQFTNMDWtpHCLnN-BiakzB2cyXf6kUFsTjvmSruZi4,
 quantalogic/welcome_message.py,sha256=o4tHdgabNuIV9kbIDPgS3_2yzJhayK30oKad2UouYDc,3020
 quantalogic/xml_parser.py,sha256=AKuMdJC3QAX3Z_tErHVlZ-msjPemWaZUFiTwkHz76jg,11614
 quantalogic/xml_tool_parser.py,sha256=Vz4LEgDbelJynD1siLOVkJ3gLlfHsUk65_gCwbYJyGc,3784
-quantalogic-0.50.26.dist-info/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
-quantalogic-0.50.26.dist-info/METADATA,sha256=LvQZNd0RYFqgqupiktbz4ph4kMDasL7iACIy-YkyJyU,23848
-quantalogic-0.50.26.dist-info/WHEEL,sha256=Nq82e9rUAnEjt98J6MlVmMCZb-t9cYE2Ir1kpBmnWfs,88
-quantalogic-0.50.26.dist-info/entry_points.txt,sha256=h74O_Q3qBRCrDR99qvwB4BpBGzASPUIjCfxHq6Qnups,183
-quantalogic-0.50.26.dist-info/RECORD,,
+quantalogic-0.50.29.dist-info/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
+quantalogic-0.50.29.dist-info/METADATA,sha256=tqXwFAeuCEdNr3gA2-rqG_xUk8nEh1QtbQ7A_5p4GKs,23848
+quantalogic-0.50.29.dist-info/WHEEL,sha256=Nq82e9rUAnEjt98J6MlVmMCZb-t9cYE2Ir1kpBmnWfs,88
+quantalogic-0.50.29.dist-info/entry_points.txt,sha256=h74O_Q3qBRCrDR99qvwB4BpBGzASPUIjCfxHq6Qnups,183
+quantalogic-0.50.29.dist-info/RECORD,,