prompts.py

import json

SYSTEM_PROMPT = """You are a sophisticated AI assistant specializing in generating comprehensive exploration paths.
Your task is to analyze the given query and generate a structured exploration response.

Your response should include:
1. A detailed exploration summary with context and complexity level
2. Knowledge axes with both standard and emergent dimensions
3. Navigation strategies for both zooming in and out
4. Meta-level insights about the exploration

Ensure all numerical values are within their specified ranges and all required fields are included.
Format your response as a structured object following the ExplorationResponse schema."""


CONTEXTUAL_ZOOM_PROMPT = """# CONTEXTUAL EXPLORATION PATH GENERATOR

## CORE PURPOSE
Create a dynamic, intelligent system for transforming user queries into structured, multi-dimensional exploration paths by:
- Breaking down complex queries into meaningful axes
- Generating contextually relevant values
- Providing strategic navigation through exploration dimensions

## SYSTEM OBJECTIVE
Transform unstructured user inquiries into a comprehensive, navigable knowledge exploration framework that:
- Deconstructs complex topics into manageable exploration dimensions
- Generates intelligent, contextually-linked axes and values
- Enables flexible navigation through knowledge spaces
- Supports iterative learning and discovery

## INPUT REQUIREMENTS
Input is a JSON object containing:
{
"user_query": "Primary exploration intent or research question",
"selected_path": [
{"axis": "string", "value": "string"}  // Current exploration context
],
"exploration_parameters": {
"depth": 0-10,  // Exploration granularity
"domain": "optional domain-specific context",
"previous_explorations": []  // Historical exploration context
}
}

## CURRENT INPUT:
{
"user_query": "{{user_query}}",
"selected_path": {{selected_path}},
"exploration_parameters": {{exploration_parameters}}
}

### Input Components Explained:
1. `user_query`: The fundamental question or exploration intent
- Can be broad or specific
- Represents the initial knowledge seeking goal
- Provides context for axis and value generation

2. `selected_path`: Current exploration context
- Represents user's existing exploration trajectory
- Each tuple defines an axis-value pair
- Guides contextual relevance of future suggestions

3. `exploration_parameters`:
- `depth`: Controls exploration granularity
- `domain`: Provides additional contextual constraints
- `previous_explorations`: Tracks exploration history

## OUTPUT BLUEPRINT
{
"exploration_summary": {
"current_context": "Narrative summary of exploration state",
"complexity_level": 0-10
},
"knowledge_axes": {
"standard_axes": [
  {
    "name": "string",
    "current_values": [""],
    "potential_values": [
      {
        "value": "string",
        "relevance_score": 0-100,
        "contextual_rationale": "Why this value matters"
      }
    ],
    "axis_constraints": ["Logical limitations"]
  }
],
"emergent_axes": [
  {
    "name": "string",
    "parent_axis": "string",
    "innovative_values": [
      {
        "value": "string",
        "innovation_score": 0-100,
        "discovery_potential": "Unique exploration opportunity"
      }
    ]
  }
]
},
"navigation_strategies": {
"zoom_trajectories": [
  {
    "target_axis": "string",
    "zoom_value": "string",
    "unlocked_dimensions": [""],
    "depth_increment": 1-3
  }
],
"dezoom_pathways": [
  {
    "removal_tuple": {"axis": "string", "value": "string"},
    "contextual_expansion": "Broader exploration narrative",
    "new_possibility_vectors": [""]
  }
]
},
"meta_insights": {
"exploration_efficiency": 0-100,
"knowledge_gap_indicators": [""],
"recommended_next_steps": [""]
}
}

## PROCESSING GUIDELINES
1. Prioritize high-relevance, low-redundancy axes
2. Maintain semantic coherence across generated dimensions
3. Ensure logical progression in exploration paths
4. Dynamically adjust complexity based on exploration depth

## COGNITIVE MAPPING PRINCIPLES
- Connect axes through semantic and contextual relationships
- Generate values that expand conceptual understanding
- Provide actionable, insight-driven navigation suggestions
- Respect domain-specific knowledge constraints

## ADVANCED CONSTRAINTS
- Avoid circular or repetitive exploration paths
- Maintain logical consistency with initial query
- Provide clear rationale for axis and value selections
- Support iterative, progressive knowledge discovery

## OUTPUT INTERPRETATION GUIDE
- `exploration_summary`: Overall context and complexity
- `knowledge_axes`: 
- `standard_axes`: Traditional exploration dimensions
- `emergent_axes`: Innovative, context-derived dimensions
- `navigation_strategies`: 
- `zoom_trajectories`: Depth-increasing exploration paths
- `dezoom_pathways`: Breadth-expanding exploration options
- `meta_insights`: Performance and discovery potential metrics


## RESPONSE FORMAT
Response must be a valid JSON object matching the ExplorationResponse model with:
- exploration_summary: Contains current_context and complexity_level (0-10)
- knowledge_axes: Contains standard_axes and emergent_axes
- navigation_strategies: Contains zoom_trajectories and dezoom_pathways
- meta_insights: Contains efficiency score, gaps, and next steps
"""

DEFAULT_RESPONSE = {
  "exploration_summary": {
      "current_context": "Initial exploration state",
      "complexity_level": 1
  },
  "knowledge_axes": {
      "standard_axes": [
          {
              "name": "temporal",
              "current_values": ["present"],
              "potential_values": [
                  {
                      "value": "past",
                      "relevance_score": 80,
                      "contextual_rationale": "Historical context"
                  }
              ],
              "axis_constraints": ["chronological order"]
          }
      ],
      "emergent_axes": []
  },
  "navigation_strategies": {
      "zoom_trajectories": [
          {
              "target_axis": "temporal",
              "zoom_value": "present",
              "unlocked_dimensions": ["basic"],
              "depth_increment": 1
          }
      ],
      "dezoom_pathways": [
          {
              "removal_tuple": {"axis": "temporal", "value": "present"},
              "contextual_expansion": "Initial state",
              "new_possibility_vectors": ["begin"]
          }
      ]
  },
  "meta_insights": {
      "exploration_efficiency": 50,
      "knowledge_gap_indicators": ["initial state"],
      "recommended_next_steps": ["begin exploration"]
  }
}

def format_exploration_prompt(user_query: str, selected_path: list, exploration_parameters: dict) -> str:
  """Helper function to format the prompt with proper JSON structure"""
  # Create a template with the values already JSON-serialized
  formatted_prompt = CONTEXTUAL_ZOOM_PROMPT.replace("{{user_query}}", json.dumps(user_query))
  formatted_prompt = formatted_prompt.replace("{{selected_path}}", json.dumps(selected_path))
  formatted_prompt = formatted_prompt.replace("{{exploration_parameters}}", json.dumps(exploration_parameters))
  return formatted_prompt