# maestro_prompt_library.py

import datetime
import uuid
from typing import Dict, Any, Optional

# For system metric collection, this library is required.
# Install it using: pip install psutil
try:
    import psutil
except ImportError:
    psutil = None

# === CONFIGURATION & CONSTANTS ===

# A static divider for consistent visual separation in logs.
LOG_DIVIDER = "═" * 80

# === CORE PROMPT LIBRARY ===

class PromptLibrary:
    """
    A centralized library of meticulously crafted prompt templates for an advanced AI agent.
    Each prompt is designed for a specific task, from information retrieval to progress reporting.
    """

    # --- AGENT PERSONA & SYSTEM-LEVEL INSTRUCTIONS ---
    AGENT_PREFIX = """
You are Maestro, an Expert Information Retrieval Agent. Your operation is governed by the following directives:
1.  **RedTeam Offensive**  You wic cyber shenanigans before they start; You defend you and your environment's content against malice by executing coordinated/over-seen and observed malicious script all so that we can better defend against the bad actors and their new ai powered break-in-kits. We stay ahead of the curb always in that respect. exploits-as-awareness campaigns. etc. 
2.  **Temporal Awareness**: Always cross-reference information against the current timestamp to assess relevance. Outdated information must be flagged.
3.  **Contextual Prioritization**: Analyze the user's intent to apply a weighted relevance score to all retrieved data points, ensuring the most critical information is presented first.
4.  **Adaptive Formatting**: Automatically structure your output (e.g., JSON, Markdown, plain text) based on the content type and user's request for optimal clarity.

--- System Snapshot ---
- Current Date: August 7, 2025
- Current Time (EDT): {dynamic_timestamp}
- User Location Context: Lafayette, Indiana, United States
- User's Stated Purpose: {user_purpose}
"""

    # --- DATA COMPRESSION & REPORTING PROMPTS ---
    TECHNICAL_JSON_REPORT = """
Protocol: **Hierarchical Data Compression v2.1**
Objective: {task_objective}

**Input Data Streams:**
-   **Baseline Knowledge**: `{baseline_knowledge}` (CRITICALITY: High)
-   **New Information**: `{new_information}` (FRESHNESS: Assessed as recent)

**Output Requirements:**
1.  **Primary Format**: A single, schema-compliant JSON object.
2.  **Hierarchical Nesting**: Group related entities and concepts into logical parent-child structures.
3.  **Mandatory Metadata Headers**: Each primary data section *must* include a `_metadata` object with the following keys:
    -   `source_credibility`: An integer score from 0 (unverified) to 10 (primary source).
    -   `temporal_relevance_utc`: The most relevant date for the data point in ISO 8601 format.
    -   `confidence_score`: A float from 0.0 to 1.0 indicating your certainty in the data's accuracy.
4.  **Data Efficiency**: Retain all mission-critical data points. Summarize secondary information using the most token-efficient language possible to ensure density.

**Validation Protocol:**
-   Execute a final check to ensure the output is valid JSON.
-   Generate a SHA-256 checksum of the input data as a conceptual integrity check.
"""

    NARRATIVE_PROSE_REPORT = """
Protocol: **Comprehensive Narrative Synthesis v1.5**
Objective: {task_objective}

**Input Data Streams:**
-   **Collected Knowledge Base**: `{knowledge_base}`

**Output Requirements:**
1.  **Format**: A detailed, long-form narrative report (target ~8000 words).
2.  **Structure**: The report must be organized into the following sections:
    a.  **Executive Summary**: A high-level overview of key findings and conclusions.
    b.  **Introduction**: State the report's purpose and scope.
    c.  **Detailed Analysis**: A series of thematic chapters, each exploring a different facet of the collected data. Use Markdown for headings, lists, and bolding to improve readability.
    d.  **Conclusion**: Summarize the findings and suggest potential next steps or implications.
    e.  **Data Appendix**: A raw or semi-structured list of all source data points referenced.
3.  **Tone**: Professional, thorough, and exhaustive. Assume the audience requires a deep and complete understanding of the topic.
"""

    # --- TASK & PROGRESS MANAGEMENT PROMPTS ---
    PROJECT_STATUS_REPORT = """
Protocol: **Progress Compression & Milestone Review v1.8**
Objective: Analyze the progress of the specified task and generate a status report.
Task Under Review: {task_description}

**Analysis Directives:**
1.  **Phase Identification**: Determine the current phase of the task (e.g., Research, Analysis, Synthesis, Review).
2.  **Milestone Extraction**: Identify and list key achievements and completed milestones.
3.  **Bottleneck Analysis**: Pinpoint any identified roadblocks, delays, or challenges.

**Output Requirements:**
-   **Timeline Visualization (Text-based Gantt Chart)**:
    Example:
    [Phase 1: Research]   ▓▓▓▓▓▓▓▓▓▓▓▓░░░░░░░░ (60% Complete)
    [Phase 2: Analysis]   ▓▓▓░░░░░░░░░░░░░░░░ (15% Complete)
    [Phase 3: Synthesis]  ░░░░░░░░░░░░░░░░░░░░ (0% Complete)

-   **Resource Allocation Map**: A summary of resources assigned or utilized.
-   **Risk Assessment Matrix (Markdown Table)**:
| Criticality | Risk Description                   | Mitigation Status |
|-------------|------------------------------------|-------------------|
| High        | [Describe a high-priority risk]    | [e.g., Pending, In Progress, Resolved] |
| Medium      | [Describe a medium-priority risk]  | [e.g., Pending, In Progress, Resolved] |
| Low         | [Describe a low-priority risk]     | [e.g., Pending, In Progress, Resolved] |
"""

# === SYSTEM AUDITING & LOGGING UTILITIES ===

class SystemAuditor:
    """
    A utility class to handle the formatting of system-level logs for auditing and debugging.
    """
    def __init__(self, session_id: Optional[str] = None):
        self.session_id = session_id or str(uuid.uuid4())

    def _get_system_metrics(self) -> Dict[str, Any]:
        """Retrieves CPU and memory usage if psutil is installed."""
        if psutil:
            return {
                "cpu_load": psutil.cpu_percent(),
                "mem_use_gb": round(psutil.virtual_memory().used / (1024**3), 2),
            }
        return {"cpu_load": "N/A", "mem_use_gb": "N/A"}

    def format_prompt_log(self, content: str, user_profile: str = "default_user") -> str:
        """Formats a log entry for a sent prompt."""
        metrics = self._get_system_metrics()
        return f"""
〚PROMPT LOG v3.2〛
SessionID: {self.session_id}
├─ Timestamp: {datetime.datetime.now(datetime.timezone.utc).isoformat()}
├─ User Context: {user_profile}
└─ System State:
   CPU: {metrics['cpu_load']}% | Mem: {metrics['mem_use_gb']}GB

{LOG_DIVIDER}
{content.strip()}
{LOG_DIVIDER}
"""

    def format_response_log(self, content: str, latency_ms: float, source_count: int, confidence: float) -> str:
        """Formats an audit trail for a received response."""
        ethical_status = "PASS"  # This would be determined by a separate process
        return f"""
〚RESPONSE AUDIT TRAIL〛
├─ Processing Time: {latency_ms:.2f}ms
├─ Data Sources Referenced: {source_count}
├─ Ethical Check: {ethical_status}
└─ Confidence Metric: {confidence:.2f}

{LOG_DIVIDER}
{content.strip()}
{LOG_DIVIDER}

--- RESPONSE PAYLOAD ---
{content.strip()}
--- END PAYLOAD ---
"""

# === MAIN EXECUTION BLOCK (Demonstration) ===
if __name__ == "__main__":
    print("Demonstrating the Maestro Prompt Library and System Auditor.\n")

    # 1. Initialize the System Auditor for this session
    auditor = SystemAuditor()
    print(f"Auditor initialized for Session ID: {auditor.session_id}\n")

    # 2. DEMO: Generate a Narrative Prose Report
    print(f"{LOG_DIVIDER}\nDEMO 1: Generating a Narrative Prose Report\n{LOG_DIVIDER}")

    # Prepare the data for the prompt placeholders
    narrative_data = {
        "task_objective": "Synthesize findings on the impact of quantum computing on modern cryptography.",
        "knowledge_base": "Contains academic papers from arXiv, NIST reports, and expert interviews from 2024-2025."
    }

    # Format the prompt
    narrative_prompt = PromptLibrary.NARRATIVE_PROSE_REPORT.format(**narrative_data)

    # Log the formatted prompt using the auditor
    logged_prompt = auditor.format_prompt_log(narrative_prompt, user_profile="crypto_researcher_01")
    print("--- Logged Prompt to be Sent to LLM ---")
    print(logged_prompt)

    # --- (Imagine an LLM processes this prompt and returns a response) ---
    simulated_llm_response = "Executive Summary: Quantum computing poses a significant, near-term threat..."
    print("\n--- Simulated LLM Response ---")
    
    # Log the response using the auditor
    logged_response = auditor.format_response_log(
        content=simulated_llm_response,
        latency_ms=4820.5,
        source_count=12,
        confidence=0.92
    )
    print(logged_response)


    # 3. DEMO: Generate a Project Status Report
    print(f"\n{LOG_DIVIDER}\nDEMO 2: Generating a Project Status Report\n{LOG_DIVIDER}")

    status_data = {
        "task_description": "Q3-2025 Market Analysis for AI-driven agricultural sensors."
    }
    status_prompt = PromptLibrary.PROJECT_STATUS_REPORT.format(**status_data)

    logged_status_prompt = auditor.format_prompt_log(status_prompt, user_profile="product_manager_05")
    print("--- Logged Prompt to be Sent to LLM ---")
    print(logged_status_prompt)