"""
Condition 2: Expert-Only (No Attributes)

Uses expert perspectives to generate ideas, but without
attribute decomposition. Each expert generates ideas directly
for the query from their professional perspective.
"""

import sys
from pathlib import Path

# Add backend to path for imports
sys.path.insert(0, str(Path(__file__).parent.parent.parent / "backend"))

from typing import List, Dict, Any
from app.services.llm_service import ollama_provider, extract_json_from_response
from app.services.expert_source_service import expert_source_service
from experiments.config import (
    MODEL, TEMPERATURE, EXPERT_COUNT, EXPERT_SOURCE,
    IDEAS_PER_EXPERT, PROMPT_LANGUAGE
)


def get_expert_idea_generation_prompt(
    query: str,
    expert_name: str,
    expert_domain: str,
    idea_count: int,
    lang: str = "en"
) -> str:
    """Generate prompt for expert-based idea generation."""
    if lang == "en":
        domain_text = f" ({expert_domain} field)" if expert_domain else ""
        return f"""/no_think
You are a {expert_name}{domain_text}.

Task: Generate {idea_count} creative and innovative ideas for "{query}" from your professional perspective.

Requirements:
1. Each idea should reflect your professional expertise and unique viewpoint
2. Think about how concepts from your field could improve or reimagine "{query}"
3. Ideas should be specific and actionable (15-30 words each)
4. Combine your professional knowledge with creative thinking

Return JSON only:
{{"ideas": ["idea 1", "idea 2", "idea 3", ...]}}

Generate exactly {idea_count} ideas from your perspective as a {expert_name}."""
    else:
        domain_text = f"（{expert_domain}領域）" if expert_domain else ""
        return f"""/no_think
你是一位{expert_name}{domain_text}。

任務：從你的專業角度，為「{query}」生成 {idea_count} 個創意點子。

要求：
1. 每個點子要反映你的專業知識和獨特觀點
2. 思考你領域的概念如何改進或重新想像「{query}」
3. 點子要具體可行（每個 15-30 字）
4. 結合專業知識和創意思維

只回傳 JSON：
{{"ideas": ["點子1", "點子2", "點子3", ...]}}

從你作為{expert_name}的角度生成正好 {idea_count} 個點子。"""


async def generate_ideas(
    query: str,
    model: str = None,
    temperature: float = None,
    expert_count: int = None,
    expert_source: str = None,
    ideas_per_expert: int = None,
    lang: str = None
) -> Dict[str, Any]:
    """
    Generate ideas using expert perspectives only (C2).

    Args:
        query: The object/concept to generate ideas for
        model: LLM model to use
        temperature: Generation temperature
        expert_count: Number of experts to use
        expert_source: Source of experts (curated, dbpedia, etc.)
        ideas_per_expert: Ideas each expert generates
        lang: Language for prompts

    Returns:
        Dict with ideas and metadata
    """
    model = model or MODEL
    temperature = temperature or TEMPERATURE
    expert_count = expert_count or EXPERT_COUNT
    expert_source = expert_source or EXPERT_SOURCE
    ideas_per_expert = ideas_per_expert or IDEAS_PER_EXPERT
    lang = lang or PROMPT_LANGUAGE

    # Get experts from curated source
    experts, actual_source = expert_source_service.get_experts(
        source=expert_source,
        count=expert_count,
        language=lang
    )

    all_ideas = []
    expert_details = []

    for expert in experts:
        expert_name = expert.get("name", "Expert")
        expert_domain = expert.get("domain", "")

        prompt = get_expert_idea_generation_prompt(
            query=query,
            expert_name=expert_name,
            expert_domain=expert_domain,
            idea_count=ideas_per_expert,
            lang=lang
        )

        response = await ollama_provider.generate(
            prompt=prompt,
            model=model,
            temperature=temperature
        )

        result = extract_json_from_response(response)
        ideas = result.get("ideas", [])

        # Tag ideas with expert source
        for idea in ideas:
            all_ideas.append({
                "idea": idea,
                "expert_name": expert_name,
                "expert_domain": expert_domain
            })

        expert_details.append({
            "name": expert_name,
            "domain": expert_domain,
            "ideas_generated": len(ideas)
        })

    return {
        "condition": "c2_expert_only",
        "query": query,
        "ideas": [item["idea"] for item in all_ideas],
        "ideas_with_source": all_ideas,
        "idea_count": len(all_ideas),
        "metadata": {
            "model": model,
            "temperature": temperature,
            "prompt_language": lang,
            "expert_count": expert_count,
            "expert_source": actual_source,
            "ideas_per_expert": ideas_per_expert,
            "experts": expert_details,
            "mechanism": "expert_perspectives_only"
        }
    }


# For testing
if __name__ == "__main__":
    import asyncio

    async def test():
        result = await generate_ideas("Chair")
        print(f"Generated {result['idea_count']} ideas from {len(result['metadata']['experts'])} experts:")
        for exp in result['metadata']['experts']:
            print(f"  - {exp['name']}: {exp['ideas_generated']} ideas")
        print("\nSample ideas:")
        for i, item in enumerate(result['ideas_with_source'][:5], 1):
            print(f"  {i}. [{item['expert_name']}] {item['idea']}")

    asyncio.run(test())