import os
from src.clients.llm_client import LLMClient
import json 
import pandas as pd
from pydantic import BaseModel, Field
from enum import Enum
import base64
from io import BytesIO
from PIL import Image
from typing import List, Optional
from functools import cache
from datetime import datetime
import pytz
from src.utils.tracer import customtracer
from src.models.common import model

def _ask_raw_hf(messages, model, response_format=None):
    """Compatibility wrapper: routes OpenAI-style messages through HF LLMClient."""
    from src.clients.llm_client import LLMClient
    import json as _json

    client = LLMClient()
    system_prompt = None
    user_text = ""
    images = []
    for msg in messages:
        role = msg.get("role", "")
        c = msg.get("content", "")
        if role == "system":
            if isinstance(c, str):
                system_prompt = c
        elif role == "user":
            if isinstance(c, str):
                user_text = c
            elif isinstance(c, list):
                for part in c:
                    if isinstance(part, dict):
                        if part.get("type") == "text":
                            user_text += part.get("text", "")
                        elif part.get("type") == "image_url":
                            url = part.get("image_url", {}).get("url", "")
                            if url.startswith("data:"):
                                images.append(url.split(",", 1)[1] if "," in url else url)
                            else:
                                images.append(url)

    if response_format is not None and hasattr(response_format, "model_json_schema"):
        result = client.call(
            prompt=user_text,
            schema=response_format,
            model=model,
            system_prompt=system_prompt,
            images=images if images else None,
            temperature=0,
        )
        return _json.dumps(result.model_dump(), ensure_ascii=False)
    else:
        return client.call_raw(
            prompt=user_text,
            model=model,
            system_prompt=system_prompt,
            images=images if images else None,
        )


# スキーマ定義�E�Eormat2fvinfo.pyから独立！E
class Meta(BaseModel):
    会社吁E str
    業畁E str
    ブランチE str
    サービス: str
    啁E��: str
    タイトル: str
    構�Eの意図: str
    訴求テーチE list[str]

class Font(str, Enum):
    font1 = "ゴシチE��"
    font2 = "明朝"
    font3 = "手書ぁE

class EvsF(str, Enum):
    EMOTIONAL = "惁E��E
    FUNCTIONAL = "機�E"

class PvsS(str, Enum):
    PROBLEM = "問題提起"
    SOLUTION = "課題解決"

class Copy(BaseModel):
    text: str
    font: Font
    color: str
    visual: str
    appeal_mode : EvsF
    forcus_stage : PvsS

class CatchCopy(BaseModel):
    main_copy: list[Copy]
    sub_copy: list[Copy]

class FVinfo(BaseModel):
    非LP: bool
    メタ: Meta
    キャチE��コピ�E: CatchCopy
    権威付け: list[str]
    ビジュアル: list[str]
    CTAボタン: list[str]

def get_openai_request(messages, format, n=1):
    """
    OpenAI API呼び出し！Eパラメータ対応！E
    
    Args:
        messages: メチE��ージリスチE
        format: レスポンスフォーマッチE
        n: 生�Eする候補数�E�デフォルチE 1�E�E
    
    Returns:
        list[str]: 常にリストで返却�E�E=1でも長ぁEのリスト！E
    """
    client = LLMClient()
    response = _ask_raw_hf([{"role":"user","content":p}], model,
        model="meta-llama/Llama-3.3-70B-Instruct",
        messages=messages,
        top_p=1,
        frequency_penalty=0,
        presence_penalty=0,
        response_format=format,
        temperature=1.2,
        n=n
    )
    
    # 常にリストで返す�E�E=1でも統一�E�E
    return [choice.message.content for choice in response.choices]

@customtracer
def format2fvinfos(p, openai_key=os.environ.get('OPENAI_KEY'), n=1):
    """
    褁E��バリアントを返すformat2fvinfo�E��E列版�E�E
    
    input1 (text): prompt text
    input2 (text): default
    input3 (number): 1
    output1 (json): variants list
    """
    print(datetime.now(pytz.timezone('Asia/Tokyo')).strftime("%Y-%m-%d %H:%M:%S"), __name__, f"n={n}")
    
    if openai_key == "default":
        os.environ['OPENAI_API_KEY'] = os.environ.get('OPENAI_KEY')
    else:
        os.environ['OPENAI_API_KEY'] = openai_key
    
    # n を整数に変換し、篁E��チェチE��
    try:
        n = int(n)
        if n < 1:
            print(f"Warning: n={n} is invalid, using n=1")
            n = 1
        elif n > 10:
            print(f"Warning: n={n} is too large, capping at 10")
            n = 10
    except (TypeError, ValueError):
        print(f"Warning: n={n} is invalid, using n=1")
        n = 1
    
    messages=[
        {
        "role": "system",
        "content": """提供したフォーマットデータから、忁E��なファーストビューの要素を生成してください、E"",
        },
        {
        "role": "user",
        "content":  [{"type": "text", "text":p}]
        },
    ]
    
    # get_openai_requestは常にリストを返すので、そのまま使用
    result = get_openai_request(messages, FVinfo, n=n)
    
    print(f"Generated {len(result)} variants")
    
    # リストをJSON斁E���Eとして返す
    return json.dumps(result, ensure_ascii=False)