"""Mock ASR 服务 — 用于测试和开发环境。

生成模拟的字幕时间轴，不依赖真实ASR服务。
"""

from __future__ import annotations

import re
from pathlib import Path
from typing import Optional

from packages.domain.subtitle import (
    SubtitleSegment,
    SubtitleTimeline,
    SubtitleWord,
)
from packages.ports.asr_service import ASRService, ASRServiceError


class MockASRService(ASRService):
    """Mock ASR 服务，生成模拟字幕数据。

    如果 audio_path 对应的目录下有同名 .txt 文件，
    就读取该文件内容作为字幕文本，按时间均匀分段。
    否则生成默认的测试字幕。
    """

    def __init__(self, mock_text: Optional[str] = None):
        self._mock_text = mock_text

    def transcribe(
        self,
        audio_path: Path,
        language: Optional[str] = None,
        with_word_timestamps: bool = True,
    ) -> SubtitleTimeline:
        if not audio_path.exists():
            raise ASRServiceError(f"音频文件不存在: {audio_path}", provider="mock")

        # 尝试读取同名 txt 文件作为字幕文本
        text = self._mock_text
        if text is None:
            txt_path = audio_path.with_suffix(".txt")
            if txt_path.exists():
                text = txt_path.read_text(encoding="utf-8").strip()
            else:
                text = "这是一段测试字幕。它用于验证ASR自动字幕功能是否正常工作。每一句话都会被正确地分段并显示在视频底部。字幕的样式可以根据用户的喜好进行自定义调整。"

        # 估算音频时长（用ffmpeg probe或者直接假设）
        # mock模式下按字数估算，每秒4个字
        total_duration = max(5.0, len(text) / 4.0)

        segments = self._text_to_segments(text, total_duration, with_word_timestamps)

        return SubtitleTimeline(
            segments=segments,
            language=language or "zh",
            total_duration=total_duration,
        )

    def _text_to_segments(
        self,
        text: str,
        total_duration: float,
        with_word_timestamps: bool,
    ) -> list[SubtitleSegment]:
        """将文本按句切分成带时间轴的字幕片段。"""
        # 按句末标点拆分
        sentences = re.split(r"(?<=[。！？!?])", text)
        sentences = [s.strip() for s in sentences if s.strip()]

        if not sentences:
            sentences = [text]

        total_chars = sum(len(s) for s in sentences)
        if total_chars == 0:
            return []

        segments = []
        current_time = 0.0

        for sentence in sentences:
            char_count = len(sentence)
            duration = total_duration * (char_count / total_chars)
            end_time = current_time + duration

            words: list[SubtitleWord] = []
            if with_word_timestamps:
                # 每个字作为一个词级单元（中文按字，英文按词）
                word_time = current_time
                word_duration = duration / char_count

                for char in sentence:
                    words.append(
                        SubtitleWord(
                            text=char,
                            start=word_time,
                            end=word_time + word_duration,
                        )
                    )
                    word_time += word_duration

            segments.append(
                SubtitleSegment(
                    text=sentence,
                    start=current_time,
                    end=end_time,
                    words=words,
                )
            )
            current_time = end_time

        return segments