xiaoxia-saas/tests/unit/test_subtitle_domain.py

"""
Subtitle 字幕领域模型单元测试
"""

import pytest

from packages.domain.subtitle import (
    SubtitleSegment,
    SubtitleTimeline,
    SubtitleWord,
)


class TestSubtitleWord:
    """SubtitleWord 测试"""

    def test_duration_positive(self):
        word = SubtitleWord(text="你好", start=1.0, end=2.5)
        assert word.duration == pytest.approx(1.5)

    def test_duration_zero(self):
        word = SubtitleWord(text="a", start=5.0, end=5.0)
        assert word.duration == 0.0

    def test_duration_negative_returns_zero(self):
        """测试结束时间小于开始时间时返回 0"""
        word = SubtitleWord(text="a", start=3.0, end=1.0)
        assert word.duration == 0.0


class TestSubtitleSegment:
    """SubtitleSegment 测试"""

    def test_duration(self):
        seg = SubtitleSegment(text="你好世界", start=0.0, end=3.0)
        assert seg.duration == pytest.approx(3.0)

    def test_duration_zero(self):
        seg = SubtitleSegment(text="test", start=5.0, end=5.0)
        assert seg.duration == 0.0

    def test_duration_negative_returns_zero(self):
        seg = SubtitleSegment(text="test", start=5.0, end=2.0)
        assert seg.duration == 0.0

    def test_char_count(self):
        seg = SubtitleSegment(text="你好世界", start=0, end=1)
        assert seg.char_count == 4

    def test_char_count_empty(self):
        seg = SubtitleSegment(text="", start=0, end=1)
        assert seg.char_count == 0

    def test_default_words_empty(self):
        seg = SubtitleSegment(text="test", start=0, end=1)
        assert seg.words == []

    def test_with_words(self):
        words = [
            SubtitleWord(text="你好", start=0.0, end=1.0),
            SubtitleWord(text="世界", start=1.0, end=2.0),
        ]
        seg = SubtitleSegment(text="你好世界", start=0.0, end=2.0, words=words)
        assert len(seg.words) == 2
        assert seg.words[0].text == "你好"
        assert seg.words[1].text == "世界"


class TestSubtitleTimelineBasics:
    """SubtitleTimeline 基础属性测试"""

    def test_empty_timeline(self):
        tl = SubtitleTimeline()
        assert tl.segment_count == 0
        assert tl.total_chars == 0
        assert tl.language == "zh"
        assert tl.total_duration == 0.0

    def test_segment_count(self):
        tl = SubtitleTimeline(segments=[
            SubtitleSegment(text="a", start=0, end=1),
            SubtitleSegment(text="b", start=1, end=2),
            SubtitleSegment(text="c", start=2, end=3),
        ])
        assert tl.segment_count == 3

    def test_total_chars(self):
        tl = SubtitleTimeline(segments=[
            SubtitleSegment(text="你好", start=0, end=1),
            SubtitleSegment(text="世界", start=1, end=2),
            SubtitleSegment(text="abcde", start=2, end=3),
        ])
        assert tl.total_chars == 9

    def test_custom_language(self):
        tl = SubtitleTimeline(language="en")
        assert tl.language == "en"

    def test_custom_total_duration(self):
        tl = SubtitleTimeline(total_duration=60.0)
        assert tl.total_duration == 60.0


class TestMergeShortSegments:
    """merge_short_segments 测试"""

    def test_single_segment_no_merge(self):
        """单个片段不需要合并"""
        tl = SubtitleTimeline(segments=[
            SubtitleSegment(text="a", start=0, end=1),
        ])
        result = tl.merge_short_segments(min_chars=8)
        assert result.segment_count == 1
        assert result.segments[0].text == "a"

    def test_empty_timeline(self):
        """空时间轴"""
        tl = SubtitleTimeline()
        result = tl.merge_short_segments(min_chars=8)
        assert result.segment_count == 0

    def test_all_short_segments_merge_into_one(self):
        """所有短片段合并成一个"""
        tl = SubtitleTimeline(segments=[
            SubtitleSegment(text="你", start=0, end=0.5),
            SubtitleSegment(text="好", start=0.5, end=1.0),
            SubtitleSegment(text="世", start=1.0, end=1.5),
            SubtitleSegment(text="界", start=1.5, end=2.0),
        ])
        result = tl.merge_short_segments(min_chars=8)
        assert result.segment_count == 1
        assert result.segments[0].text == "你好世界"
        assert result.segments[0].start == 0
        assert result.segments[0].end == 2.0

    def test_merge_short_segments_preserves_timing(self):
        """合并后时间轴正确"""
        tl = SubtitleTimeline(segments=[
            SubtitleSegment(text="你好", start=1.0, end=2.0),
            SubtitleSegment(text="世界", start=2.0, end=3.5),
        ])
        result = tl.merge_short_segments(min_chars=10)
        assert result.segment_count == 1
        assert result.segments[0].start == 1.0
        assert result.segments[0].end == 3.5

    def test_merge_short_segments_with_words(self):
        """合并后词级信息保留"""
        w1 = SubtitleWord(text="你好", start=0.0, end=1.0)
        w2 = SubtitleWord(text="世界", start=1.0, end=2.0)
        tl = SubtitleTimeline(segments=[
            SubtitleSegment(text="你好", start=0.0, end=1.0, words=[w1]),
            SubtitleSegment(text="世界", start=1.0, end=2.0, words=[w2]),
        ])
        result = tl.merge_short_segments(min_chars=10)
        assert len(result.segments[0].words) == 2
        assert result.segments[0].words[0].text == "你好"
        assert result.segments[0].words[1].text == "世界"

    def test_multiple_merged_groups(self):
        """多个合并组 — 短段会和后续段累积到够数才提交"""
        tl = SubtitleTimeline(segments=[
            SubtitleSegment(text="一二三四五六七八", start=0, end=2),  # 8字，够数，提交
            SubtitleSegment(text="九", start=2, end=2.5),  # 1字，入buffer
            SubtitleSegment(text="十", start=2.5, end=3),  # 1字，入buffer（共2字）
            SubtitleSegment(text="一二三四五六七八九十", start=3, end=5),  # 10字，入buffer后共12字，够数提交
        ])
        result = tl.merge_short_segments(min_chars=8)
        # 第1段："一二三四五六七八"（8字直接提交）
        # 第2段："九十" + "一二三四五六七八九十" 累积到12字一起提交
        assert result.segment_count == 2
        assert result.segments[0].text == "一二三四五六七八"
        assert result.segments[1].text == "九十一二三四五六七八九十"

    def test_remaining_short_merged_with_last(self):
        """剩余短片段合并到最后一段"""
        tl = SubtitleTimeline(segments=[
            SubtitleSegment(text="一二三四五六七八", start=0, end=2),  # 8字
            SubtitleSegment(text="一二三", start=2, end=3),  # 3字，不够
        ])
        result = tl.merge_short_segments(min_chars=8)
        # 最后的3字会合并到上一段（因为 < min_chars）
        assert result.segment_count == 1
        assert result.segments[0].text == "一二三四五六七八一二三"

    def test_custom_min_chars(self):
        """自定义最小字数 — 累积到够数就提交，剩余短的合并到最后"""
        tl = SubtitleTimeline(segments=[
            SubtitleSegment(text="一二", start=0, end=1),
            SubtitleSegment(text="三四", start=1, end=2),
            SubtitleSegment(text="五六", start=2, end=3),
        ])
        # min_chars=3：
        # "一二"(2字) → 不够
        # +"三四"(共4字) → 够了，提交"一二三四"，buffer清空
        # "五六"(2字) → 循环结束，剩余<min_chars且merged非空 → 合并到最后一段
        # 结果：1段 "一二三四五六"
        result = tl.merge_short_segments(min_chars=3)
        assert result.segment_count == 1
        assert result.segments[0].text == "一二三四五六"

    def test_preserves_language_and_duration(self):
        """合并后保留语言和总时长"""
        tl = SubtitleTimeline(
            segments=[SubtitleSegment(text="a", start=0, end=1)],
            language="en",
            total_duration=60.0,
        )
        result = tl.merge_short_segments(min_chars=8)
        assert result.language == "en"
        assert result.total_duration == 60.0

    def test_does_not_modify_original(self):
        """不修改原时间轴"""
        segments = [
            SubtitleSegment(text="a", start=0, end=1),
            SubtitleSegment(text="b", start=1, end=2),
        ]
        tl = SubtitleTimeline(segments=segments)
        result = tl.merge_short_segments(min_chars=5)
        # 原时间轴不变
        assert tl.segment_count == 2
        assert result is not tl


class TestSplitLongSegments:
    """split_long_segments 测试"""

    def test_short_segments_no_split(self):
        """短片段不需要拆分"""
        tl = SubtitleTimeline(segments=[
            SubtitleSegment(text="你好", start=0, end=1),
        ])
        result = tl.split_long_segments(max_chars=20)
        assert result.segment_count == 1
        assert result.segments[0].text == "你好"

    def test_single_long_segment_split_by_punctuation(self):
        """长片段按标点拆分"""
        text = "你好世界。今天天气真好，我们出去玩吧！"
        tl = SubtitleTimeline(segments=[
            SubtitleSegment(text=text, start=0, end=10.0),
        ])
        result = tl.split_long_segments(max_chars=10)
        # 应该被拆成多段
        assert result.segment_count > 1
        # 每段都不超过 max_chars（除了硬切的情况）
        for seg in result.segments:
            assert seg.char_count <= len(text)  # 至少比原文短

    def test_split_preserves_total_text(self):
        """拆分后总文本不变"""
        text = "你好世界。今天天气真好，我们出去玩吧！明天再见。"
        tl = SubtitleTimeline(segments=[
            SubtitleSegment(text=text, start=0, end=10.0),
        ])
        result = tl.split_long_segments(max_chars=8)
        merged_text = "".join(s.text for s in result.segments)
        assert merged_text == text

    def test_split_time_proportional(self):
        """拆分后时间按字数比例分配"""
        text = "一二三四五六七八九十。"  # 11字
        tl = SubtitleTimeline(segments=[
            SubtitleSegment(text=text, start=0, end=10.0),
        ])
        result = tl.split_long_segments(max_chars=5)
        # 总时长不变
        assert result.segments[0].start == 0.0
        assert result.segments[-1].end == pytest.approx(10.0)
        # 各段首尾相接
        for i in range(len(result.segments) - 1):
            assert result.segments[i].end == pytest.approx(result.segments[i + 1].start)

    def test_split_with_words(self):
        """拆分时词级信息正确分配"""
        words = [
            SubtitleWord(text="你好", start=0.0, end=1.0),
            SubtitleWord(text="世界", start=1.0, end=2.0),
            SubtitleWord(text="你好吗", start=2.0, end=3.5),
        ]
        tl = SubtitleTimeline(segments=[
            SubtitleSegment(text="你好世界。你好吗？", start=0.0, end=3.5, words=words),
        ])
        result = tl.split_long_segments(max_chars=4)
        # 第一段应该有前几个词
        assert len(result.segments) >= 2
        total_words = sum(len(s.words) for s in result.segments)
        assert total_words == 3  # 词的总数不变

    def test_multiple_mixed_segments(self):
        """混合长短片段"""
        tl = SubtitleTimeline(segments=[
            SubtitleSegment(text="短", start=0, end=1),  # 短
            SubtitleSegment(text="一二三四五六七八九十一二三四五六七八九十", start=1, end=5),  # 长
            SubtitleSegment(text="也短", start=5, end=6),  # 短
        ])
        result = tl.split_long_segments(max_chars=10)
        assert result.segment_count >= 3  # 至少3段（中间被拆成多段）
        # 第一段还是原来的短的
        assert result.segments[0].text == "短"
        # 最后一段还是原来的短的
        assert result.segments[-1].text == "也短"

    def test_no_punctuation_hard_split(self):
        """没有标点时硬切"""
        text = "一二三四五六七八九十一二三四五六七八九十一二三四五"
        tl = SubtitleTimeline(segments=[
            SubtitleSegment(text=text, start=0, end=10.0),
        ])
        result = tl.split_long_segments(max_chars=10)
        assert result.segment_count >= 3
        for seg in result.segments:
            # 硬切的每段应该 <= max_chars
            assert seg.char_count <= 10

    def test_preserves_language_and_duration(self):
        """拆分后保留语言和总时长"""
        tl = SubtitleTimeline(
            segments=[SubtitleSegment(text="a", start=0, end=1)],
            language="ja",
            total_duration=30.0,
        )
        result = tl.split_long_segments(max_chars=20)
        assert result.language == "ja"
        assert result.total_duration == 30.0

    def test_does_not_modify_original(self):
        """不修改原时间轴"""
        original_text = "一二三四五六七八九十一二三四五六七八九十"
        tl = SubtitleTimeline(segments=[
            SubtitleSegment(text=original_text, start=0, end=5),
        ])
        result = tl.split_long_segments(max_chars=8)
        assert tl.segment_count == 1
        assert tl.segments[0].text == original_text
        assert result is not tl


class TestSplitTextByPunctuation:
    """_split_text_by_punctuation 静态方法测试"""

    def test_short_text_no_split(self):
        result = SubtitleTimeline._split_text_by_punctuation("你好世界", 10)
        assert result == ["你好世界"]

    def test_split_at_sentence_end(self):
        """在句末标点处断开"""
        result = SubtitleTimeline._split_text_by_punctuation("你好。世界。", 5)
        assert len(result) == 2
        assert result[0] == "你好。"
        assert result[1] == "世界。"

    def test_split_at_comma(self):
        """在逗号处断开（超过最大长度时）"""
        text = "一二三四五六七八，二二三四五六七八。"
        result = SubtitleTimeline._split_text_by_punctuation(text, 10)
        assert len(result) >= 2

    def test_no_punctuation_hard_split(self):
        """没有标点时硬切"""
        result = SubtitleTimeline._split_text_by_punctuation("一二三四五六七八九十", 5)
        assert len(result) == 2
        assert result[0] == "一二三四五"
        assert result[1] == "六七八九十"

    def test_empty_text(self):
        # 空字符串循环不执行，current为空不append，返回空列表
        result = SubtitleTimeline._split_text_by_punctuation("", 10)
        assert result == []

    def test_mixed_punctuation(self):
        """混合标点"""
        text = "你好！吃饭了吗？是的，我吃过了。"
        result = SubtitleTimeline._split_text_by_punctuation(text, 6)
        # 验证所有段加起来等于原文
        assert "".join(result) == text

    def test_sentence_end_with_min_length(self):
        """句末标点断句的「半长门槛」只在未超max_chars时生效；
        超过max_chars回溯找标点时，即使首段很短也会断开。"""
        # "你好。" 3字 < max_chars//2(5)，未超max_chars时不会主动断开
        # 但加上后面的"世界很大很美好"后超过10字，回溯找标点找到"。"，强制断开
        text = "你好。世界很大很美好。"
        result = SubtitleTimeline._split_text_by_punctuation(text, 10)
        # 超过max_chars时回溯断开，首段可能很短
        assert len(result) == 2
        assert result[0] == "你好。"
        assert result[1] == "世界很大很美好。"
        # 总文本不变
        assert "".join(result) == text

    def test_exclamation_and_question_marks(self):
        """感叹号和问号也算句末标点"""
        text = "你好吗！我很好！你呢？"
        result = SubtitleTimeline._split_text_by_punctuation(text, 4)
        assert len(result) >= 3


class TestMergeSegments:
    """_merge_segments 静态方法测试"""

    def test_merge_two_segments(self):
        result = SubtitleTimeline._merge_segments([
            SubtitleSegment(text="你好", start=0.0, end=1.0),
            SubtitleSegment(text="世界", start=1.0, end=2.0),
        ])
        assert result.text == "你好世界"
        assert result.start == 0.0
        assert result.end == 2.0

    def test_merge_empty_list(self):
        result = SubtitleTimeline._merge_segments([])
        assert result.text == ""
        assert result.start == 0
        assert result.end == 0

    def test_merge_single_segment(self):
        seg = SubtitleSegment(text="test", start=1.0, end=2.0)
        result = SubtitleTimeline._merge_segments([seg])
        assert result.text == "test"
        assert result.start == 1.0
        assert result.end == 2.0

    def test_merge_preserves_words(self):
        w1 = SubtitleWord(text="你好", start=0.0, end=1.0)
        w2 = SubtitleWord(text="世界", start=1.0, end=2.0)
        result = SubtitleTimeline._merge_segments([
            SubtitleSegment(text="你好", start=0.0, end=1.0, words=[w1]),
            SubtitleSegment(text="世界", start=1.0, end=2.0, words=[w2]),
        ])
        assert len(result.words) == 2
        assert result.words[0].text == "你好"
        assert result.words[1].text == "世界"

    def test_merge_non_contiguous_segments(self):
        """合并非连续片段（有间隙）"""
        result = SubtitleTimeline._merge_segments([
            SubtitleSegment(text="a", start=0.0, end=1.0),
            SubtitleSegment(text="b", start=3.0, end=4.0),
        ])
        assert result.start == 0.0
        assert result.end == 4.0
        assert result.text == "ab"


class TestMergeAndSplitRoundtrip:
    """合并和拆分的组合测试"""

    def test_split_then_merge_approximate(self):
        """拆分后再合并，总字数和总时长基本一致"""
        original_text = "你好世界。今天天气真好，我们出去玩吧！明天见。"
        tl = SubtitleTimeline(segments=[
            SubtitleSegment(text=original_text, start=0.0, end=10.0),
        ])
        split = tl.split_long_segments(max_chars=5)
        merged = split.merge_short_segments(min_chars=50)  # 足够大的min_chars让它们都合并
        assert merged.segment_count == 1
        assert merged.segments[0].text == original_text
        assert merged.segments[0].start == 0.0
        assert merged.segments[0].end == pytest.approx(10.0)