From fc76ded3e4b5ed8ba68a3c12e29d6ceeb5608595 Mon Sep 17 00:00:00 2001
From: hmo <hmo@yoin.fun>
Date: Mon, 4 May 2026 00:22:55 +0800
Subject: [PATCH] feat: add Pipeline.add_clip_by_title

---
 src/core/pipeline.py | 69 ++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 69 insertions(+)

diff --git a/src/core/pipeline.py b/src/core/pipeline.py
index c5328f2..4291599 100644
--- a/src/core/pipeline.py
+++ b/src/core/pipeline.py
@@ -584,6 +584,75 @@ class Pipeline:
         # 保存更新后的 config
         self._save_config()
 
+    def add_clip_by_title(self, new_title):
+        """
+        用新标题在转录中匹配时间段，判断合并或新增。
+
+        Args:
+            new_title: 新知识点标题
+
+        Returns:
+            tuple: (clip_index, matched) — 新 clip 在 self.clips 中的索引，matched 是否匹配成功
+        """
+        # 1. 加载 corrected_transcript.json
+        transcript_path = os.path.join(self.inter_dir, 'corrected_transcript.json')
+        if not os.path.exists(transcript_path):
+            logger.warning(f"corrected_transcript.json not found")
+            return None, False
+
+        with open(transcript_path, 'r', encoding='utf-8') as f:
+            corrected_segments = json.load(f)
+
+        # 2. 匹配标题
+        from .ppt_parser import PPTParser
+        parser = PPTParser.__new__(PPTParser)
+        parser.inter_dir = self.inter_dir
+        result = parser._find_title_in_transcript(new_title, corrected_segments)
+
+        if result is None:
+            # 匹配不到：不加入 clips（用户可在 GUI 中看到未匹配状态）
+            logger.info(f"标题 '{new_title}' 在转录中未找到匹配")
+            return None, False
+
+        start, end = result
+
+        # 3. 构建新 clip
+        new_clip = {
+            'title': new_title,
+            'start': start,
+            'end': end,
+            'matched': True,
+        }
+
+        # 4. 判断是否与现有 clip 重叠
+        overlapped_index = None
+        for i, clip in enumerate(self.clips):
+            if clip.get('matched', True) is False:
+                continue
+            if start < clip['end'] and end > clip['start']:
+                overlapped_index = i
+                break
+
+        if overlapped_index is not None:
+            # 有重叠：合并到现有 clip
+            self.clips.append(new_clip)
+            self.clips = parser._merge_overlapping_clips(self.clips)
+        else:
+            # 无重叠：直接追加
+            self.clips.append(new_clip)
+            # 重新排序
+            self.clips = sorted(self.clips, key=lambda c: c['start'])
+
+        # 找到新 clip 的索引
+        new_clip_index = next(
+            i for i, c in enumerate(self.clips)
+            if c['title'] == new_title and c['start'] == start
+        )
+
+        # 5. 保存 config
+        self._save_config()
+        return new_clip_index, True
+
     # ==================== 主流程 ====================
 
     def run(self) -> str: