init

backend/app/services/docx_parser.py

@@ -0,0 +1,93 @@
+import re
+from docx import Document
+
+
+def parse_transcript(file_path: str) -> list[dict]:
+    """
+    解析直播回放文字稿 DOCX
+    格式：发言者(时间戳): 内容
+
+    返回：[{index, speaker, timestamp, content, raw_text}]
+    """
+    doc = Document(file_path)
+    paragraphs = []
+    for i, para in enumerate(doc.paragraphs):
+        text = para.text.strip()
+        if not text:
+            continue
+
+        # 解析格式：发言者(时间戳): 内容
+        match = re.match(r'^(.+?)\((\d{2}:\d{2}:\d{2})\)\s*[:：]\s*(.*)', text)
+        if match:
+            speaker = match.group(1).strip()
+            timestamp = match.group(2)
+            content = match.group(3).strip()
+        else:
+            # 尝试格式：发言者: 内容
+            match2 = re.match(r'^(.+?)\s*[:：]\s*(.*)', text)
+            if match2:
+                speaker = match2.group(1).strip()
+                timestamp = ""
+                content = match2.group(2).strip()
+            else:
+                speaker = ""
+                timestamp = ""
+                content = text
+
+        paragraphs.append({
+            "index": i,
+            "speaker": speaker,
+            "timestamp": timestamp,
+            "content": content,
+            "raw_text": text,
+        })
+
+    return paragraphs
+
+
+def is_teacher(speaker: str) -> bool:
+    """判断是否为老师/助教发言"""
+    teacher_keywords = ["卢慧老师", "静静", "督导", "老师"]
+    return any(kw in speaker for kw in teacher_keywords)
+
+
+def parse_person_info(file_path: str) -> dict:
+    """
+    解析讲述者个人信息 DOCX
+    返回：{name, info_text, photos: [path]}
+    """
+    doc = Document(file_path)
+    texts = []
+    photos = []
+
+    # 提取文本
+    for para in doc.paragraphs:
+        text = para.text.strip()
+        if text:
+            texts.append(text)
+
+    # 提取图片
+    import os
+    photo_dir = os.path.join(os.path.dirname(file_path), "photos")
+    os.makedirs(photo_dir, exist_ok=True)
+
+    for i, rel in enumerate(doc.part.rels.values()):
+        if "image" in rel.reltype:
+            image = rel.target_part
+            photo_name = f"{os.path.splitext(os.path.basename(file_path))[0]}_photo_{i}.{image.content_type.split('/')[-1]}"
+            photo_path = os.path.join(photo_dir, photo_name)
+            with open(photo_path, "wb") as f:
+                f.write(image.blob)
+            photos.append(photo_path)
+
+    # 从文件名提取姓名
+    filename = os.path.basename(file_path)
+    name = re.sub(r'[-_].*$', '', filename).strip()
+    # 去掉扩展名
+    name = os.path.splitext(name)[0].strip()
+
+    return {
+        "name": name,
+        "info_text": "\n".join(texts),
+        "photos": photos,
+    }