feat(script): 添加帧间隔输入参数并更新相关逻辑

- 在 GenerateScriptRequest 和 ScriptGenerator 中新增 frame_interval_input 参数，默认为 5
- 更新 VideoProcessor 的 process_video_pipeline 方法，调整帧提取间隔逻辑
- 修改前端界面，增加帧间隔输入框并更新相关翻译文本
- 优化 generate_script_docu 函数，使用新的帧间隔参数进行视频处理

app/models/schema_v2.py

@@ -6,6 +6,7 @@ class GenerateScriptRequest(BaseModel):
     video_path: str
     video_theme: Optional[str] = ""
     custom_prompt: Optional[str] = ""
+    frame_interval_input: Optional[int] = 5
     skip_seconds: Optional[int] = 0
     threshold: Optional[int] = 30
     vision_batch_size: Optional[int] = 5

app/services/script_service.py

@@ -22,6 +22,7 @@ class ScriptGenerator:
         video_path: str,
         video_theme: str = "",
         custom_prompt: str = "",
+        frame_interval_input: int = 5,
         skip_seconds: int = 0,
         threshold: int = 30,
         vision_batch_size: int = 5,

app/utils/video_processor.py

@@ -292,10 +292,7 @@ class VideoProcessor:
     def process_video_pipeline(self,
                               output_dir: str,
                               skip_seconds: float = 0.0,
-                              threshold: int = 20,  # 此参数保留但不使用
+                              interval_seconds: float = 5.0,  # 帧提取间隔（秒）
-                              compressed_width: int = 320,  # 此参数保留但不使用
-                              keep_temp: bool = False,  # 此参数保留但不使用
-                              interval_seconds: float = 5.0,
                               use_hw_accel: bool = True) -> None:
         """
         执行简化的视频处理流程，直接从原视频按固定时间间隔提取帧
@@ -303,9 +300,6 @@ class VideoProcessor:
         Args:
             output_dir: 输出目录
             skip_seconds: 跳过视频开头的秒数
-            threshold: 保留参数，不使用
-            compressed_width: 保留参数，不使用
-            keep_temp: 保留参数，不使用
             interval_seconds: 帧提取间隔（秒）
             use_hw_accel: 是否使用硬件加速
         """
@@ -314,7 +308,7 @@ class VideoProcessor:
         
         try:
             # 直接从原视频提取关键帧
-            logger.info("从视频直接提取关键帧...")
+            logger.info(f"从视频间隔 {interval_seconds} 秒提取关键帧...")
             self.extract_frames_by_interval(
                 output_dir,
                 interval_seconds=interval_seconds,

webui/components/script_settings.py

@@ -47,7 +47,7 @@ def render_script_file(tr, params):
         (tr("None"), ""), 
         (tr("Auto Generate"), "auto"), 
         (tr("Short Generate"), "short"),
-        (tr("Upload Script"), "upload_script")  # 新增上传脚本选项
+        (tr("Upload Script"), "upload_script")
     ]
 
     # 获取已有脚本文件
@@ -214,38 +214,16 @@ def render_script_buttons(tr, params):
     # 根据脚本类型显示不同的设置
     if script_path != "short":
         # 非短视频模式下显示原有的三个输入框
-        input_cols = st.columns(3)
+        input_cols = st.columns(2)
         
         with input_cols[0]:
-            skip_seconds = st.number_input(
+            st.number_input(
-                "skip_seconds",
+                tr("Frame Interval (seconds)"),
                 min_value=0,
-                value=st.session_state.get('skip_seconds', config.frames.get('skip_seconds', 0)),
+                value=st.session_state.get('frame_interval_input', config.frames.get('frame_interval_input', 5)),
-                help=tr("Skip the first few seconds"),
+                help=tr("Frame Interval (seconds) (More keyframes consume more tokens)"),
-                key="skip_seconds_input"
+                key="frame_interval_input"
             )
-            st.session_state['skip_seconds'] = skip_seconds
-            
-        with input_cols[1]:
-            threshold = st.number_input(
-                "threshold",
-                min_value=0,
-                value=st.session_state.get('threshold', config.frames.get('threshold', 30)),
-                help=tr("Difference threshold"),
-                key="threshold_input"
-            )
-            st.session_state['threshold'] = threshold
-            
-        with input_cols[2]:
-            vision_batch_size = st.number_input(
-                "vision_batch_size",
-                min_value=1,
-                max_value=20,
-                value=st.session_state.get('vision_batch_size', config.frames.get('vision_batch_size', 5)),
-                help=tr("Vision processing batch size"),
-                key="vision_batch_size_input"
-            )
-            st.session_state['vision_batch_size'] = vision_batch_size
 
     # 生成/加载按钮
     if script_path == "auto":
@@ -259,7 +237,8 @@ def render_script_buttons(tr, params):
 
     if st.button(button_name, key="script_action", disabled=not script_path):
         if script_path == "auto":
-            generate_script_docu(tr, params)
+            # 执行纪录片视频脚本生成（视频无字幕无配音）
+            generate_script_docu(params)
         elif script_path == "short":
             # 获取自定义片段数量参数
             custom_clips = st.session_state.get('custom_clips', 5)

webui/i18n/en.json

@@ -85,6 +85,7 @@
     "TTS Provider": "TTS Provider",
     "Hide Log": "Hide Log",
     "Upload Local Files": "Upload Local Files",
-    "File Uploaded Successfully": "File Uploaded Successfully"
+    "File Uploaded Successfully": "File Uploaded Successfully",
+    "Frame Interval (seconds)": "Frame Interval (seconds) (More keyframes consume more tokens)"
   }
 }

webui/i18n/zh.json

@@ -192,6 +192,8 @@
     "Generate Short Video Script": "AI生成短剧混剪脚本",
     "Adjust the volume of the original audio": "调整原始音频的音量",
     "Original Volume": "视频音量",
-    "Auto Generate": "纪录片解说 (画面解说)"
+    "Auto Generate": "纪录片解说 (画面解说)",
+    "Frame Interval (seconds)": "帧间隔 (秒)",
+    "Frame Interval (seconds) (More keyframes consume more tokens)": "帧间隔 (秒) (更多关键帧消耗更多令牌)"
   }
 }

webui/tools/generate_script_docu.py

@@ -17,9 +17,11 @@ from app.utils import utils, video_processor, qwenvl_analyzer
 from webui.tools.base import create_vision_analyzer, get_batch_files, get_batch_timestamps, chekc_video_config
 
 
-def generate_script_docu(tr, params):
+def generate_script_docu(params):
     """
     生成 纪录片 视频脚本
+    要求: 原视频无字幕无配音
+    适合场景: 纪录片、动物搞笑解说、荒野建造等
     """
     progress_bar = st.progress(0)
     status_text = st.empty()
@@ -69,8 +71,7 @@ def generate_script_docu(tr, params):
                     # 处理视频并提取关键帧
                     processor.process_video_pipeline(
                         output_dir=video_keyframes_dir,
-                        skip_seconds=st.session_state.get('skip_seconds'),
+                        interval_seconds=st.session_state.get('frame_interval_input'),
-                        threshold=st.session_state.get('threshold')
                     )
 
                     # 获取所有关键文件路径
@@ -194,6 +195,7 @@ def generate_script_docu(tr, params):
                     _, _, timestamp_range = get_batch_timestamps(batch_files, prev_batch_files)
 
                     frame_content = {
+                        "_id": i + 1,
                         "timestamp": timestamp_range,
                         "picture": result['response'],
                         "narration": "",