fix
This commit is contained in:
root
238
python_core/scene_detection/workflows/single_workflow_manager.py
Normal file
238
python_core/scene_detection/workflows/single_workflow_manager.py
Normal file
@@ -0,0 +1,238 @@
|
||||
"""
|
||||
单个文件场景检测工作流管理器
|
||||
"""
|
||||
|
||||
from pathlib import Path
|
||||
from typing import Dict, Any, Optional, List
|
||||
|
||||
from python_core.utils.logger import logger
|
||||
from .single_workflow_nodes import SingleWorkflowNodes
|
||||
from ..types.single_workflow_state import SingleSceneDetectionWorkflowState
|
||||
from ..types.enums import DetectorType, OutputFormat
|
||||
|
||||
|
||||
class SingleWorkflowManager:
|
||||
"""单个文件场景检测工作流管理器"""
|
||||
|
||||
def __init__(self):
|
||||
self.workflow_nodes = SingleWorkflowNodes()
|
||||
|
||||
def detect_and_split_single_video(
|
||||
self,
|
||||
video_path: Path,
|
||||
output_dir: Optional[Path] = None,
|
||||
detector_type: DetectorType = DetectorType.CONTENT,
|
||||
threshold: float = 30.0,
|
||||
min_scene_length: float = 1.0,
|
||||
output_format: OutputFormat = OutputFormat.JSON,
|
||||
enable_ai_analysis: bool = False,
|
||||
enable_video_splitting: bool = True,
|
||||
use_advanced_split: bool = True,
|
||||
split_quality: int = 23,
|
||||
split_preset: str = "fast",
|
||||
max_video_duration: float = 60.0,
|
||||
project_id: Optional[str] = None,
|
||||
project_directory: Optional[Path] = None,
|
||||
material_tags: Optional[List[str]] = None,
|
||||
request_id: Optional[str] = None
|
||||
) -> Dict[str, Any]:
|
||||
"""
|
||||
单个视频文件的场景检测和切分
|
||||
|
||||
Args:
|
||||
video_path: 视频文件路径
|
||||
output_dir: 输出目录
|
||||
detector_type: 检测器类型
|
||||
threshold: 检测阈值
|
||||
min_scene_length: 最小场景长度
|
||||
output_format: 输出格式
|
||||
enable_ai_analysis: 是否启用AI分析
|
||||
enable_video_splitting: 是否启用视频切分
|
||||
use_advanced_split: 是否使用高级切分
|
||||
split_quality: 切分质量
|
||||
split_preset: 切分预设
|
||||
max_video_duration: 最大视频时长
|
||||
project_id: 项目ID（用于项目素材导入）
|
||||
project_directory: 项目目录（用于项目素材导入）
|
||||
material_tags: 素材标签
|
||||
request_id: 请求ID（用于JSON-RPC进度报告）
|
||||
|
||||
Returns:
|
||||
Dict: 处理结果
|
||||
"""
|
||||
|
||||
# 验证输入
|
||||
if not video_path.exists():
|
||||
raise FileNotFoundError(f"视频文件不存在: {video_path}")
|
||||
|
||||
if not video_path.suffix.lower() in {'.mp4', '.avi', '.mov', '.mkv', '.wmv', '.flv', '.webm', '.m4v'}:
|
||||
raise ValueError(f"不支持的视频格式: {video_path.suffix}")
|
||||
|
||||
# 设置默认输出目录
|
||||
if output_dir is None:
|
||||
output_dir = video_path.parent / f"{video_path.stem}_scenes"
|
||||
|
||||
# 确保输出目录存在
|
||||
output_dir.mkdir(parents=True, exist_ok=True)
|
||||
|
||||
logger.info(f"🚀 开始单个视频场景检测和切分")
|
||||
logger.info(f"📁 视频文件: {video_path}")
|
||||
logger.info(f"📂 输出目录: {output_dir}")
|
||||
logger.info(f"🎯 检测器: {detector_type.value}, 阈值: {threshold}")
|
||||
logger.info(f"✂️ 视频切分: {'启用' if enable_video_splitting else '禁用'}")
|
||||
|
||||
if project_id:
|
||||
logger.info(f"📦 项目导入: {project_id}")
|
||||
|
||||
# 创建工作流状态
|
||||
state = SingleSceneDetectionWorkflowState(
|
||||
video_path=video_path,
|
||||
output_dir=output_dir,
|
||||
detector_type=detector_type.value,
|
||||
threshold=threshold,
|
||||
min_scene_length=min_scene_length,
|
||||
output_format=output_format.value,
|
||||
enable_ai_analysis=enable_ai_analysis,
|
||||
enable_video_splitting=enable_video_splitting,
|
||||
use_advanced_split=use_advanced_split,
|
||||
split_quality=split_quality,
|
||||
split_preset=split_preset,
|
||||
max_video_duration=max_video_duration,
|
||||
project_id=project_id,
|
||||
project_directory=project_directory,
|
||||
material_tags=material_tags or [],
|
||||
request_id=request_id,
|
||||
enable_jsonrpc=request_id is not None
|
||||
)
|
||||
|
||||
try:
|
||||
# 执行工作流
|
||||
result = self.workflow_nodes.process_single_video(state)
|
||||
|
||||
# 构建返回结果
|
||||
response = {
|
||||
"success": result["success"],
|
||||
"video_path": str(video_path),
|
||||
"output_dir": str(output_dir),
|
||||
"processing_time": result.get("processing_time", 0),
|
||||
"total_scenes": result.get("total_scenes", 0),
|
||||
"total_segments": result.get("total_segments", 0),
|
||||
"state": state.to_dict()
|
||||
}
|
||||
|
||||
if result["success"]:
|
||||
# 转换检测结果为字典
|
||||
detection_result_dict = None
|
||||
if result.get("detection_result"):
|
||||
from dataclasses import asdict
|
||||
detection_result_dict = asdict(result["detection_result"])
|
||||
|
||||
response.update({
|
||||
"detection_result": detection_result_dict,
|
||||
"split_results": [
|
||||
{
|
||||
"scene_index": sr.scene_index,
|
||||
"output_path": str(sr.output_path),
|
||||
"start_time": sr.start_time,
|
||||
"end_time": sr.end_time,
|
||||
"duration": sr.duration,
|
||||
"file_size": sr.file_size,
|
||||
"success": sr.success,
|
||||
"error": sr.error
|
||||
}
|
||||
for sr in result.get("split_results", [])
|
||||
]
|
||||
})
|
||||
|
||||
logger.info(f"✅ 单个视频处理成功!")
|
||||
logger.info(f"📊 处理统计:")
|
||||
logger.info(f" 视频文件: {video_path.name}")
|
||||
logger.info(f" 处理时间: {result.get('processing_time', 0):.1f}s")
|
||||
logger.info(f" 场景数量: {result.get('total_scenes', 0)}")
|
||||
logger.info(f" 切分片段: {result.get('total_segments', 0)}")
|
||||
|
||||
else:
|
||||
response["error"] = result.get("error", "未知错误")
|
||||
logger.error(f"❌ 单个视频处理失败: {response['error']}")
|
||||
|
||||
return response
|
||||
|
||||
except Exception as e:
|
||||
error_msg = f"工作流执行失败: {str(e)}"
|
||||
logger.error(f"❌ {error_msg}")
|
||||
|
||||
return {
|
||||
"success": False,
|
||||
"error": error_msg,
|
||||
"video_path": str(video_path),
|
||||
"output_dir": str(output_dir),
|
||||
"processing_time": state.processing_time or 0,
|
||||
"state": state.to_dict()
|
||||
}
|
||||
|
||||
def detect_and_split_for_project(
|
||||
self,
|
||||
video_path: Path,
|
||||
project_id: str,
|
||||
project_directory: Path,
|
||||
material_tags: Optional[List[str]] = None,
|
||||
detector_type: DetectorType = DetectorType.CONTENT,
|
||||
threshold: float = 30.0,
|
||||
min_scene_length: float = 1.0,
|
||||
split_quality: int = 23,
|
||||
split_preset: str = "fast",
|
||||
max_video_duration: float = 2.0,
|
||||
request_id: Optional[str] = None
|
||||
) -> Dict[str, Any]:
|
||||
"""
|
||||
为项目导入单个视频素材
|
||||
|
||||
Args:
|
||||
video_path: 视频文件路径
|
||||
project_id: 项目ID
|
||||
project_directory: 项目目录
|
||||
material_tags: 素材标签
|
||||
其他参数: 检测和切分配置
|
||||
|
||||
Returns:
|
||||
Dict: 处理结果
|
||||
"""
|
||||
|
||||
# 创建临时输出目录
|
||||
temp_output_dir = project_directory / "temp" / f"{video_path.stem}_processing"
|
||||
|
||||
return self.detect_and_split_single_video(
|
||||
video_path=video_path,
|
||||
output_dir=temp_output_dir,
|
||||
detector_type=detector_type,
|
||||
threshold=threshold,
|
||||
min_scene_length=min_scene_length,
|
||||
output_format=OutputFormat.JSON,
|
||||
enable_ai_analysis=False,
|
||||
enable_video_splitting=True,
|
||||
use_advanced_split=True,
|
||||
split_quality=split_quality,
|
||||
split_preset=split_preset,
|
||||
max_video_duration=max_video_duration,
|
||||
project_id=project_id,
|
||||
project_directory=project_directory,
|
||||
material_tags=material_tags,
|
||||
request_id=request_id
|
||||
)
|
||||
|
||||
def get_supported_formats(self) -> List[str]:
|
||||
"""获取支持的视频格式"""
|
||||
return ['.mp4', '.avi', '.mov', '.mkv', '.wmv', '.flv', '.webm', '.m4v']
|
||||
|
||||
def validate_video_file(self, video_path: Path) -> bool:
|
||||
"""验证视频文件"""
|
||||
if not video_path.exists():
|
||||
return False
|
||||
|
||||
if not video_path.is_file():
|
||||
return False
|
||||
|
||||
if video_path.suffix.lower() not in self.get_supported_formats():
|
||||
return False
|
||||
|
||||
return True
|
||||
315
python_core/scene_detection/workflows/single_workflow_nodes.py
Normal file
315
python_core/scene_detection/workflows/single_workflow_nodes.py
Normal file
@@ -0,0 +1,315 @@
|
||||
"""
|
||||
单个文件场景检测工作流节点
|
||||
"""
|
||||
|
||||
import os
|
||||
import json
|
||||
import hashlib
|
||||
from pathlib import Path
|
||||
from typing import Dict, Any
|
||||
from datetime import datetime
|
||||
|
||||
from python_core.utils.logger import logger
|
||||
from python_core.utils.jsonrpc_enhanced import ProgressLevel
|
||||
|
||||
from ..services.detector_service import SceneDetectorService
|
||||
from ..services.ai_analysis_service import AIAnalysisService
|
||||
from ..utils.result_saver import ResultSaver
|
||||
from ..types.single_workflow_state import SingleSceneDetectionWorkflowState, SingleVideoTask
|
||||
from python_core.services.ffmpeg_slice_service_sync import FfmpegSliceService, SliceOptions, SliceSegment
|
||||
from python_core.services.project_material_service import ProjectMaterialService
|
||||
|
||||
|
||||
class SingleWorkflowNodes:
|
||||
"""单个文件场景检测工作流节点"""
|
||||
|
||||
def __init__(self):
|
||||
self.detector_service = SceneDetectorService()
|
||||
self.ai_service = AIAnalysisService()
|
||||
self.result_saver = ResultSaver()
|
||||
self.splitter_service = FfmpegSliceService()
|
||||
self.material_service = ProjectMaterialService()
|
||||
|
||||
def process_single_video(self, state: SingleSceneDetectionWorkflowState) -> Dict[str, Any]:
|
||||
"""处理单个视频文件"""
|
||||
try:
|
||||
state.mark_task_started()
|
||||
state.update_progress("started", 0.0, f"开始处理视频: {state.video_path.name}")
|
||||
|
||||
logger.info(f"🎬 开始处理单个视频: {state.video_path.name}")
|
||||
|
||||
# 1. 场景检测
|
||||
state.update_progress("detecting", 20.0, "正在进行场景检测...")
|
||||
|
||||
# 转换检测器类型
|
||||
from ..types.enums import DetectorType
|
||||
detector_type_enum = DetectorType(state.detector_type)
|
||||
|
||||
detection_result = self.detector_service.detect_scenes(
|
||||
video_path=state.video_path,
|
||||
detector_type=detector_type_enum,
|
||||
threshold=state.threshold,
|
||||
min_scene_length=state.min_scene_length
|
||||
)
|
||||
|
||||
logger.info(f"✅ 场景检测完成: 检测到 {detection_result.total_scenes} 个场景")
|
||||
|
||||
# 2. 保存检测结果
|
||||
if state.output_dir:
|
||||
state.update_progress("saving", 40.0, "保存检测结果...")
|
||||
|
||||
# 转换输出格式
|
||||
from ..types.enums import OutputFormat
|
||||
output_format_enum = OutputFormat(state.output_format)
|
||||
|
||||
# 创建结果文件路径
|
||||
result_filename = f"{state.video_path.stem}_scenes.{output_format_enum.value}"
|
||||
result_path = state.output_dir / result_filename
|
||||
|
||||
self.result_saver.save_results(
|
||||
detection_result,
|
||||
result_path,
|
||||
output_format_enum
|
||||
)
|
||||
|
||||
# 3. AI分析（如果启用）
|
||||
if state.enable_ai_analysis:
|
||||
state.update_progress("ai_analysis", 60.0, "正在进行AI分析...")
|
||||
try:
|
||||
ai_result = self.ai_service.analyze_scenes(detection_result)
|
||||
detection_result.ai_analysis = ai_result
|
||||
logger.info("✅ AI分析完成")
|
||||
except Exception as e:
|
||||
logger.warning(f"⚠️ AI分析失败: {e}")
|
||||
|
||||
# 4. 视频切分（如果启用）
|
||||
split_results = []
|
||||
if state.enable_video_splitting and state.output_dir:
|
||||
state.update_progress("splitting", 70.0, "正在进行视频切分...")
|
||||
split_results = self._handle_video_splitting(state, detection_result)
|
||||
|
||||
# 5. 项目素材管理（如果是项目导入）
|
||||
if state.project_id and state.project_directory:
|
||||
state.update_progress("project_import", 90.0, "导入到项目素材库...")
|
||||
self._handle_project_material_import(state, detection_result, split_results)
|
||||
|
||||
# 标记完成
|
||||
state.mark_task_completed(detection_result, split_results)
|
||||
state.update_progress("completed", 100.0, "处理完成")
|
||||
|
||||
logger.info(f"🎉 视频处理完成: {state.video_path.name}")
|
||||
|
||||
return {
|
||||
"success": True,
|
||||
"detection_result": detection_result,
|
||||
"split_results": split_results,
|
||||
"processing_time": state.processing_time,
|
||||
"total_scenes": state.total_scenes,
|
||||
"total_segments": state.total_split_segments
|
||||
}
|
||||
|
||||
except Exception as e:
|
||||
error_msg = f"处理视频失败: {str(e)}"
|
||||
logger.error(f"❌ {error_msg}")
|
||||
state.mark_task_failed(error_msg)
|
||||
state.update_progress("failed", 0.0, error_msg)
|
||||
|
||||
return {
|
||||
"success": False,
|
||||
"error": error_msg,
|
||||
"processing_time": state.processing_time
|
||||
}
|
||||
|
||||
def _handle_video_splitting(self, state: SingleSceneDetectionWorkflowState, detection_result) -> list:
|
||||
"""处理视频切分"""
|
||||
try:
|
||||
# 创建切分选项
|
||||
slice_options = SliceOptions(
|
||||
crf=state.split_quality,
|
||||
preset=state.split_preset
|
||||
)
|
||||
|
||||
# 转换场景为SliceSegment
|
||||
segments = [
|
||||
SliceSegment(start=scene.start_time, end=scene.end_time)
|
||||
for scene in detection_result.scenes
|
||||
]
|
||||
|
||||
# 创建输出目录
|
||||
scenes_dir = state.output_dir / "scenes"
|
||||
scenes_dir.mkdir(parents=True, exist_ok=True)
|
||||
|
||||
# 检查是否无场景，如果是则跳过切分
|
||||
if len(segments) <= 0:
|
||||
logger.info(f"🔄 检测到 0 个场景切换，跳过切分，直接使用源文件")
|
||||
|
||||
# 直接复制源文件作为结果
|
||||
import shutil
|
||||
source_file = state.video_path
|
||||
target_file = scenes_dir / f"{state.video_path.stem}_scene_001.mp4"
|
||||
|
||||
try:
|
||||
shutil.copy2(source_file, target_file)
|
||||
logger.info(f"✅ 源文件已复制到: {target_file}")
|
||||
|
||||
# 获取源文件元数据
|
||||
metadata = self.splitter_service.get_video_metadata(str(source_file))
|
||||
|
||||
# 创建模拟的切分结果
|
||||
slice_results = [(str(target_file), metadata)]
|
||||
|
||||
# 修正场景统计：无场景切换 = 1个场景（整个视频）
|
||||
detection_result.total_scenes = 1
|
||||
|
||||
except Exception as e:
|
||||
logger.error(f"❌ 复制源文件失败: {e}")
|
||||
slice_results = []
|
||||
else:
|
||||
logger.info(f"🎬 检测到 {len(segments)} 个场景，开始切分")
|
||||
|
||||
# 生成输出路径
|
||||
base_output_path = str(scenes_dir / f"{state.video_path.stem}_scene")
|
||||
|
||||
# 使用同步方法进行切分
|
||||
slice_results = self.splitter_service.slice_video(
|
||||
media_path=str(state.video_path),
|
||||
segments=segments,
|
||||
options=slice_options,
|
||||
output_path=base_output_path
|
||||
)
|
||||
|
||||
# 转换为兼容格式
|
||||
split_results_raw = []
|
||||
|
||||
# 创建一个简单的结果对象
|
||||
class SplitResult:
|
||||
def __init__(self, scene_index, output_path, start_time, end_time, duration, file_size, success, error=None):
|
||||
self.scene_index = scene_index
|
||||
self.output_path = Path(output_path)
|
||||
self.start_time = start_time
|
||||
self.end_time = end_time
|
||||
self.duration = duration
|
||||
self.file_size = file_size
|
||||
self.success = success
|
||||
self.error = error
|
||||
|
||||
for i, (output_path, metadata) in enumerate(slice_results):
|
||||
if len(detection_result.scenes) == 0:
|
||||
# 无场景切换的情况：整个视频作为一个场景
|
||||
split_result = SplitResult(
|
||||
scene_index=0,
|
||||
output_path=output_path,
|
||||
start_time=0.0,
|
||||
end_time=metadata.duration,
|
||||
duration=metadata.duration,
|
||||
file_size=metadata.size,
|
||||
success=Path(output_path).exists(),
|
||||
error=None if Path(output_path).exists() else "文件不存在"
|
||||
)
|
||||
split_results_raw.append(split_result)
|
||||
else:
|
||||
# 有场景切换的情况
|
||||
scene = detection_result.scenes[i] if i < len(detection_result.scenes) else None
|
||||
if scene:
|
||||
split_result = SplitResult(
|
||||
scene_index=i,
|
||||
output_path=output_path,
|
||||
start_time=scene.start_time,
|
||||
end_time=scene.end_time,
|
||||
duration=scene.duration,
|
||||
file_size=metadata.size,
|
||||
success=Path(output_path).exists(),
|
||||
error=None if Path(output_path).exists() else "文件不存在"
|
||||
)
|
||||
split_results_raw.append(split_result)
|
||||
|
||||
# 4. 添加视频时长检查，如果时长大于最大视频时长那么就要进行二次切分
|
||||
final_split_results = []
|
||||
|
||||
for split_result in split_results_raw:
|
||||
try:
|
||||
# 检查切分后的视频时长
|
||||
video_path = str(split_result.output_path)
|
||||
|
||||
if split_result.duration > state.max_video_duration:
|
||||
logger.info(f"⚠️ 片段 {split_result.scene_index + 1} 时长 {split_result.duration:.2f}s 超过限制 {state.max_video_duration:.2f}s，进行二次切分")
|
||||
|
||||
# 进行二次切分
|
||||
secondary_results = self.splitter_service.check_and_split_by_duration(
|
||||
video_path=video_path,
|
||||
max_duration=state.max_video_duration,
|
||||
options=slice_options,
|
||||
output_dir=str(split_result.output_path.parent)
|
||||
)
|
||||
|
||||
# 处理二次切分结果
|
||||
for i, (secondary_path, secondary_metadata) in enumerate(secondary_results):
|
||||
if len(secondary_results) > 1:
|
||||
# 如果进行了二次切分，创建新的结果对象
|
||||
secondary_split_result = SplitResult(
|
||||
scene_index=split_result.scene_index,
|
||||
output_path=secondary_path,
|
||||
start_time=split_result.start_time + (i * state.max_video_duration),
|
||||
end_time=min(split_result.start_time + ((i + 1) * state.max_video_duration), split_result.end_time),
|
||||
duration=secondary_metadata.duration,
|
||||
file_size=secondary_metadata.size,
|
||||
success=Path(secondary_path).exists(),
|
||||
error=None if Path(secondary_path).exists() else "二次切分文件不存在"
|
||||
)
|
||||
final_split_results.append(secondary_split_result)
|
||||
|
||||
# 删除原始的过长片段（如果二次切分成功）
|
||||
if Path(video_path).exists() and len(secondary_results) > 1:
|
||||
try:
|
||||
Path(video_path).unlink()
|
||||
logger.info(f"🗑️ 已删除过长的原始片段: {Path(video_path).name}")
|
||||
except Exception as e:
|
||||
logger.warning(f"⚠️ 删除原始片段失败: {e}")
|
||||
else:
|
||||
# 如果没有进行二次切分（时长检查通过），保持原结果
|
||||
final_split_results.append(split_result)
|
||||
else:
|
||||
# 时长未超过限制，直接添加到最终结果
|
||||
final_split_results.append(split_result)
|
||||
|
||||
except Exception as e:
|
||||
logger.error(f"❌ 处理片段 {split_result.scene_index + 1} 的时长检查时出错: {e}")
|
||||
# 出错时保持原结果
|
||||
final_split_results.append(split_result)
|
||||
|
||||
# 使用最终的切分结果
|
||||
split_results_raw = final_split_results
|
||||
|
||||
# 更新统计信息
|
||||
logger.info(f"📊 最终生成 {len(split_results_raw)} 个视频片段（包含二次切分）")
|
||||
|
||||
return split_results_raw
|
||||
|
||||
except Exception as e:
|
||||
logger.error(f"❌ 视频切分失败: {e}")
|
||||
return []
|
||||
|
||||
def _handle_project_material_import(self, state: SingleSceneDetectionWorkflowState, detection_result, split_results):
|
||||
"""处理项目素材导入"""
|
||||
try:
|
||||
if not state.project_directory:
|
||||
logger.warning("⚠️ 项目目录未设置，跳过素材导入")
|
||||
return
|
||||
|
||||
# 使用项目素材管理服务
|
||||
result = self.material_service.import_video_materials(
|
||||
video_segments=split_results,
|
||||
project_id=state.project_id,
|
||||
project_directory=state.project_directory,
|
||||
source_video_path=str(state.video_path),
|
||||
material_tags=state.material_tags
|
||||
)
|
||||
|
||||
if result["success"]:
|
||||
logger.info(f"📝 项目素材导入完成: 新增 {result['imported_count']} 个素材，跳过 {result['skipped_count']} 个重复素材")
|
||||
else:
|
||||
logger.error(f"❌ 项目素材导入失败: {result.get('error', '未知错误')}")
|
||||
|
||||
except Exception as e:
|
||||
logger.error(f"❌ 项目素材导入失败: {e}")
|
||||
|
||||
Reference in New Issue
Block a user