feat: Add comprehensive timeline editor with frame editing and regeneration capabilities

This commit is contained in:
empty
2026-01-05 14:48:43 +08:00
parent 7d78dcd078
commit ca018a9b1f
68 changed files with 14904 additions and 57 deletions

View File

@@ -54,6 +54,8 @@ from api.routers import (
files_router,
resources_router,
frame_router,
editor_router,
publish_router,
)
@@ -133,6 +135,8 @@ app.include_router(tasks_router, prefix=api_config.api_prefix)
app.include_router(files_router, prefix=api_config.api_prefix)
app.include_router(resources_router, prefix=api_config.api_prefix)
app.include_router(frame_router, prefix=api_config.api_prefix)
app.include_router(editor_router, prefix=api_config.api_prefix)
app.include_router(publish_router, prefix=api_config.api_prefix)
@app.get("/")
@@ -153,6 +157,8 @@ async def root():
"files": f"{api_config.api_prefix}/files",
"resources": f"{api_config.api_prefix}/resources",
"frame": f"{api_config.api_prefix}/frame",
"editor": f"{api_config.api_prefix}/editor",
"publish": f"{api_config.api_prefix}/publish",
}
}

View File

@@ -24,6 +24,8 @@ from api.routers.tasks import router as tasks_router
from api.routers.files import router as files_router
from api.routers.resources import router as resources_router
from api.routers.frame import router as frame_router
from api.routers.editor import router as editor_router
from api.routers.publish import router as publish_router
__all__ = [
"health_router",
@@ -36,5 +38,8 @@ __all__ = [
"files_router",
"resources_router",
"frame_router",
"editor_router",
"publish_router",
]

579
api/routers/editor.py Normal file
View File

@@ -0,0 +1,579 @@
# Copyright (C) 2025 AIDC-AI
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
# http://www.apache.org/licenses/LICENSE-2.0
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
"""
Editor API router for timeline editor operations
Provides endpoints for:
- Fetching storyboard data
- Reordering frames
- Updating frame duration
- Generating preview
"""
from fastapi import APIRouter, HTTPException, Path
from loguru import logger
from api.schemas.editor import (
StoryboardSchema,
StoryboardFrameSchema,
ReorderFramesRequest,
UpdateDurationRequest,
PreviewRequest,
PreviewResponse,
UpdateFrameRequest,
UpdateFrameResponse,
RegenerateImageRequest,
RegenerateImageResponse,
RegenerateAudioRequest,
RegenerateAudioResponse,
)
router = APIRouter(prefix="/editor", tags=["Editor"])
def _path_to_url(file_path: str, base_url: str = "http://localhost:8000") -> str:
"""Convert local file path to URL accessible through API"""
if not file_path:
return None
import os
from pathlib import Path
# Normalize path separators
file_path = file_path.replace("\\", "/")
# Extract relative path from output directory
parts = file_path.split("/")
try:
output_idx = parts.index("output")
relative_parts = parts[output_idx + 1:]
relative_path = "/".join(relative_parts)
except ValueError:
relative_path = Path(file_path).name
return f"{base_url}/api/files/{relative_path}"
# In-memory cache for demo (in production, use database)
_storyboard_cache: dict = {}
# Demo data for testing
_demo_storyboard = {
"id": "demo-1",
"title": "演示视频",
"total_duration": 15.5,
"final_video_path": None,
"created_at": None,
"frames": [
{"id": "frame-0", "index": 0, "order": 0, "narration": "在一个宁静的早晨,阳光洒满了整个城市", "image_prompt": "A peaceful morning", "duration": 3.2},
{"id": "frame-1", "index": 1, "order": 1, "narration": "小明决定出门去探索这个美丽的世界", "image_prompt": "A young man stepping out", "duration": 2.8},
{"id": "frame-2", "index": 2, "order": 2, "narration": "他走过熟悉的街道,感受着微风的吹拂", "image_prompt": "Walking through streets", "duration": 3.5},
{"id": "frame-3", "index": 3, "order": 3, "narration": "公园里的花朵正在盛开,散发着迷人的芬芳", "image_prompt": "Blooming flowers", "duration": 3.0},
{"id": "frame-4", "index": 4, "order": 4, "narration": "这是新的一天的开始,充满了无限可能", "image_prompt": "New day begins", "duration": 3.0},
],
}
# Import task manager
from api.tasks.manager import task_manager
@router.get("/storyboard/{storyboard_id}", response_model=StoryboardSchema)
async def get_storyboard(storyboard_id: str = Path(..., description="Storyboard/task ID")):
"""
Get storyboard by ID
Supports:
- 'demo-1': Returns demo data for testing
- Any task_id: Loads real storyboard from completed video generation tasks
- History tasks: Loads from persistence service
"""
# Return demo data for demo-1
if storyboard_id == "demo-1":
if "demo-1" not in _storyboard_cache:
_storyboard_cache["demo-1"] = _demo_storyboard.copy()
return _storyboard_cache["demo-1"]
# Try to get from cache first
if storyboard_id in _storyboard_cache:
return _storyboard_cache[storyboard_id]
# Try to load from task manager (in-memory task)
task = task_manager.get_task(storyboard_id)
if task and task.result:
# Extract storyboard from task result
result = task.result
# Handle different result formats
storyboard_data = None
if hasattr(result, 'storyboard'):
storyboard_data = result.storyboard
elif isinstance(result, dict) and 'storyboard' in result:
storyboard_data = result['storyboard']
if storyboard_data:
# Convert to editor schema format
schema = _convert_storyboard_to_schema(storyboard_id, storyboard_data)
_storyboard_cache[storyboard_id] = schema
logger.info(f"Loaded storyboard from task {storyboard_id}")
return schema
# Try to load from persistence service (history)
try:
from pixelle_video.services.persistence import PersistenceService
persistence = PersistenceService(output_dir="output")
# Load storyboard from disk (await since we're in an async function)
storyboard = await persistence.load_storyboard(storyboard_id)
if storyboard:
schema = _convert_storyboard_to_schema(storyboard_id, storyboard)
_storyboard_cache[storyboard_id] = schema
logger.info(f"Loaded storyboard from persistence {storyboard_id}")
return schema
except Exception as e:
logger.warning(f"Failed to load from persistence: {e}")
raise HTTPException(status_code=404, detail=f"Storyboard {storyboard_id} not found")
def _convert_storyboard_to_schema(storyboard_id: str, storyboard) -> dict:
"""Convert internal Storyboard model to API schema format."""
frames = []
# Handle both object and dict formats
if hasattr(storyboard, 'frames'):
frame_list = storyboard.frames
title = getattr(storyboard, 'title', storyboard_id)
total_duration = getattr(storyboard, 'total_duration', 0)
final_video_path = getattr(storyboard, 'final_video_path', None)
created_at = getattr(storyboard, 'created_at', None)
elif isinstance(storyboard, dict):
frame_list = storyboard.get('frames', [])
title = storyboard.get('title', storyboard_id)
total_duration = storyboard.get('total_duration', 0)
final_video_path = storyboard.get('final_video_path')
created_at = storyboard.get('created_at')
else:
frame_list = []
title = storyboard_id
total_duration = 0
final_video_path = None
created_at = None
for i, frame in enumerate(frame_list):
if hasattr(frame, 'narration'):
# Object format
frames.append({
"id": f"frame-{i}",
"index": getattr(frame, 'index', i),
"order": i,
"narration": frame.narration or "",
"image_prompt": getattr(frame, 'image_prompt', ""),
"image_path": _path_to_url(getattr(frame, 'image_path', None)),
"audio_path": _path_to_url(getattr(frame, 'audio_path', None)),
"video_segment_path": _path_to_url(getattr(frame, 'video_segment_path', None)),
"duration": getattr(frame, 'duration', 3.0),
})
elif isinstance(frame, dict):
# Dict format
frames.append({
"id": f"frame-{i}",
"index": frame.get('index', i),
"order": i,
"narration": frame.get('narration', ""),
"image_prompt": frame.get('image_prompt', ""),
"image_path": _path_to_url(frame.get('image_path')),
"audio_path": _path_to_url(frame.get('audio_path')),
"video_segment_path": _path_to_url(frame.get('video_segment_path')),
"duration": frame.get('duration', 3.0),
})
return {
"id": storyboard_id,
"title": title,
"frames": frames,
"total_duration": total_duration or sum(f.get('duration', 3.0) for f in frames),
"final_video_path": final_video_path,
"created_at": created_at.isoformat() if created_at else None,
}
@router.patch("/storyboard/{storyboard_id}/reorder", response_model=StoryboardSchema)
async def reorder_frames(
storyboard_id: str = Path(..., description="Storyboard/task ID"),
request: ReorderFramesRequest = None
):
"""
Reorder frames in storyboard
Updates the order of frames based on the provided frame ID list.
"""
if storyboard_id not in _storyboard_cache:
raise HTTPException(status_code=404, detail=f"Storyboard {storyboard_id} not found in cache")
storyboard = _storyboard_cache[storyboard_id]
frames = storyboard["frames"]
# Create ID to frame mapping
frame_map = {f["id"]: f for f in frames}
# Validate all IDs exist
for frame_id in request.order:
if frame_id not in frame_map:
raise HTTPException(status_code=400, detail=f"Frame {frame_id} not found")
# Reorder frames
reordered = []
for idx, frame_id in enumerate(request.order):
frame = frame_map[frame_id].copy()
frame["order"] = idx
reordered.append(frame)
storyboard["frames"] = reordered
_storyboard_cache[storyboard_id] = storyboard
logger.info(f"Reordered {len(reordered)} frames in storyboard {storyboard_id}")
return storyboard
@router.patch(
"/storyboard/{storyboard_id}/frames/{frame_id}/duration",
response_model=StoryboardFrameSchema
)
async def update_frame_duration(
storyboard_id: str = Path(..., description="Storyboard/task ID"),
frame_id: str = Path(..., description="Frame ID"),
request: UpdateDurationRequest = None
):
"""
Update frame duration
Changes the duration of a specific frame and recalculates total duration.
"""
if storyboard_id not in _storyboard_cache:
raise HTTPException(status_code=404, detail=f"Storyboard {storyboard_id} not found in cache")
storyboard = _storyboard_cache[storyboard_id]
frames = storyboard["frames"]
# Find and update frame
updated_frame = None
for frame in frames:
if frame["id"] == frame_id:
frame["duration"] = request.duration
updated_frame = frame
break
if not updated_frame:
raise HTTPException(status_code=404, detail=f"Frame {frame_id} not found")
# Recalculate total duration
storyboard["total_duration"] = sum(f["duration"] for f in frames)
_storyboard_cache[storyboard_id] = storyboard
logger.info(f"Updated frame {frame_id} duration to {request.duration}s")
return updated_frame
@router.post("/storyboard/{storyboard_id}/preview", response_model=PreviewResponse)
async def generate_preview(
storyboard_id: str = Path(..., description="Storyboard/task ID"),
request: PreviewRequest = None
):
"""
Generate preview video for selected frames
Creates a preview video from the specified frame range.
"""
if storyboard_id not in _storyboard_cache:
raise HTTPException(status_code=404, detail=f"Storyboard {storyboard_id} not found in cache")
storyboard = _storyboard_cache[storyboard_id]
frames = storyboard["frames"]
# Determine frame range
start = request.start_frame if request else 0
end = request.end_frame if request and request.end_frame else len(frames)
if start >= len(frames):
raise HTTPException(status_code=400, detail="Start frame out of range")
preview_frames = frames[start:end]
total_duration = sum(f["duration"] for f in preview_frames)
# TODO: Implement actual preview generation logic
# For now, return mock response
preview_path = f"/output/{storyboard_id}/preview_{start}_{end}.mp4"
logger.info(f"Generated preview for frames {start}-{end} ({len(preview_frames)} frames)")
return PreviewResponse(
preview_path=preview_path,
duration=total_duration,
frames_count=len(preview_frames)
)
def _storyboard_to_schema(storyboard_id: str, storyboard) -> dict:
"""Convert internal Storyboard to API schema format"""
frames = []
for i, frame in enumerate(storyboard.frames):
frames.append({
"id": f"frame-{i}",
"index": frame.index,
"order": i,
"narration": frame.narration,
"image_prompt": frame.image_prompt,
"image_path": frame.image_path,
"audio_path": frame.audio_path,
"video_segment_path": frame.video_segment_path,
"duration": frame.duration,
})
return {
"id": storyboard_id,
"title": storyboard.title,
"frames": frames,
"total_duration": storyboard.total_duration,
"final_video_path": storyboard.final_video_path,
"created_at": storyboard.created_at,
}
@router.put(
"/storyboard/{storyboard_id}/frames/{frame_id}",
response_model=UpdateFrameResponse
)
async def update_frame(
storyboard_id: str = Path(..., description="Storyboard/task ID"),
frame_id: str = Path(..., description="Frame ID"),
request: UpdateFrameRequest = None
):
"""
Update frame content (narration and/or image prompt)
Updates the text content of a frame without regenerating media.
"""
if storyboard_id not in _storyboard_cache:
raise HTTPException(status_code=404, detail=f"Storyboard {storyboard_id} not found in cache")
storyboard = _storyboard_cache[storyboard_id]
frames = storyboard["frames"]
# Find and update frame
updated_frame = None
for frame in frames:
if frame["id"] == frame_id:
if request.narration is not None:
frame["narration"] = request.narration
if request.image_prompt is not None:
frame["image_prompt"] = request.image_prompt
updated_frame = frame
break
if not updated_frame:
raise HTTPException(status_code=404, detail=f"Frame {frame_id} not found")
_storyboard_cache[storyboard_id] = storyboard
logger.info(f"Updated frame {frame_id} content")
return UpdateFrameResponse(
id=frame_id,
narration=updated_frame["narration"],
image_prompt=updated_frame.get("image_prompt"),
updated=True
)
@router.post(
"/storyboard/{storyboard_id}/frames/{frame_id}/regenerate-image",
response_model=RegenerateImageResponse
)
async def regenerate_frame_image(
storyboard_id: str = Path(..., description="Storyboard/task ID"),
frame_id: str = Path(..., description="Frame ID"),
request: RegenerateImageRequest = None
):
"""
Regenerate image for a frame
Uses the frame's image_prompt (or override) to generate a new image.
Requires ComfyUI service to be running.
"""
if storyboard_id not in _storyboard_cache:
raise HTTPException(status_code=404, detail=f"Storyboard {storyboard_id} not found")
storyboard = _storyboard_cache[storyboard_id]
frames = storyboard["frames"]
# Find frame
target_frame = None
frame_index = 0
for i, frame in enumerate(frames):
if frame["id"] == frame_id:
target_frame = frame
frame_index = i
break
if not target_frame:
raise HTTPException(status_code=404, detail=f"Frame {frame_id} not found")
# Get prompt to use
prompt = request.image_prompt if request and request.image_prompt else target_frame.get("image_prompt", "")
if not prompt:
raise HTTPException(status_code=400, detail="No image prompt available")
try:
# Import and use PixelleVideo core for image generation
from api.dependencies import get_pixelle_video
from pixelle_video.models.storyboard import StoryboardFrame, StoryboardConfig
pixelle_video = get_pixelle_video()
# Generate image using ComfyKit
result = await pixelle_video.comfy(
workflow="image_gen",
prompt=prompt,
task_id=storyboard_id,
)
if result and result.get("images"):
# Download and save image
image_url = result["images"][0]
import aiohttp
import os
output_dir = f"output/{storyboard_id}"
os.makedirs(output_dir, exist_ok=True)
image_path = f"{output_dir}/frame_{frame_index}_regenerated.png"
async with aiohttp.ClientSession() as session:
async with session.get(image_url) as resp:
if resp.status == 200:
with open(image_path, 'wb') as f:
f.write(await resp.read())
# Update frame
target_frame["image_path"] = _path_to_url(image_path)
_storyboard_cache[storyboard_id] = storyboard
logger.info(f"Regenerated image for frame {frame_id}")
return RegenerateImageResponse(
image_path=target_frame["image_path"],
success=True
)
else:
raise HTTPException(status_code=500, detail="Image generation failed")
except ImportError as e:
logger.error(f"Failed to import dependencies: {e}")
raise HTTPException(status_code=500, detail="Image generation service not available")
except Exception as e:
logger.error(f"Image regeneration failed: {e}")
raise HTTPException(status_code=500, detail=str(e))
@router.post(
"/storyboard/{storyboard_id}/frames/{frame_id}/regenerate-audio",
response_model=RegenerateAudioResponse
)
async def regenerate_frame_audio(
storyboard_id: str = Path(..., description="Storyboard/task ID"),
frame_id: str = Path(..., description="Frame ID"),
request: RegenerateAudioRequest = None
):
"""
Regenerate audio for a frame
Uses the frame's narration (or override) to generate new audio via TTS.
"""
if storyboard_id not in _storyboard_cache:
raise HTTPException(status_code=404, detail=f"Storyboard {storyboard_id} not found")
storyboard = _storyboard_cache[storyboard_id]
frames = storyboard["frames"]
# Find frame
target_frame = None
frame_index = 0
for i, frame in enumerate(frames):
if frame["id"] == frame_id:
target_frame = frame
frame_index = i
break
if not target_frame:
raise HTTPException(status_code=404, detail=f"Frame {frame_id} not found")
# Get narration to use
narration = request.narration if request and request.narration else target_frame.get("narration", "")
if not narration:
raise HTTPException(status_code=400, detail="No narration text available")
try:
from api.dependencies import get_pixelle_video
import os
pixelle_video = get_pixelle_video()
# Create output path
output_dir = f"output/{storyboard_id}"
os.makedirs(output_dir, exist_ok=True)
audio_path = f"{output_dir}/frame_{frame_index}_audio_regenerated.mp3"
# Generate audio using TTS service
voice = request.voice if request and request.voice else None
result_path = await pixelle_video.tts(
text=narration,
voice=voice,
output_path=audio_path
)
# Get audio duration
from mutagen.mp3 import MP3
try:
audio = MP3(result_path)
duration = audio.info.length
except:
duration = 3.0 # Default duration
# Update frame
target_frame["audio_path"] = _path_to_url(result_path)
target_frame["duration"] = duration
# Recalculate total duration
storyboard["total_duration"] = sum(f.get("duration", 3.0) for f in frames)
_storyboard_cache[storyboard_id] = storyboard
logger.info(f"Regenerated audio for frame {frame_id}, duration: {duration}s")
return RegenerateAudioResponse(
audio_path=target_frame["audio_path"],
duration=duration,
success=True
)
except ImportError as e:
logger.error(f"Failed to import dependencies: {e}")
raise HTTPException(status_code=500, detail="TTS service not available")
except Exception as e:
logger.error(f"Audio regeneration failed: {e}")
raise HTTPException(status_code=500, detail=str(e))

427
api/routers/publish.py Normal file
View File

@@ -0,0 +1,427 @@
# Copyright (C) 2025 AIDC-AI
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
# http://www.apache.org/licenses/LICENSE-2.0
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
"""
Publish API router for multi-platform video distribution.
Endpoints:
- POST /publish/export - Format conversion and export
- POST /publish/bilibili - Publish to Bilibili (TODO)
- POST /publish/youtube - Publish to YouTube (TODO)
- GET /publish/tasks/{id} - Get task status
"""
import uuid
from datetime import datetime
from fastapi import APIRouter, HTTPException, Path, BackgroundTasks
from loguru import logger
from api.schemas.publish import (
PublishRequest,
PublishResultSchema,
PublishTaskSchema,
PublishStatusEnum,
PlatformRequirementsSchema,
VideoMetadataSchema,
)
from pixelle_video.services.publishing import (
VideoMetadata,
PublishStatus,
PublishTask,
Platform,
)
from pixelle_video.services.publishing.export_publisher import ExportPublisher
router = APIRouter(prefix="/publish", tags=["Publish"])
# In-memory task storage (use Redis in production)
_publish_tasks: dict = {}
# Publisher instances
_export_publisher = ExportPublisher()
@router.post("/export", response_model=PublishResultSchema)
async def export_video(
request: PublishRequest,
background_tasks: BackgroundTasks
):
"""
Convert video to platform-optimized format and export.
Optimizes for:
- Portrait 9:16 aspect ratio (1080x1920)
- H.264 codec
- ≤128MB file size
For manual upload to Douyin/Kuaishou.
"""
# Create task
task_id = str(uuid.uuid4())[:8]
metadata = VideoMetadata(
title=request.metadata.title,
description=request.metadata.description,
tags=request.metadata.tags,
category=request.metadata.category,
cover_path=request.metadata.cover_path,
privacy=request.metadata.privacy,
platform_options=request.metadata.platform_options,
)
task = PublishTask(
id=task_id,
video_path=request.video_path,
platform=Platform.EXPORT,
metadata=metadata,
status=PublishStatus.PENDING,
)
_publish_tasks[task_id] = task
logger.info(f"📤 Starting export task {task_id} for: {metadata.title}")
# Execute synchronously for now (can be moved to background)
result = await _export_publisher.publish(
request.video_path,
metadata,
progress_callback=lambda p, m: logger.info(f"Export {task_id}: {p:.0%} - {m}")
)
# Update task
task.status = PublishStatus(result.status.value)
task.result = result
task.updated_at = datetime.now()
return PublishResultSchema(
success=result.success,
platform="export",
status=PublishStatusEnum(result.status.value),
export_path=result.export_path,
error_message=result.error_message,
)
@router.get("/tasks/{task_id}", response_model=PublishTaskSchema)
async def get_publish_task(task_id: str = Path(..., description="Task ID")):
"""Get publishing task status"""
if task_id not in _publish_tasks:
raise HTTPException(status_code=404, detail=f"Task {task_id} not found")
task = _publish_tasks[task_id]
return PublishTaskSchema(
id=task.id,
platform=task.platform.value,
status=PublishStatusEnum(task.status.value),
result=PublishResultSchema(
success=task.result.success,
platform=task.result.platform.value,
status=PublishStatusEnum(task.result.status.value),
export_path=task.result.export_path,
error_message=task.result.error_message,
) if task.result else None,
created_at=task.created_at.isoformat(),
updated_at=task.updated_at.isoformat() if task.updated_at else None,
)
@router.get("/requirements/{platform}", response_model=PlatformRequirementsSchema)
async def get_platform_requirements(platform: str = Path(..., description="Platform name")):
"""Get platform-specific requirements"""
requirements = {
"export": {
"max_file_size_mb": 128,
"max_duration_seconds": 900,
"supported_formats": ["mp4"],
"recommended_resolution": (1080, 1920),
"recommended_codec": "h264",
},
"bilibili": {
"max_file_size_mb": 4096,
"max_duration_seconds": 14400, # 4 hours
"supported_formats": ["mp4", "flv", "webm"],
"recommended_resolution": (1920, 1080), # Landscape
"recommended_codec": "h264",
},
"youtube": {
"max_file_size_mb": 256000, # 256GB
"max_duration_seconds": 43200, # 12 hours
"supported_formats": ["mp4", "mov", "avi", "webm"],
"recommended_resolution": (1920, 1080),
"recommended_codec": "h264",
},
}
if platform not in requirements:
raise HTTPException(status_code=404, detail=f"Platform {platform} not supported")
return requirements[platform]
from pixelle_video.services.publishing.bilibili_publisher import BilibiliPublisher
# Bilibili publisher instance
_bilibili_publisher = BilibiliPublisher()
@router.post("/bilibili", response_model=PublishResultSchema)
async def publish_to_bilibili(request: PublishRequest):
"""
Publish video to Bilibili.
Requires environment variables:
- BILIBILI_ACCESS_TOKEN or
- BILIBILI_SESSDATA + BILIBILI_BILI_JCT
"""
if not await _bilibili_publisher.validate_credentials():
raise HTTPException(
status_code=400,
detail="B站凭证未配置。请设置 BILIBILI_ACCESS_TOKEN 或 BILIBILI_SESSDATA 环境变量"
)
metadata = VideoMetadata(
title=request.metadata.title,
description=request.metadata.description,
tags=request.metadata.tags,
category=request.metadata.category,
cover_path=request.metadata.cover_path,
privacy=request.metadata.privacy,
platform_options=request.metadata.platform_options,
)
logger.info(f"📤 Starting Bilibili upload: {metadata.title}")
result = await _bilibili_publisher.publish(
request.video_path,
metadata,
progress_callback=lambda p, m: logger.info(f"Bilibili: {p:.0%} - {m}")
)
return PublishResultSchema(
success=result.success,
platform="bilibili",
status=PublishStatusEnum(result.status.value),
video_url=result.video_url,
platform_video_id=result.platform_video_id,
error_message=result.error_message,
)
from pixelle_video.services.publishing.youtube_publisher import YouTubePublisher
# YouTube publisher instance
_youtube_publisher = YouTubePublisher()
@router.post("/youtube", response_model=PublishResultSchema)
async def publish_to_youtube(request: PublishRequest):
"""
Publish video to YouTube.
Requires:
- config/youtube_client_secrets.json (OAuth 2.0 credentials)
- First-time auth will open browser for authorization
"""
if not await _youtube_publisher.validate_credentials():
raise HTTPException(
status_code=400,
detail="YouTube 凭证未配置。请添加 config/youtube_client_secrets.json"
)
metadata = VideoMetadata(
title=request.metadata.title,
description=request.metadata.description,
tags=request.metadata.tags,
category=request.metadata.category,
cover_path=request.metadata.cover_path,
privacy=request.metadata.privacy,
platform_options=request.metadata.platform_options,
)
logger.info(f"📤 Starting YouTube upload: {metadata.title}")
result = await _youtube_publisher.publish(
request.video_path,
metadata,
progress_callback=lambda p, m: logger.info(f"YouTube: {p:.0%} - {m}")
)
return PublishResultSchema(
success=result.success,
platform="youtube",
status=PublishStatusEnum(result.status.value),
video_url=result.video_url,
platform_video_id=result.platform_video_id,
error_message=result.error_message,
)
# ============================================================
# Async Task Queue Endpoints
# ============================================================
from pixelle_video.services.publishing.task_manager import get_publish_manager, TaskPriority
from pydantic import BaseModel
from typing import List
class AsyncPublishRequest(BaseModel):
"""Request for async publishing"""
video_path: str
platform: str # export, bilibili, youtube
metadata: VideoMetadataSchema
priority: str = "normal" # low, normal, high
class QueuedTaskSchema(BaseModel):
"""Schema for queued task"""
id: str
platform: str
status: str
progress: float
progress_message: str
retries: int
created_at: str
started_at: str = None
completed_at: str = None
class QueueStatusSchema(BaseModel):
"""Queue status overview"""
pending: int
active: int
completed: int
failed: int
workers: int
@router.post("/async", response_model=dict)
async def publish_async(request: AsyncPublishRequest):
"""
Submit a publish task to the background queue.
Returns immediately with task ID for tracking.
"""
manager = get_publish_manager()
# Ensure manager is running
if not manager._running:
await manager.start()
# Map platform string to enum
platform_map = {
"export": Platform.EXPORT,
"bilibili": Platform.BILIBILI,
"youtube": Platform.YOUTUBE,
}
platform = platform_map.get(request.platform.lower())
if not platform:
raise HTTPException(status_code=400, detail=f"Invalid platform: {request.platform}")
# Map priority
priority_map = {
"low": TaskPriority.LOW,
"normal": TaskPriority.NORMAL,
"high": TaskPriority.HIGH,
}
priority = priority_map.get(request.priority.lower(), TaskPriority.NORMAL)
metadata = VideoMetadata(
title=request.metadata.title,
description=request.metadata.description,
tags=request.metadata.tags,
category=request.metadata.category,
cover_path=request.metadata.cover_path,
privacy=request.metadata.privacy,
platform_options=request.metadata.platform_options,
)
task_id = await manager.enqueue(
video_path=request.video_path,
platform=platform,
metadata=metadata,
priority=priority,
)
return {
"task_id": task_id,
"status": "queued",
"message": f"Task queued for {request.platform}",
}
@router.get("/queue/status", response_model=QueueStatusSchema)
async def get_queue_status():
"""Get queue status overview."""
manager = get_publish_manager()
all_tasks = manager.get_all_tasks()
pending = sum(1 for t in all_tasks if t.task.status == PublishStatus.PENDING)
active = len(manager.get_active_tasks())
completed = sum(1 for t in all_tasks if t.task.status == PublishStatus.PUBLISHED)
failed = sum(1 for t in all_tasks if t.task.status == PublishStatus.FAILED)
return QueueStatusSchema(
pending=pending,
active=active,
completed=completed,
failed=failed,
workers=manager.max_workers,
)
@router.get("/queue/tasks", response_model=List[QueuedTaskSchema])
async def list_queued_tasks():
"""List all tasks in queue."""
manager = get_publish_manager()
result = []
for qt in manager.get_all_tasks():
result.append(QueuedTaskSchema(
id=qt.task.id,
platform=qt.task.platform.value,
status=qt.task.status.value,
progress=qt.progress,
progress_message=qt.progress_message,
retries=qt.retries,
created_at=qt.created_at.isoformat(),
started_at=qt.started_at.isoformat() if qt.started_at else None,
completed_at=qt.completed_at.isoformat() if qt.completed_at else None,
))
return result
@router.get("/queue/tasks/{task_id}", response_model=QueuedTaskSchema)
async def get_queued_task(task_id: str = Path(..., description="Task ID")):
"""Get specific queued task status."""
manager = get_publish_manager()
qt = manager.get_task(task_id)
if not qt:
raise HTTPException(status_code=404, detail=f"Task {task_id} not found")
return QueuedTaskSchema(
id=qt.task.id,
platform=qt.task.platform.value,
status=qt.task.status.value,
progress=qt.progress,
progress_message=qt.progress_message,
retries=qt.retries,
created_at=qt.created_at.isoformat(),
started_at=qt.started_at.isoformat() if qt.started_at else None,
completed_at=qt.completed_at.isoformat() if qt.completed_at else None,
)

View File

@@ -239,8 +239,13 @@ async def generate_video_async(
"prompt_prefix": request_body.prompt_prefix,
"bgm_path": request_body.bgm_path,
"bgm_volume": request_body.bgm_volume,
# Progress callback can be added here if needed
# "progress_callback": lambda event: task_manager.update_progress(...)
# Progress callback support
"progress_callback": lambda event: task_manager.update_progress(
task_id=task.task_id,
current=int(event.progress * 100),
total=100,
message=f"{event.event_type}" + (f" - {event.action}" if event.action else "")
)
}
# Add TTS workflow if specified
@@ -268,10 +273,31 @@ async def generate_video_async(
# Convert path to URL
video_url = path_to_url(request, result.video_path)
# Convert storyboard to dict for serialization
storyboard_data = {
"title": result.storyboard.title,
"total_duration": result.storyboard.total_duration,
"final_video_path": result.storyboard.final_video_path,
"created_at": result.storyboard.created_at.isoformat() if result.storyboard.created_at else None,
"frames": [
{
"index": f.index,
"narration": f.narration,
"image_prompt": f.image_prompt,
"audio_path": f.audio_path,
"image_path": f.image_path,
"video_segment_path": f.video_segment_path,
"duration": f.duration,
}
for f in result.storyboard.frames
]
}
return {
"video_url": video_url,
"duration": result.duration,
"file_size": file_size
"file_size": file_size,
"storyboard": storyboard_data
}
# Start execution

110
api/schemas/editor.py Normal file
View File

@@ -0,0 +1,110 @@
# Copyright (C) 2025 AIDC-AI
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
# http://www.apache.org/licenses/LICENSE-2.0
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
"""
Editor API schemas for timeline editor
"""
from pydantic import BaseModel, Field
from typing import List, Optional
from datetime import datetime
class StoryboardFrameSchema(BaseModel):
"""Schema for a single storyboard frame"""
id: str
index: int
order: int
narration: str
image_prompt: Optional[str] = None
image_path: Optional[str] = None
audio_path: Optional[str] = None
video_segment_path: Optional[str] = None
duration: float = 0.0
class Config:
from_attributes = True
class StoryboardSchema(BaseModel):
"""Schema for complete storyboard"""
id: str
title: str
frames: List[StoryboardFrameSchema]
total_duration: float
final_video_path: Optional[str] = None
created_at: Optional[datetime] = None
class Config:
from_attributes = True
class ReorderFramesRequest(BaseModel):
"""Request to reorder frames"""
order: List[str] = Field(..., description="List of frame IDs in new order")
class UpdateDurationRequest(BaseModel):
"""Request to update frame duration"""
duration: float = Field(..., ge=0.1, le=60.0, description="New duration in seconds")
class PreviewRequest(BaseModel):
"""Request to generate preview"""
start_frame: int = Field(0, ge=0, description="Start frame index")
end_frame: Optional[int] = Field(None, description="End frame index (None = to end)")
class PreviewResponse(BaseModel):
"""Response with preview video path"""
preview_path: str
duration: float
frames_count: int
class UpdateFrameRequest(BaseModel):
"""Request to update frame content"""
narration: Optional[str] = Field(None, description="Updated narration text")
image_prompt: Optional[str] = Field(None, description="Updated image generation prompt")
class UpdateFrameResponse(BaseModel):
"""Response after updating frame"""
id: str
narration: str
image_prompt: Optional[str]
updated: bool = True
class RegenerateImageRequest(BaseModel):
"""Request to regenerate frame image"""
image_prompt: Optional[str] = Field(None, description="Override prompt for regeneration")
class RegenerateImageResponse(BaseModel):
"""Response after regenerating image"""
image_path: str
success: bool = True
class RegenerateAudioRequest(BaseModel):
"""Request to regenerate frame audio"""
narration: Optional[str] = Field(None, description="Override narration for regeneration")
voice: Optional[str] = Field(None, description="Voice to use for TTS")
class RegenerateAudioResponse(BaseModel):
"""Response after regenerating audio"""
audio_path: str
duration: float
success: bool = True

81
api/schemas/publish.py Normal file
View File

@@ -0,0 +1,81 @@
# Copyright (C) 2025 AIDC-AI
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
# http://www.apache.org/licenses/LICENSE-2.0
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
"""
Publish API schemas
"""
from pydantic import BaseModel, Field
from typing import List, Optional, Dict, Any
from enum import Enum
class PlatformEnum(str, Enum):
export = "export"
bilibili = "bilibili"
youtube = "youtube"
class PublishStatusEnum(str, Enum):
pending = "pending"
converting = "converting"
uploading = "uploading"
processing = "processing"
published = "published"
failed = "failed"
class VideoMetadataSchema(BaseModel):
"""Video metadata for publishing"""
title: str = Field(..., min_length=1, max_length=100)
description: str = Field("", max_length=5000)
tags: List[str] = Field(default_factory=list)
category: Optional[str] = None
cover_path: Optional[str] = None
privacy: str = Field("public", pattern="^(public|private|unlisted)$")
platform_options: Dict[str, Any] = Field(default_factory=dict)
class PublishRequest(BaseModel):
"""Request to publish a video"""
video_path: str = Field(..., description="Path to the video file")
metadata: VideoMetadataSchema
class PublishResultSchema(BaseModel):
"""Result of a publishing operation"""
success: bool
platform: str
status: PublishStatusEnum
video_url: Optional[str] = None
platform_video_id: Optional[str] = None
error_message: Optional[str] = None
export_path: Optional[str] = None
class PublishTaskSchema(BaseModel):
"""A publishing task"""
id: str
platform: str
status: PublishStatusEnum
result: Optional[PublishResultSchema] = None
created_at: str
updated_at: Optional[str] = None
class PlatformRequirementsSchema(BaseModel):
"""Platform requirements"""
max_file_size_mb: int
max_duration_seconds: Optional[int] = None
supported_formats: List[str] = []
recommended_resolution: tuple = (1080, 1920)
recommended_codec: str = "h264"

View File

@@ -186,7 +186,7 @@ class TaskManager:
message: str = ""
):
"""
Update task progress
Update task progress and add to logs
Args:
task_id: Task ID
@@ -205,6 +205,17 @@ class TaskManager:
percentage=percentage,
message=message
)
# Add to logs if message is new
if message:
# Check last log to avoid duplicates
if not task.logs or task.logs[-1].get("message") != message:
task.logs.append({
"timestamp": datetime.now().isoformat(),
"message": message,
"percentage": round(percentage, 1)
})
logger.debug(f"Task {task_id} log: {message} ({percentage:.1f}%)")
def cancel_task(self, task_id: str) -> bool:
"""

View File

@@ -16,7 +16,7 @@ Task data models
from datetime import datetime
from enum import Enum
from typing import Any, Optional
from typing import Any, Optional, List
from pydantic import BaseModel, Field
@@ -51,10 +51,13 @@ class Task(BaseModel):
# Progress tracking
progress: Optional[TaskProgress] = None
# Result
# Results and Errors
result: Optional[Any] = None
error: Optional[str] = None
# Event logs/History
logs: List[dict] = Field(default_factory=list)
# Metadata
created_at: datetime = Field(default_factory=datetime.now)
started_at: Optional[datetime] = None