Files
AI-Video/pixelle_video/pipelines/standard.py

548 lines
24 KiB
Python

# 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.
"""
Standard Video Generation Pipeline
Standard workflow for generating short videos from topic or fixed script.
This is the default pipeline for general-purpose video generation.
"""
from datetime import datetime
from pathlib import Path
from typing import Optional, Callable, Literal, Dict, Any
from loguru import logger
from pixelle_video.pipelines.base import BasePipeline
from pixelle_video.models.progress import ProgressEvent
from pixelle_video.models.storyboard import (
Storyboard,
StoryboardFrame,
StoryboardConfig,
ContentMetadata,
VideoGenerationResult
)
from pixelle_video.utils.content_generators import (
generate_title,
generate_narrations_from_topic,
split_narration_script,
generate_image_prompts,
)
class StandardPipeline(BasePipeline):
"""
Standard video generation pipeline
Workflow:
1. Generate/determine title
2. Generate narrations (from topic or split fixed script)
3. Generate image prompts for each narration
4. For each frame:
- Generate audio (TTS)
- Generate image
- Compose frame with template
- Create video segment
5. Concatenate all segments
6. Add BGM (optional)
Supports two modes:
- "generate": LLM generates narrations from topic
- "fixed": Use provided script as-is (each line = one narration)
"""
async def __call__(
self,
# === Input ===
text: str,
# === Processing Mode ===
mode: Literal["generate", "fixed"] = "generate",
# === Optional Title ===
title: Optional[str] = None,
# === Basic Config ===
n_scenes: int = 5, # Only used in generate mode; ignored in fixed mode
# === TTS Parameters ===
tts_inference_mode: Optional[str] = None, # "local" or "comfyui"
tts_voice: Optional[str] = None, # For local mode: Edge TTS voice ID
tts_speed: Optional[float] = None, # Speed multiplier (0.5-2.0)
tts_workflow: Optional[str] = None, # For ComfyUI mode: workflow path
ref_audio: Optional[str] = None, # For ComfyUI mode: reference audio
# Deprecated (kept for backward compatibility)
voice_id: Optional[str] = None,
output_path: Optional[str] = None,
# === LLM Parameters ===
min_narration_words: int = 5,
max_narration_words: int = 20,
min_image_prompt_words: int = 30,
max_image_prompt_words: int = 60,
# === Image Parameters ===
# Note: image_width and image_height are now auto-determined from template meta tags
image_workflow: Optional[str] = None,
# === Video Parameters ===
video_fps: int = 30,
# === Frame Template (determines video size) ===
frame_template: Optional[str] = None,
template_params: Optional[Dict[str, Any]] = None, # Custom template parameters
# === Image Style ===
prompt_prefix: Optional[str] = None,
# === BGM Parameters ===
bgm_path: Optional[str] = None,
bgm_volume: float = 0.2,
bgm_mode: Literal["once", "loop"] = "loop",
# === Advanced Options ===
content_metadata: Optional[ContentMetadata] = None,
progress_callback: Optional[Callable[[ProgressEvent], None]] = None,
) -> VideoGenerationResult:
"""
Generate short video from text input
Args:
text: Text input (required)
- For generate mode: topic/theme (e.g., "如何提高学习效率")
- For fixed mode: complete narration script (each line is a narration)
mode: Processing mode (default "generate")
- "generate": LLM generates narrations from topic, creates n_scenes
- "fixed": Use existing script as-is, each line becomes a narration
Note: In fixed mode, n_scenes is ignored (uses actual line count)
title: Video title (optional)
- If provided, use it as the video title
- If not provided:
* generate mode → use text as title
* fixed mode → LLM generates title from script
n_scenes: Number of storyboard scenes (default 5)
Only effective in generate mode; ignored in fixed mode
voice_id: TTS voice ID (default "[Chinese] zh-CN Yunjian")
tts_workflow: TTS workflow filename (e.g., "tts_edge.json", None = use default)
tts_speed: TTS speed multiplier (1.0 = normal, 1.2 = 20% faster, default 1.2)
ref_audio: Reference audio path for voice cloning (optional)
output_path: Output video path (auto-generated if None)
min_narration_words: Min narration length (generate mode only)
max_narration_words: Max narration length (generate mode only)
min_image_prompt_words: Min image prompt length
max_image_prompt_words: Max image prompt length
image_workflow: Image workflow filename (e.g., "image_flux.json", None = use default)
Note: Image/video size is now auto-determined from template meta tags
video_fps: Video frame rate (default 30)
frame_template: HTML template path with size (None = use default "1080x1920/default.html")
Format: "SIZExSIZE/template.html" (e.g., "1080x1920/default.html", "1920x1080/modern.html")
Video size is automatically determined from template path
prompt_prefix: Image prompt prefix (overrides config.yaml if provided)
e.g., "anime style, vibrant colors" or "" for no prefix
bgm_path: BGM path (filename like "default.mp3", custom path, or None)
bgm_volume: BGM volume 0.0-1.0 (default 0.2)
bgm_mode: BGM mode "once" or "loop" (default "loop")
content_metadata: Content metadata (optional, for display)
progress_callback: Progress callback function(ProgressEvent)
Returns:
VideoGenerationResult with video path and metadata
"""
# ========== Step 0: Process text and determine title ==========
logger.info(f"🚀 Starting StandardPipeline in '{mode}' mode")
logger.info(f" Text length: {len(text)} chars")
# Determine final title
if title:
final_title = title
logger.info(f" Title: '{title}' (user-specified)")
else:
self._report_progress(progress_callback, "generating_title", 0.01)
if mode == "generate":
final_title = await generate_title(self.llm, text, strategy="auto")
logger.info(f" Title: '{final_title}' (auto-generated)")
else: # fixed
final_title = await generate_title(self.llm, text, strategy="llm")
logger.info(f" Title: '{final_title}' (LLM-generated)")
# ========== Step 0.5: Create isolated task directory ==========
from pixelle_video.utils.os_util import (
create_task_output_dir,
get_task_final_video_path
)
task_dir, task_id = create_task_output_dir()
logger.info(f"📁 Task directory created: {task_dir}")
logger.info(f" Task ID: {task_id}")
# Determine final video path
user_specified_output = None
if output_path is None:
output_path = get_task_final_video_path(task_id)
else:
user_specified_output = output_path
output_path = get_task_final_video_path(task_id)
logger.info(f" Will copy final video to: {user_specified_output}")
# Determine TTS inference mode and parameters
# Priority: explicit params > backward compatibility > config defaults
if tts_inference_mode is None:
# Check if user provided ComfyUI-specific params
if tts_workflow is not None or ref_audio is not None:
tts_inference_mode = "comfyui"
# Check if user provided old voice_id param (backward compatibility)
elif voice_id is not None:
tts_inference_mode = "comfyui"
if tts_voice is None:
tts_voice = voice_id
else:
# Use config default
tts_config = self.core.config.get("comfyui", {}).get("tts", {})
tts_inference_mode = tts_config.get("inference_mode", "local")
# Set voice_id based on mode for StoryboardConfig
final_voice_id = None
if tts_inference_mode == "local":
final_voice_id = tts_voice or voice_id
else: # comfyui
final_voice_id = voice_id # For ComfyUI, might be None
# Determine frame template
# Priority: explicit param > config default > hardcoded default
if frame_template is None:
template_config = self.core.config.get("template", {})
frame_template = template_config.get("default_template", "1080x1920/default.html")
# Read media size from template meta tags
from pixelle_video.services.frame_html import HTMLFrameGenerator
from pixelle_video.utils.template_util import resolve_template_path
template_path = resolve_template_path(frame_template)
temp_generator = HTMLFrameGenerator(template_path)
image_width, image_height = temp_generator.get_media_size()
logger.info(f"📐 Media size from template: {image_width}x{image_height}")
# Create storyboard config
config = StoryboardConfig(
task_id=task_id,
n_storyboard=n_scenes,
min_narration_words=min_narration_words,
max_narration_words=max_narration_words,
min_image_prompt_words=min_image_prompt_words,
max_image_prompt_words=max_image_prompt_words,
video_fps=video_fps,
tts_inference_mode=tts_inference_mode,
voice_id=final_voice_id,
tts_workflow=tts_workflow,
tts_speed=tts_speed,
ref_audio=ref_audio,
image_width=image_width,
image_height=image_height,
image_workflow=image_workflow,
frame_template=frame_template,
template_params=template_params
)
# Create storyboard
storyboard = Storyboard(
title=final_title,
config=config,
content_metadata=content_metadata,
created_at=datetime.now()
)
# ========== Step 0.8: Check template requirements ==========
template_media_type = self._check_template_media_type(config.frame_template)
if template_media_type == "video":
logger.info(f"🎬 Template requires video generation")
elif template_media_type == "image":
logger.info(f"📸 Template requires image generation")
else: # static
logger.info(f"⚡ Static template - skipping media generation pipeline")
logger.info(f" 💡 Benefits: Faster generation + Lower cost + No ComfyUI dependency")
try:
# ========== Step 1: Generate/Split narrations ==========
if mode == "generate":
self._report_progress(progress_callback, "generating_narrations", 0.05)
narrations = await generate_narrations_from_topic(
self.llm,
topic=text,
n_scenes=n_scenes,
min_words=min_narration_words,
max_words=max_narration_words
)
logger.info(f"✅ Generated {len(narrations)} narrations")
else: # fixed
self._report_progress(progress_callback, "splitting_script", 0.05)
narrations = await split_narration_script(text)
logger.info(f"✅ Split script into {len(narrations)} segments (by lines)")
logger.info(f" Note: n_scenes={n_scenes} is ignored in fixed mode")
# ========== Step 2: Generate media prompts (conditional) ==========
if template_media_type == "video":
# Video template: generate video prompts
self._report_progress(progress_callback, "generating_video_prompts", 0.15)
from pixelle_video.utils.content_generators import generate_video_prompts
# Override prompt_prefix if provided
original_prefix = None
if prompt_prefix is not None:
image_config = self.core.config.get("comfyui", {}).get("image", {})
original_prefix = image_config.get("prompt_prefix")
image_config["prompt_prefix"] = prompt_prefix
logger.info(f"Using custom prompt_prefix: '{prompt_prefix}'")
try:
# Create progress callback wrapper for video prompt generation
def video_prompt_progress(completed: int, total: int, message: str):
batch_progress = completed / total if total > 0 else 0
overall_progress = 0.15 + (batch_progress * 0.15)
self._report_progress(
progress_callback,
"generating_video_prompts",
overall_progress,
extra_info=message
)
# Generate base video prompts
base_image_prompts = await generate_video_prompts(
self.llm,
narrations=narrations,
min_words=min_image_prompt_words,
max_words=max_image_prompt_words,
progress_callback=video_prompt_progress
)
# Apply prompt prefix
from pixelle_video.utils.prompt_helper import build_image_prompt
image_config = self.core.config.get("comfyui", {}).get("image", {})
prompt_prefix_to_use = prompt_prefix if prompt_prefix is not None else image_config.get("prompt_prefix", "")
image_prompts = []
for base_prompt in base_image_prompts:
final_prompt = build_image_prompt(base_prompt, prompt_prefix_to_use)
image_prompts.append(final_prompt)
finally:
# Restore original prompt_prefix
if original_prefix is not None:
image_config["prompt_prefix"] = original_prefix
logger.info(f"✅ Generated {len(image_prompts)} video prompts")
elif template_media_type == "image":
# Image template: generate image prompts
self._report_progress(progress_callback, "generating_image_prompts", 0.15)
# Override prompt_prefix if provided
original_prefix = None
if prompt_prefix is not None:
image_config = self.core.config.get("comfyui", {}).get("image", {})
original_prefix = image_config.get("prompt_prefix")
image_config["prompt_prefix"] = prompt_prefix
logger.info(f"Using custom prompt_prefix: '{prompt_prefix}'")
try:
# Create progress callback wrapper for image prompt generation
def image_prompt_progress(completed: int, total: int, message: str):
batch_progress = completed / total if total > 0 else 0
overall_progress = 0.15 + (batch_progress * 0.15)
self._report_progress(
progress_callback,
"generating_image_prompts",
overall_progress,
extra_info=message
)
# Generate base image prompts
base_image_prompts = await generate_image_prompts(
self.llm,
narrations=narrations,
min_words=min_image_prompt_words,
max_words=max_image_prompt_words,
progress_callback=image_prompt_progress
)
# Apply prompt prefix
from pixelle_video.utils.prompt_helper import build_image_prompt
image_config = self.core.config.get("comfyui", {}).get("image", {})
prompt_prefix_to_use = prompt_prefix if prompt_prefix is not None else image_config.get("prompt_prefix", "")
image_prompts = []
for base_prompt in base_image_prompts:
final_prompt = build_image_prompt(base_prompt, prompt_prefix_to_use)
image_prompts.append(final_prompt)
finally:
# Restore original prompt_prefix
if original_prefix is not None:
image_config["prompt_prefix"] = original_prefix
logger.info(f"✅ Generated {len(image_prompts)} image prompts")
else: # text
# Text-only template: skip media prompt generation
image_prompts = [None] * len(narrations)
self._report_progress(progress_callback, "preparing_frames", 0.15)
logger.info(f"⚡ Skipped media prompt generation (text-only template)")
logger.info(f" 💡 Savings: {len(narrations)} LLM calls + {len(narrations)} media generations")
# ========== Step 3: Create frames ==========
for i, (narration, image_prompt) in enumerate(zip(narrations, image_prompts)):
frame = StoryboardFrame(
index=i,
narration=narration,
image_prompt=image_prompt,
created_at=datetime.now()
)
storyboard.frames.append(frame)
# ========== Step 4: Process each frame ==========
for i, frame in enumerate(storyboard.frames):
base_progress = 0.2
frame_range = 0.6
per_frame_progress = frame_range / len(storyboard.frames)
# Create frame-specific progress callback
def frame_progress_callback(event: ProgressEvent):
overall_progress = base_progress + (per_frame_progress * i) + (per_frame_progress * event.progress)
if progress_callback:
adjusted_event = ProgressEvent(
event_type=event.event_type,
progress=overall_progress,
frame_current=event.frame_current,
frame_total=event.frame_total,
step=event.step,
action=event.action
)
progress_callback(adjusted_event)
# Report frame start
self._report_progress(
progress_callback,
"processing_frame",
base_progress + (per_frame_progress * i),
frame_current=i+1,
frame_total=len(storyboard.frames)
)
processed_frame = await self.core.frame_processor(
frame=frame,
storyboard=storyboard,
config=config,
total_frames=len(storyboard.frames),
progress_callback=frame_progress_callback
)
storyboard.total_duration += processed_frame.duration
logger.info(f"✅ Frame {i+1} completed ({processed_frame.duration:.2f}s)")
# ========== Step 5: Concatenate videos ==========
self._report_progress(progress_callback, "concatenating", 0.85)
segment_paths = [frame.video_segment_path for frame in storyboard.frames]
from pixelle_video.services.video import VideoService
video_service = VideoService()
final_video_path = video_service.concat_videos(
videos=segment_paths,
output=output_path,
bgm_path=bgm_path,
bgm_volume=bgm_volume,
bgm_mode=bgm_mode
)
storyboard.final_video_path = final_video_path
storyboard.completed_at = datetime.now()
# Copy to user-specified path if provided
if user_specified_output:
import shutil
Path(user_specified_output).parent.mkdir(parents=True, exist_ok=True)
shutil.copy2(final_video_path, user_specified_output)
logger.info(f"📹 Final video copied to: {user_specified_output}")
final_video_path = user_specified_output
storyboard.final_video_path = user_specified_output
logger.success(f"🎬 Video generation completed: {final_video_path}")
# ========== Step 6: Create result ==========
self._report_progress(progress_callback, "completed", 1.0)
video_path_obj = Path(final_video_path)
file_size = video_path_obj.stat().st_size
result = VideoGenerationResult(
video_path=final_video_path,
storyboard=storyboard,
duration=storyboard.total_duration,
file_size=file_size
)
logger.info(f"✅ Generated video: {final_video_path}")
logger.info(f" Duration: {storyboard.total_duration:.2f}s")
logger.info(f" Size: {file_size / (1024*1024):.2f} MB")
logger.info(f" Frames: {len(storyboard.frames)}")
return result
except Exception as e:
logger.error(f"❌ Video generation failed: {e}")
raise
def _check_template_media_type(self, frame_template: str) -> str:
"""
Check template media type requirement
This is checked at pipeline level to avoid unnecessary:
- LLM calls (generating media prompts)
- Media generation API calls
- ComfyUI dependency
Template naming convention:
- static_*.html: Static style template (returns "static")
- image_*.html: Image template (returns "image")
- video_*.html: Video template (returns "video")
Args:
frame_template: Template path (e.g., "1080x1920/image_default.html" or "1080x1920/video_default.html")
Returns:
"static", "image", or "video"
"""
from pixelle_video.utils.template_util import get_template_type
# Determine type by template filename prefix
template_name = Path(frame_template).name
template_type = get_template_type(template_name)
logger.debug(f"Template '{frame_template}' is {template_type} template")
return template_type