JianYing MCP

# -*- coding: utf-8 -*- """ Author: jian wei File Name: effect_manager.py """ import logging from typing import List, Any, Optional, Dict import difflib from pyJianYingDraft import VideoSceneEffectType, TextIntro, TextOutro, TextLoopAnim, IntroType, OutroType, GroupAnimationType, VideoCharacterEffectType from pyJianYingDraft.metadata.audio_effect_meta import ToneEffectType, AudioSceneEffectType, SpeechToSongType from pyJianYingDraft.metadata.filter_meta import FilterType from pyJianYingDraft.metadata.mask_meta import MaskType from pyJianYingDraft.metadata.transition_meta import TransitionType from pyJianYingDraft.metadata.font_meta import FontType class JianYingResourceManager: """ 剪映资源管理器 提供获取各种特效、字体、滤镜等资源的统一接口 """ # 特效类型映射 - 共15个特效类型 EFFECT_TYPE_MAPPING = { "VIDEO_SCENE": VideoSceneEffectType, # 画面特效类型 "ToneEffectType": ToneEffectType, # 音频音色 "AudioSceneEffectType": AudioSceneEffectType, # 音频场景 "filter_type": FilterType, # 滤镜 "SpeechToSongType": SpeechToSongType, # 语音转歌曲 "mask_type": MaskType, # 蒙版 "TransitionType": TransitionType, # 转场 "Font": FontType, # 字体 "TextIntro": TextIntro, # 文字入场 "TextOutro": TextOutro, # 文字出场 "TextLoopAnim": TextLoopAnim, # 文字循环动画 "GroupAnimationType": GroupAnimationType, # 组合动画 "VIDEO_CHARACTER": VideoCharacterEffectType, # 视频人物特效 "IntroType": IntroType, # 视频/图片入场动画类型 "OutroType": OutroType # 视频/图片出场动画类型 } DESCRIPTIONS = { "VIDEO_SCENE": "视频画面特效，包含各种视觉效果和场景特效", "ToneEffectType": "音频音色特效，用于改变声音的音调和音色", "AudioSceneEffectType": "音频场景特效，提供各种环境音效和声音处理", "filter_type": "滤镜特效，用于调整画面色彩和风格", "SpeechToSongType": "语音转歌曲特效", "mask_type": "蒙版特效", "TransitionType": "转场特效", "Font": "字体特效", "TextIntro": "文字入场特效", "TextOutro": "文字出场特效", "TextLoopAnim": "文字循环动画特效", "GroupAnimationType": "组合动画特效", "VIDEO_CHARACTER": "视频人物特效", "IntroType": "视频/图片入场动画类型", "OutroType": "视频/图片出场动画类型" } def __init__(self): """初始化资源管理器""" pass def get_all_types(self) -> Dict[str, str]: """ 获取所有的资源类型列表 Returns: Dict[str, str]: 包含所有资源类型及其描述的字典 """ return self.DESCRIPTIONS.copy() def find_by_type(self, effect_type: str, is_vip: Optional[bool] = None, limit: int = None, keyword: str = None) -> List[Dict[str, Any]]: """ 根据类型获取特效资源 Args: effect_type (str): 特效类型，支持的类型见 EFFECT_TYPE_MAPPING is_vip (Optional[bool]): 是否只获取VIP资源，None表示获取所有 limit(int)：返回数量，默认为None,即全部返回，有的特效数量比较多，建议加上 keyword (str): 模糊匹配关键词，用于搜索特效名称，None表示不过滤 Returns: List[Dict[str, Any]]: 特效数据字典列表，每个字典包含原始特效对象的所有属性（除了md5） Raises: ValueError: 当特效类型参数无效时抛出异常 """ try: if not effect_type: raise ValueError("没有指定资源类型") if effect_type not in self.EFFECT_TYPE_MAPPING: raise ValueError(f"无效的特效类型。支持的类型: {list(self.EFFECT_TYPE_MAPPING.keys())}") enum_class = self.EFFECT_TYPE_MAPPING[effect_type] result = self._extract_effects_from_enum(enum_class, is_vip) # 模糊匹配过滤 if keyword: result = self._fuzzy_match_filter(result, keyword) if limit: result = result[:limit] return result except ValueError as ve: raise ve except Exception as e: raise Exception(f"获取特效数据失败: {str(e)}") def _extract_effects_from_enum(self, enum_class, is_vip: Optional[bool] = None) -> List[Dict[str, Any]]: """ 从枚举类中提取特效数据，返回原始特效对象的所有属性（除了md5） Args: enum_class: 特效枚举类 is_vip (Optional[bool]): VIP过滤条件 Returns: List[Dict[str, Any]]: 特效数据字典列表，包含原始对象的所有属性 """ result = [] for effect_enum_member in enum_class: effect_meta = effect_enum_member.value # 创建数据字典，包含所有属性（除了md5） effect_data = {} # 获取所有属性，排除私有属性和md5 for attr_name in dir(effect_meta): if attr_name.startswith('_') or attr_name == 'md5': continue try: attr_value = getattr(effect_meta, attr_name) # 跳过方法和函数 if callable(attr_value): continue # 特殊处理时长字段（微秒转秒） if attr_name == 'duration' and isinstance(attr_value, int) and attr_value > 1000000: effect_data[attr_name] = attr_value / 1000000.0 # 跳过params参数 elif attr_name == 'params': continue else: effect_data[attr_name] = attr_value except: # 如果获取属性失败，跳过 continue # VIP过滤 effect_is_vip = effect_data.get('is_vip', False) if is_vip is not None and effect_is_vip != is_vip: continue if effect_data: # 确保有数据才添加 result.append(effect_data) return result def _fuzzy_match_filter(self, effects: List[Dict[str, Any]], keyword: str) -> List[Dict[str, Any]]: """ 智能模糊匹配过滤，使用文本相似度算法 支持多种匹配模式： 1. 完全匹配 - 最高优先级 2. 开头匹配 - 高优先级 3. 包含匹配 - 中等优先级 4. 文本相似度匹配 - 使用difflib或rapidfuzz 5. 字符序列匹配 - 支持部分字符匹配 Args: effects: 特效列表 keyword: 搜索关键词 Returns: List[Dict[str, Any]]: 按相似度排序的匹配特效列表 """ if not keyword: return effects scored_results = [] for effect in effects: # 获取特效名称 effect_name = "" if 'name' in effect and effect['name']: effect_name = str(effect['name']) elif 'title' in effect and effect['title']: effect_name = str(effect['title']) if not effect_name: continue # 计算匹配得分 match_score = self._calculate_similarity_score(keyword, effect_name) # 只保留有一定相似度的结果 if match_score >= 30: # 相似度阈值30% effect_copy = effect.copy() effect_copy['_match_score'] = match_score scored_results.append(effect_copy) # 按匹配得分排序，得分高的在前 scored_results.sort(key=lambda x: x.get('_match_score', 0), reverse=True) # 移除临时的匹配得分字段 for effect in scored_results: if '_match_score' in effect: del effect['_match_score'] return scored_results def _calculate_similarity_score(self, keyword: str, effect_name: str) -> float: """计算文本相似度得分""" if keyword == effect_name: return 100.0 if keyword in effect_name: return 90.0 score = difflib.SequenceMatcher(None, keyword, effect_name).ratio() * 100 return score if score >= 30 else 0.0 if __name__ == "__main__": manager = JianYingResourceManager() results = manager.find_by_type("TextLoopAnim", keyword="色差故障", limit=3) for i in results: print(i)