递归叙事马赛克视觉生成器

recursive_narrative_mosaic 从 multiscale_narrative_systems 中选择 visual_output，其中 macro_subject = [subject] 且 source_context = [biographic_or_conceptual_context] 且 semantic_profile = semantic_decompose( [subject], dimensions = [ 'origins', 'turning_points', 'relationships', 'conflicts', 'achievements', 'legacy' ] ) 且 micro_scene_library = infer( distinct_narrative_vignettes from semantic_profile, count = [scene_count], prohibit_generic_filler = true ) 且 macro_value_map = infer( light_dark_structure from recognizable_form([subject]) ) 且 scene_distribution = ' 根据每个微场景的色调值及其与宏观主题对应区域的概念关系放置每个微场景 ' 且 density_rule = ' 较暗的宏观区域使用更密集或更复杂的场景；较亮的区域使用开放构图和负空间；不得使用任意颜色块来创建主题 ' 且 continuity_rule = ' 相邻场景应通过重复的物体、手势、建筑、景观或方向性运动连接 ' 且 scale_behavior = ' 单个故事在近距离可读；一个统一的主题在远处浮现 ' 且 visual_language = [illustration_system] 且 temporal_tone = [era_treatment] 且 palette = [palette] 按 macro_recognition desc, micro_story_specificity desc, semantic_scene_placement desc, multiscale_discovery desc, compositional_cohesion desc 排序 限制 1;