"""Renderer manifest generation for retention-led AutoShortsBot videos.""" from __future__ import annotations import json from pathlib import Path from typing import Any from autoshorts.script_drafts import _slug RESOLUTION = {"width": 1080, "height": 1920} ASPECT_RATIO = "9:16" FPS = 30 ASSET_REQUIREMENT_HINTS = { "kinetic_text": "Generated by renderer from text and style instructions.", "screen_recording": "Needs captured or synthesized vertical screen recording; avoid fake unreadable UI text.", "card_animation": "Generated by renderer from concise text cards; animate one idea per beat.", "stock_broll": "Needs human/phone/context B-roll; avoid generic corporate stock feel.", "generated_background": "Optional abstract background only; must not become the main content.", "cinematic_ai_video": "Needs photorealistic vertical AI-video clip; no baked text, logos, or watermarks.", "mechanism_animation": "Needs renderer-generated mechanism/layer animation with visible movement every 1-2 seconds.", "kinetic_text_overlay": "Generated by renderer as large beat-synced caption overlays, not paragraph cards.", } def _seconds_to_ms(value: int | float) -> int: return round(float(value) * 1000) def _duration_ms(start: int | float, end: int | float) -> int: return max(500, _seconds_to_ms(end) - _seconds_to_ms(start)) def _placeholder(asset_type: str, video_id: str, scene: int) -> str: return f"placeholder://{video_id}/scene-{scene:02d}/{asset_type}" def _visual_style(scene: dict[str, Any]) -> dict[str, Any]: base = { "safe_area_pct": 8, "font": "bold_sans", "caption_case": "sentence", "motion": "push_zoom" if scene["scene"] == 1 else "cut_or_slide", "contrast": "high", } if scene["asset_type"] == "card_animation": base.update({"layout": "minimal_cards", "max_words_on_screen": 18}) elif scene["asset_type"] == "screen_recording": base.update({"layout": "screen_plus_caption", "blur_sensitive_ui": True}) elif scene["asset_type"] == "stock_broll": base.update({"layout": "broll_under_caption", "pace": "fast_cut"}) elif scene["asset_type"] == "cinematic_ai_video": base.update({"layout": "cinematic_under_caption", "no_baked_text": True, "pace": "fast_cut"}) elif scene["asset_type"] == "mechanism_animation": base.update({"layout": "xray_layer_reveal", "max_words_on_screen": 10, "layered_motion": True}) elif scene["asset_type"] == "kinetic_text_overlay": base.update({"layout": "kinetic_full_frame", "max_words_on_screen": 12, "beat_synced": True}) return base def _visual_timeline_items(plan: dict[str, Any]) -> list[dict[str, Any]]: video_id = plan["id"] items: list[dict[str, Any]] = [] for scene in plan["scenes"]: items.append( { "id": f"scene_{scene['scene']:02d}", "track": "visual", "type": scene["asset_type"], "asset_ref": _placeholder(scene["asset_type"], video_id, scene["scene"]), "start_ms": _seconds_to_ms(scene["start"]), "duration_ms": _duration_ms(scene["start"], scene["end"]), "text": scene["on_screen_text"], "visual_direction": scene["visual_direction"], "visual_motion": scene.get("visual_motion", "cut_or_slide"), "sound_cue": scene.get("sound_cue", "none"), "retention_goal": scene["retention_goal"], "style": _visual_style(scene), } ) return items def _caption_timeline_items(plan: dict[str, Any]) -> list[dict[str, Any]]: beats = plan.get("caption_beats", []) total_ms = int(plan["estimated_seconds"]) * 1000 items: list[dict[str, Any]] = [] for index, beat in enumerate(beats): start_ms = _seconds_to_ms(beat["start"]) next_start_ms = _seconds_to_ms(beats[index + 1]["start"]) if index + 1 < len(beats) else total_ms items.append( { "id": f"caption_{index + 1:02d}", "track": "captions", "type": "kinetic_caption", "start_ms": start_ms, "duration_ms": max(750, next_start_ms - start_ms), "text": beat["text"], "style": { "safe_area_pct": 8, "position": "lower_third", "max_chars_per_line": 28, "highlight_keywords": True, }, } ) return items def _asset_requirements(plan: dict[str, Any]) -> dict[str, list[dict[str, Any]]]: requirements: dict[str, list[dict[str, Any]]] = {key: [] for key in ASSET_REQUIREMENT_HINTS} for scene in plan["scenes"]: asset_type = scene["asset_type"] requirements.setdefault(asset_type, []).append( { "scene": scene["scene"], "asset_ref": _placeholder(asset_type, plan["id"], scene["scene"]), "brief": scene["visual_direction"], "hint": ASSET_REQUIREMENT_HINTS.get(asset_type, "Provide asset matching the visual direction."), } ) return requirements def build_render_manifest(plan: dict[str, Any]) -> dict[str, Any]: """Convert a human shot plan into a structured renderer contract.""" timeline = _visual_timeline_items(plan) + _caption_timeline_items(plan) track_order = {"visual": 0, "captions": 1} timeline.sort(key=lambda item: (item["start_ms"], track_order.get(item["track"], 99), item["id"])) return { "schema_version": "render_manifest.v1", "video": { "id": plan["id"], "title": plan["title"], "pillar": plan["pillar"], "duration_ms": int(plan["estimated_seconds"]) * 1000, "aspect_ratio": ASPECT_RATIO, "resolution": RESOLUTION, "fps": FPS, }, "timeline": timeline, "asset_requirements": _asset_requirements(plan), "render_constraints": { "avoid_static_slideshow": True, "pattern_interrupt_every_seconds": 3, "visual_change_every_seconds": plan.get("final_render_intent", {}).get("visual_change_every_seconds", 3), "first_frame_hook_required": True, "max_words_per_visual_card": 18, "max_static_cards": plan.get("final_render_intent", {}).get("static_card_limit", 1), "human_approval_required_before_publish": True, }, "production_notes": plan.get("production_notes", []), } def _manifest_filename(plan: dict[str, Any]) -> str: title_slug = _slug(plan["title"]) video_id = plan["id"] if video_id.endswith(title_slug): return f"{video_id}.json" return f"{video_id}-{title_slug}.json" def write_render_manifests(plans: list[dict[str, Any]], output_dir: Path) -> list[Path]: """Write render manifests as pretty JSON files.""" output_dir.mkdir(parents=True, exist_ok=True) paths: list[Path] = [] for plan in plans: manifest = build_render_manifest(plan) path = output_dir / _manifest_filename(plan) path.write_text(json.dumps(manifest, indent=2, ensure_ascii=False) + "\n", encoding="utf-8") paths.append(path) return paths