| 1 | #!/usr/bin/env python3 |
| 2 | """ |
| 3 | PPT Master - Narration Sync Tool |
| 4 | |
| 5 | Derive click-free narration timing from the canonical animation config and |
| 6 | page-local SRT cues, then merge those subtitles against timing values read |
| 7 | from the final narrated PPTX. |
| 8 | See workflows/stages/generate-audio.md for the owning stage. |
| 9 | |
| 10 | Usage: |
| 11 | python3 scripts/narration_sync.py fingerprint <project_path> |
| 12 | python3 scripts/narration_sync.py animations <project_path> |
| 13 | python3 scripts/narration_sync.py subtitles <project_path> --pptx <pptx> --force |
| 14 | python3 scripts/narration_sync.py subtitles <project_path> --pptx <pptx> \ |
| 15 | --video <powerpoint_video> --force |
| 16 | |
| 17 | Examples: |
| 18 | python3 scripts/narration_sync.py fingerprint projects/demo |
| 19 | python3 scripts/narration_sync.py animations projects/demo |
| 20 | python3 scripts/narration_sync.py subtitles projects/demo --pptx exports/demo.pptx --force |
| 21 | python3 scripts/narration_sync.py subtitles projects/demo \ |
| 22 | --pptx exports/demo.pptx --video exports/demo.mp4 --force |
| 23 | |
| 24 | Dependencies: |
| 25 | ffprobe for animation-window validation. Optional exported-video timeline |
| 26 | calibration additionally requires ffmpeg and numpy. |
| 27 | """ |
| 28 | |
| 29 | from __future__ import annotations |
| 30 | |
| 31 | import argparse |
| 32 | import copy |
| 33 | import hashlib |
| 34 | import json |
| 35 | import math |
| 36 | import os |
| 37 | import posixpath |
| 38 | import re |
| 39 | import shutil |
| 40 | import subprocess |
| 41 | import sys |
| 42 | import tempfile |
| 43 | import zipfile |
| 44 | from dataclasses import dataclass |
| 45 | from pathlib import Path |
| 46 | from typing import Any |
| 47 | from xml.etree import ElementTree as ET |
| 48 | |
| 49 | _SCRIPTS_DIR = Path(__file__).resolve().parent |
| 50 | if str(_SCRIPTS_DIR) not in sys.path: |
| 51 | sys.path.insert(0, str(_SCRIPTS_DIR)) |
| 52 | |
| 53 | from console_encoding import configure_utf8_stdio # noqa: E402 |
| 54 | from pptx_animations import ( # noqa: E402 |
| 55 | ANIMATION_TIMING_OPTION_FIELDS, |
| 56 | animation_seconds_to_milliseconds, |
| 57 | normalize_animation_effect, |
| 58 | normalize_animation_trigger, |
| 59 | ) |
| 60 | from pptx_transitions import ( # noqa: E402 |
| 61 | DEFAULT_TRANSITION_DURATION, |
| 62 | normalize_transition_effect_request, |
| 63 | read_slide_transition_xml, |
| 64 | validate_seconds, |
| 65 | ) |
| 66 | from svg_to_pptx.animation_config import ( # noqa: E402 |
| 67 | animation_group_effect_entries, |
| 68 | scan_project_targets, |
| 69 | scan_svg_targets, |
| 70 | validate_animation_config_errors, |
| 71 | validate_transition_config, |
| 72 | ) |
| 73 | from svg_to_pptx.pptx_package.narration import ( # noqa: E402 |
| 74 | DEFAULT_NARRATION_START_FLOOR, |
| 75 | NARRATION_EXTENSIONS, |
| 76 | narration_lead_in_seconds, |
| 77 | probe_audio_duration, |
| 78 | read_narration_start_delay_xml, |
| 79 | ) |
| 80 | |
| 81 | configure_utf8_stdio() |
| 82 | |
| 83 | |
| 84 | _PML_NS = "http://schemas.openxmlformats.org/presentationml/2006/main" |
| 85 | _REL_NS = "http://schemas.openxmlformats.org/package/2006/relationships" |
| 86 | _DOC_REL_NS = ( |
| 87 | "http://schemas.openxmlformats.org/officeDocument/2006/relationships" |
| 88 | ) |
| 89 | _TIMING_RE = re.compile( |
| 90 | r"^(?P<start>\d+:\d{2}:\d{2},\d{3})\s+-->\s+" |
| 91 | r"(?P<end>\d+:\d{2}:\d{2},\d{3})(?:\s+.*)?$" |
| 92 | ) |
| 93 | _ALIGNMENT_SAMPLE_RATE = 8000 |
| 94 | _ALIGNMENT_FINE_HZ = 1000 |
| 95 | _ALIGNMENT_COARSE_FACTOR = 10 |
| 96 | _ALIGNMENT_SEARCH_WINDOW_MS = 2000 |
| 97 | _ALIGNMENT_REFINE_WINDOW_MS = 100 |
| 98 | _ALIGNMENT_TEMPLATE_MAX_MS = 12_000 |
| 99 | _ALIGNMENT_MIN_CORRELATION = 0.85 |
| 100 | _ALIGNMENT_END_TOLERANCE_MS = 100 |
| 101 | |
| 102 | |
| 103 | @dataclass(frozen=True) |
| 104 | class SubtitleCue: |
| 105 | """One parsed SRT cue on a millisecond timeline.""" |
| 106 | |
| 107 | start_ms: int |
| 108 | end_ms: int |
| 109 | text: str |
| 110 | |
| 111 | |
| 112 | @dataclass(frozen=True) |
| 113 | class TimingPlanEntry: |
| 114 | """One ordered animation target and its optional SRT cue anchor.""" |
| 115 | |
| 116 | group_id: str |
| 117 | cue_number: int | None |
| 118 | |
| 119 | |
| 120 | @dataclass(frozen=True) |
| 121 | class AnimationGroupState: |
| 122 | """One effective canonical animation row before narration timing.""" |
| 123 | |
| 124 | group_id: str |
| 125 | effect_index: int | None |
| 126 | order: int |
| 127 | source_index: int |
| 128 | duration_ms: int |
| 129 | original_delay_ms: int |
| 130 | trigger: str |
| 131 | |
| 132 | |
| 133 | @dataclass(frozen=True) |
| 134 | class SlideAnimationSettings: |
| 135 | """Effective animation settings inherited by one slide.""" |
| 136 | |
| 137 | effect: str | None |
| 138 | duration_ms: int |
| 139 | stagger_ms: int |
| 140 | trigger: str |
| 141 | timing_options: dict[str, Any] |
| 142 | |
| 143 | |
| 144 | @dataclass(frozen=True) |
| 145 | class AnimationBuildResult: |
| 146 | """Summary of one narration animation derivation.""" |
| 147 | |
| 148 | slide_count: int |
| 149 | group_count: int |
| 150 | anchored_count: int |
| 151 | fallback_count: int |
| 152 | ignored_cue_count: int |
| 153 | svg_fallback_slide_count: int |
| 154 | timing_plan_written: bool |
| 155 | |
| 156 | |
| 157 | @dataclass(frozen=True) |
| 158 | class SubtitleMergeResult: |
| 159 | """Summary of one page-local SRT merge.""" |
| 160 | |
| 161 | slide_count: int |
| 162 | cue_count: int |
| 163 | powerpoint_timeline_ms: int |
| 164 | minimum_video_adjustment_ms: int | None = None |
| 165 | maximum_video_adjustment_ms: int | None = None |
| 166 | minimum_video_correlation: float | None = None |
| 167 | |
| 168 | |
| 169 | @dataclass(frozen=True) |
| 170 | class PowerPointTiming: |
| 171 | """One slide's transition, narration, and advance timing in milliseconds.""" |
| 172 | |
| 173 | transition_ms: int |
| 174 | narration_delay_ms: int |
| 175 | advance_ms: int |
| 176 | |
| 177 | |
| 178 | @dataclass(frozen=True) |
| 179 | class VideoSlideTiming: |
| 180 | """One slide mapped from the PPTX clock to an exported-video clock.""" |
| 181 | |
| 182 | slide_name: str |
| 183 | transition_ms: int |
| 184 | narration_delay_ms: int |
| 185 | advance_ms: int |
| 186 | powerpoint_slide_start_ms: int |
| 187 | powerpoint_narration_start_ms: int |
| 188 | video_slide_start_ms: int |
| 189 | video_narration_start_ms: int |
| 190 | adjustment_ms: int |
| 191 | correlation: float |
| 192 | audio_path: Path |
| 193 | |
| 194 | |
| 195 | @dataclass(frozen=True) |
| 196 | class VideoTimelineCalibration: |
| 197 | """Page-level calibration between one narrated PPTX and exported video.""" |
| 198 | |
| 199 | slides: tuple[VideoSlideTiming, ...] |
| 200 | powerpoint_timeline_ms: int |
| 201 | |
| 202 | |
| 203 | def _timestamp_to_ms(value: str) -> int: |
| 204 | hours_text, minutes_text, remainder = value.split(":") |
| 205 | seconds_text, milliseconds_text = remainder.split(",") |
| 206 | hours = int(hours_text) |
| 207 | minutes = int(minutes_text) |
| 208 | seconds = int(seconds_text) |
| 209 | milliseconds = int(milliseconds_text) |
| 210 | if minutes >= 60 or seconds >= 60: |
| 211 | raise ValueError(f"Invalid SRT timestamp: {value}") |
| 212 | return (((hours * 60) + minutes) * 60 + seconds) * 1000 + milliseconds |
| 213 | |
| 214 | |
| 215 | def _ms_to_timestamp(value: int) -> str: |
| 216 | if value < 0: |
| 217 | raise ValueError(f"SRT timestamp cannot be negative: {value}") |
| 218 | hours, remainder = divmod(value, 3_600_000) |
| 219 | minutes, remainder = divmod(remainder, 60_000) |
| 220 | seconds, milliseconds = divmod(remainder, 1000) |
| 221 | return f"{hours:02d}:{minutes:02d}:{seconds:02d},{milliseconds:03d}" |
| 222 | |
| 223 | |
| 224 | def _parse_srt(path: Path) -> list[SubtitleCue]: |
| 225 | text = path.read_text(encoding="utf-8-sig") |
| 226 | blocks = re.split(r"\r?\n\s*\r?\n", text.strip()) |
| 227 | cues: list[SubtitleCue] = [] |
| 228 | previous_end = -1 |
| 229 | |
| 230 | for block_number, block in enumerate(blocks, 1): |
| 231 | lines = block.splitlines() |
| 232 | if len(lines) < 3: |
| 233 | raise ValueError(f"{path}: malformed SRT block {block_number}") |
| 234 | try: |
| 235 | cue_number = int(lines[0].strip()) |
| 236 | except ValueError as exc: |
| 237 | raise ValueError( |
| 238 | f"{path}: invalid cue number in block {block_number}" |
| 239 | ) from exc |
| 240 | if cue_number != block_number: |
| 241 | raise ValueError( |
| 242 | f"{path}: cue numbers must be consecutive from 1; " |
| 243 | f"block {block_number} is numbered {cue_number}" |
| 244 | ) |
| 245 | timing_match = _TIMING_RE.match(lines[1].strip()) |
| 246 | if timing_match is None: |
| 247 | raise ValueError( |
| 248 | f"{path}: invalid cue timing in block {block_number}" |
| 249 | ) |
| 250 | start_ms = _timestamp_to_ms(timing_match.group("start")) |
| 251 | end_ms = _timestamp_to_ms(timing_match.group("end")) |
| 252 | cue_text = "\n".join(lines[2:]).strip() |
| 253 | if not cue_text: |
| 254 | raise ValueError(f"{path}: empty cue text in block {block_number}") |
| 255 | if end_ms <= start_ms: |
| 256 | raise ValueError( |
| 257 | f"{path}: cue {block_number} must end after it starts" |
| 258 | ) |
| 259 | if start_ms < previous_end: |
| 260 | raise ValueError( |
| 261 | f"{path}: cue {block_number} overlaps the preceding cue" |
| 262 | ) |
| 263 | cues.append(SubtitleCue(start_ms, end_ms, cue_text)) |
| 264 | previous_end = end_ms |
| 265 | |
| 266 | if not cues: |
| 267 | raise ValueError(f"No subtitle cues found: {path}") |
| 268 | return cues |
| 269 | |
| 270 | |
| 271 | def _format_srt(cues: list[SubtitleCue]) -> str: |
| 272 | blocks = [] |
| 273 | for index, cue in enumerate(cues, 1): |
| 274 | blocks.append( |
| 275 | f"{index}\n" |
| 276 | f"{_ms_to_timestamp(cue.start_ms)} --> " |
| 277 | f"{_ms_to_timestamp(cue.end_ms)}\n" |
| 278 | f"{cue.text}" |
| 279 | ) |
| 280 | return "\n\n".join(blocks) + "\n" |
| 281 | |
| 282 | |
| 283 | def _atomic_write_text(path: Path, text: str) -> None: |
| 284 | path.parent.mkdir(parents=True, exist_ok=True) |
| 285 | descriptor, temporary_name = tempfile.mkstemp( |
| 286 | prefix=f".{path.name}.", |
| 287 | suffix=".tmp", |
| 288 | dir=str(path.parent), |
| 289 | ) |
| 290 | temporary_path = Path(temporary_name) |
| 291 | try: |
| 292 | stream = os.fdopen(descriptor, "w", encoding="utf-8", newline="\n") |
| 293 | descriptor = -1 |
| 294 | with stream: |
| 295 | stream.write(text) |
| 296 | stream.flush() |
| 297 | os.fsync(stream.fileno()) |
| 298 | os.replace(temporary_path, path) |
| 299 | finally: |
| 300 | if descriptor >= 0: |
| 301 | os.close(descriptor) |
| 302 | temporary_path.unlink(missing_ok=True) |
| 303 | |
| 304 | |
| 305 | def _project_path(project_path: Path, value: str | None, default: Path) -> Path: |
| 306 | path = Path(value) if value else default |
| 307 | return path if path.is_absolute() else project_path / path |
| 308 | |
| 309 | |
| 310 | def _project_input_path(project_path: Path, value: str) -> Path: |
| 311 | path = Path(value) |
| 312 | if path.is_absolute(): |
| 313 | return path |
| 314 | return project_path / path |
| 315 | |
| 316 | |
| 317 | def _require_replaceable(path: Path, force: bool) -> None: |
| 318 | if path.exists() and not force: |
| 319 | raise FileExistsError(f"Output already exists: {path}; pass --force to replace it") |
| 320 | |
| 321 | |
| 322 | def _reject_output_alias( |
| 323 | output_path: Path, |
| 324 | input_paths: list[Path], |
| 325 | *, |
| 326 | label: str, |
| 327 | ) -> None: |
| 328 | output_resolved = output_path.resolve() |
| 329 | for input_path in input_paths: |
| 330 | if output_resolved == input_path.resolve(): |
| 331 | raise ValueError( |
| 332 | f"{label} output must not overwrite an input file: {output_path}" |
| 333 | ) |
| 334 | |
| 335 | |
| 336 | def _subtitle_fingerprint(slide_names: list[str], subtitle_dir: Path) -> str: |
| 337 | digest = hashlib.sha256() |
| 338 | for slide_name in slide_names: |
| 339 | path = subtitle_dir / f"{slide_name}.srt" |
| 340 | digest.update(slide_name.encode("utf-8")) |
| 341 | digest.update(b"\0") |
| 342 | digest.update(path.read_bytes()) |
| 343 | digest.update(b"\0") |
| 344 | return digest.hexdigest() |
| 345 | |
| 346 | |
| 347 | def _finite_non_negative_seconds(value: object, field: str) -> float: |
| 348 | if ( |
| 349 | isinstance(value, bool) |
| 350 | or not isinstance(value, (int, float)) |
| 351 | or not math.isfinite(float(value)) |
| 352 | or value < 0 |
| 353 | ): |
| 354 | raise ValueError(f'{field} must be a finite non-negative number') |
| 355 | return float(value) |
| 356 | |
| 357 | |
| 358 | def _load_timing_plan( |
| 359 | path: Path, |
| 360 | ) -> tuple[str, float, float | None, dict[str, list[TimingPlanEntry]]]: |
| 361 | raw = json.loads(path.read_text(encoding="utf-8")) |
| 362 | if not isinstance(raw, dict): |
| 363 | raise ValueError(f"Narration timing plan must be a JSON object: {path}") |
| 364 | unknown_top = set(raw) - { |
| 365 | "version", |
| 366 | "srt_sha256", |
| 367 | "narration_padding", |
| 368 | "narration_start_floor", |
| 369 | "slides", |
| 370 | } |
| 371 | if unknown_top: |
| 372 | raise ValueError( |
| 373 | f"Narration timing plan has unknown top-level field(s): " |
| 374 | f"{', '.join(sorted(unknown_top))}" |
| 375 | ) |
| 376 | if raw.get("version") != 1: |
| 377 | raise ValueError( |
| 378 | f"Unsupported narration timing plan version: {raw.get('version')!r}" |
| 379 | ) |
| 380 | srt_sha256 = raw.get("srt_sha256") |
| 381 | if ( |
| 382 | not isinstance(srt_sha256, str) |
| 383 | or re.fullmatch(r"[0-9a-f]{64}", srt_sha256) is None |
| 384 | ): |
| 385 | raise ValueError( |
| 386 | 'Narration timing plan field "srt_sha256" must be a lowercase ' |
| 387 | "SHA-256 digest of the ordered page-local SRT files" |
| 388 | ) |
| 389 | narration_padding = _finite_non_negative_seconds( |
| 390 | raw.get("narration_padding"), |
| 391 | 'Narration timing plan field "narration_padding"', |
| 392 | ) |
| 393 | narration_start_floor = ( |
| 394 | _finite_non_negative_seconds( |
| 395 | raw["narration_start_floor"], |
| 396 | 'Narration timing plan field "narration_start_floor"', |
| 397 | ) |
| 398 | if "narration_start_floor" in raw |
| 399 | else None |
| 400 | ) |
| 401 | slides = raw.get("slides") |
| 402 | if not isinstance(slides, dict): |
| 403 | raise ValueError('Narration timing plan field "slides" must be an object') |
| 404 | |
| 405 | result: dict[str, list[TimingPlanEntry]] = {} |
| 406 | for slide_name, slide_raw in slides.items(): |
| 407 | if not isinstance(slide_name, str) or not slide_name: |
| 408 | raise ValueError("Narration timing plan slide names must be non-empty strings") |
| 409 | if not isinstance(slide_raw, dict) or set(slide_raw) != {"groups"}: |
| 410 | raise ValueError( |
| 411 | f'Narration timing plan slide "{slide_name}" must contain only "groups"' |
| 412 | ) |
| 413 | groups = slide_raw["groups"] |
| 414 | if not isinstance(groups, list): |
| 415 | raise ValueError( |
| 416 | f'Narration timing plan slide "{slide_name}" groups must be a list' |
| 417 | ) |
| 418 | |
| 419 | entries: list[TimingPlanEntry] = [] |
| 420 | seen_groups: set[str] = set() |
| 421 | for position, entry_raw in enumerate(groups, 1): |
| 422 | if not isinstance(entry_raw, dict): |
| 423 | raise ValueError( |
| 424 | f'Narration timing plan "{slide_name}" group #{position} ' |
| 425 | "must be an object" |
| 426 | ) |
| 427 | unknown_fields = set(entry_raw) - {"id", "cue"} |
| 428 | if unknown_fields: |
| 429 | raise ValueError( |
| 430 | f'Narration timing plan "{slide_name}" group #{position} ' |
| 431 | f"has unknown field(s): {', '.join(sorted(unknown_fields))}" |
| 432 | ) |
| 433 | group_id = entry_raw.get("id") |
| 434 | if not isinstance(group_id, str) or not group_id.strip(): |
| 435 | raise ValueError( |
| 436 | f'Narration timing plan "{slide_name}" group #{position} ' |
| 437 | 'field "id" must be a non-empty string' |
| 438 | ) |
| 439 | if group_id in seen_groups: |
| 440 | raise ValueError( |
| 441 | f'Narration timing plan "{slide_name}" repeats group "{group_id}"' |
| 442 | ) |
| 443 | cue_number = entry_raw.get("cue") |
| 444 | if cue_number is not None and ( |
| 445 | isinstance(cue_number, bool) |
| 446 | or not isinstance(cue_number, int) |
| 447 | or cue_number <= 0 |
| 448 | ): |
| 449 | raise ValueError( |
| 450 | f'Narration timing plan "{slide_name}/{group_id}" cue ' |
| 451 | "must be a positive integer or null" |
| 452 | ) |
| 453 | entries.append(TimingPlanEntry(group_id, cue_number)) |
| 454 | seen_groups.add(group_id) |
| 455 | result[slide_name] = entries |
| 456 | return srt_sha256, narration_padding, narration_start_floor, result |
| 457 | |
| 458 | |
| 459 | def _load_canonical_animation_config(path: Path) -> dict[str, Any]: |
| 460 | """Load and field-validate the read-only canonical animation sidecar.""" |
| 461 | if not path.is_file(): |
| 462 | raise FileNotFoundError( |
| 463 | f"Canonical animation config is missing: {path}. " |
| 464 | "Complete the customize-animations stage first; narration sync " |
| 465 | "does not create or replace animations.json." |
| 466 | ) |
| 467 | raw = json.loads(path.read_text(encoding="utf-8")) |
| 468 | if not isinstance(raw, dict): |
| 469 | raise ValueError(f"Canonical animation config must be a JSON object: {path}") |
| 470 | if raw.get("version", 1) != 1: |
| 471 | raise ValueError( |
| 472 | f"Unsupported canonical animation config version: {raw.get('version')!r}" |
| 473 | ) |
| 474 | errors = list( |
| 475 | dict.fromkeys( |
| 476 | [ |
| 477 | *validate_transition_config(raw), |
| 478 | *validate_animation_config_errors(raw), |
| 479 | ] |
| 480 | ) |
| 481 | ) |
| 482 | if errors: |
| 483 | details = "\n".join(f"- {error}" for error in errors) |
| 484 | raise ValueError( |
| 485 | f"Canonical animation config is invalid: {path}\n{details}" |
| 486 | ) |
| 487 | return raw |
| 488 | |
| 489 | |
| 490 | def _page_subtitle_names(subtitle_dir: Path) -> list[str]: |
| 491 | """Return ordered page-local SRT stems, excluding the merged sidecar.""" |
| 492 | if not subtitle_dir.is_dir(): |
| 493 | raise FileNotFoundError(f"Page-local SRT directory not found: {subtitle_dir}") |
| 494 | names = [ |
| 495 | path.stem |
| 496 | for path in sorted(subtitle_dir.glob("*.srt")) |
| 497 | if path.stem != "total" |
| 498 | ] |
| 499 | if not names: |
| 500 | raise FileNotFoundError(f"No page-local SRT files found under: {subtitle_dir}") |
| 501 | return names |
| 502 | |
| 503 | |
| 504 | def _animation_slide_names( |
| 505 | project_path: Path, |
| 506 | config: dict[str, Any], |
| 507 | subtitle_dir: Path, |
| 508 | ) -> list[str]: |
| 509 | """Resolve the page roster without parsing SVG unless a slide is missing.""" |
| 510 | slides = config.get("slides", {}) |
| 511 | if not isinstance(slides, dict): |
| 512 | raise ValueError('Canonical animations.json field "slides" must be an object') |
| 513 | canonical_names = list(slides) |
| 514 | subtitle_names = _page_subtitle_names(subtitle_dir) |
| 515 | subtitle_set = set(subtitle_names) |
| 516 | |
| 517 | missing_subtitles = [ |
| 518 | slide_name for slide_name in canonical_names if slide_name not in subtitle_set |
| 519 | ] |
| 520 | if missing_subtitles: |
| 521 | raise FileNotFoundError( |
| 522 | "Missing page-local SRT for canonical animation slide(s): " |
| 523 | + ", ".join(missing_subtitles) |
| 524 | ) |
| 525 | |
| 526 | canonical_set = set(canonical_names) |
| 527 | extra_subtitles = [ |
| 528 | slide_name for slide_name in subtitle_names if slide_name not in canonical_set |
| 529 | ] |
| 530 | unexpected = [ |
| 531 | slide_name |
| 532 | for slide_name in extra_subtitles |
| 533 | if not (project_path / "svg_output" / f"{slide_name}.svg").is_file() |
| 534 | ] |
| 535 | if unexpected: |
| 536 | raise ValueError( |
| 537 | "Page-local SRT has no matching canonical animation slide or SVG: " |
| 538 | + ", ".join(unexpected) |
| 539 | ) |
| 540 | if not extra_subtitles: |
| 541 | return canonical_names |
| 542 | return subtitle_names |
| 543 | |
| 544 | |
| 545 | def _animation_scope( |
| 546 | scope: dict[str, Any], |
| 547 | *, |
| 548 | label: str, |
| 549 | ) -> dict[str, Any]: |
| 550 | value = scope.get("animation", {}) |
| 551 | if not isinstance(value, dict): |
| 552 | raise ValueError(f'{label} field "animation" must be an object') |
| 553 | return value |
| 554 | |
| 555 | |
| 556 | def _transition_scope( |
| 557 | scope: dict[str, Any], |
| 558 | *, |
| 559 | label: str, |
| 560 | ) -> dict[str, Any]: |
| 561 | value = scope.get("transition", {}) |
| 562 | if not isinstance(value, dict): |
| 563 | raise ValueError(f'{label} field "transition" must be an object') |
| 564 | return value |
| 565 | |
| 566 | |
| 567 | def _effective_transition_duration_ms( |
| 568 | config: dict[str, Any], |
| 569 | slide_cfg: dict[str, Any], |
| 570 | ) -> int: |
| 571 | """Resolve the destination slide's effective transition duration.""" |
| 572 | defaults = config.get("defaults", {}) |
| 573 | if not isinstance(defaults, dict): |
| 574 | raise ValueError('Canonical animations.json field "defaults" must be an object') |
| 575 | default_transition = _transition_scope( |
| 576 | defaults, |
| 577 | label="Canonical animations.json defaults", |
| 578 | ) |
| 579 | default_effect, _default_options = normalize_transition_effect_request( |
| 580 | default_transition.get("effect", "fade"), |
| 581 | default_transition.get("effect_options"), |
| 582 | ) |
| 583 | default_duration = validate_seconds( |
| 584 | default_transition.get("duration", DEFAULT_TRANSITION_DURATION), |
| 585 | "canonical transition duration", |
| 586 | allow_zero=default_effect is None, |
| 587 | ) |
| 588 | |
| 589 | slide_transition = _transition_scope( |
| 590 | slide_cfg, |
| 591 | label="Canonical animations.json slide", |
| 592 | ) |
| 593 | if "effect" in slide_transition: |
| 594 | effect, _effect_options = normalize_transition_effect_request( |
| 595 | slide_transition["effect"], |
| 596 | slide_transition.get("effect_options"), |
| 597 | ) |
| 598 | else: |
| 599 | effect = default_effect |
| 600 | duration = validate_seconds( |
| 601 | slide_transition.get("duration", default_duration), |
| 602 | "canonical slide transition duration", |
| 603 | allow_zero=effect is None, |
| 604 | ) |
| 605 | return 0 if effect is None else round(duration * 1000) |
| 606 | |
| 607 | |
| 608 | def _effective_slide_animation( |
| 609 | config: dict[str, Any], |
| 610 | slide_cfg: dict[str, Any], |
| 611 | ) -> SlideAnimationSettings: |
| 612 | defaults = config.get("defaults", {}) |
| 613 | if not isinstance(defaults, dict): |
| 614 | raise ValueError('Canonical animations.json field "defaults" must be an object') |
| 615 | default_animation = _animation_scope( |
| 616 | defaults, |
| 617 | label="Canonical animations.json defaults", |
| 618 | ) |
| 619 | slide_animation = _animation_scope( |
| 620 | slide_cfg, |
| 621 | label="Canonical animations.json slide", |
| 622 | ) |
| 623 | effect = normalize_animation_effect( |
| 624 | slide_animation.get( |
| 625 | "effect", |
| 626 | default_animation.get("effect", "none"), |
| 627 | ) |
| 628 | ) |
| 629 | duration_ms = animation_seconds_to_milliseconds( |
| 630 | slide_animation.get( |
| 631 | "duration", |
| 632 | default_animation.get("duration", 0.4), |
| 633 | ), |
| 634 | "canonical animation duration", |
| 635 | allow_zero=False, |
| 636 | ) |
| 637 | stagger_ms = animation_seconds_to_milliseconds( |
| 638 | slide_animation.get( |
| 639 | "stagger", |
| 640 | default_animation.get("stagger", 0.5), |
| 641 | ), |
| 642 | "canonical animation stagger", |
| 643 | allow_zero=True, |
| 644 | ) |
| 645 | trigger = normalize_animation_trigger( |
| 646 | slide_animation.get( |
| 647 | "trigger", |
| 648 | default_animation.get("trigger", "after-previous"), |
| 649 | ) |
| 650 | ) |
| 651 | timing_options = { |
| 652 | field: ( |
| 653 | slide_animation[field] |
| 654 | if field in slide_animation |
| 655 | else default_animation[field] |
| 656 | ) |
| 657 | for field in ANIMATION_TIMING_OPTION_FIELDS |
| 658 | if field in slide_animation or field in default_animation |
| 659 | } |
| 660 | return SlideAnimationSettings( |
| 661 | effect=effect, |
| 662 | duration_ms=duration_ms, |
| 663 | stagger_ms=stagger_ms, |
| 664 | trigger=trigger, |
| 665 | timing_options=timing_options, |
| 666 | ) |
| 667 | |
| 668 | |
| 669 | def _animation_playback_duration_ms( |
| 670 | duration_ms: int, |
| 671 | timing_options: dict[str, Any], |
| 672 | *, |
| 673 | label: str, |
| 674 | ) -> int: |
| 675 | """Return schedule duration after repeat and auto-reverse parameters.""" |
| 676 | if 'repeat_duration' in timing_options: |
| 677 | return animation_seconds_to_milliseconds( |
| 678 | timing_options['repeat_duration'], |
| 679 | f'{label} repeat_duration', |
| 680 | allow_zero=False, |
| 681 | ) |
| 682 | one_play = duration_ms * ( |
| 683 | 2 if timing_options.get('auto_reverse') is True else 1 |
| 684 | ) |
| 685 | repeat_count = timing_options.get('repeat_count', 1) |
| 686 | if ( |
| 687 | isinstance(repeat_count, bool) |
| 688 | or not isinstance(repeat_count, (int, float)) |
| 689 | or not math.isfinite(float(repeat_count)) |
| 690 | or float(repeat_count) <= 0 |
| 691 | ): |
| 692 | raise ValueError(f'{label} repeat_count must be a positive number') |
| 693 | return max(1, round(one_play * float(repeat_count))) |
| 694 | |
| 695 | |
| 696 | def _active_group_effect_entries( |
| 697 | slide_name: str, |
| 698 | group_id: str, |
| 699 | group_cfg: dict[str, Any], |
| 700 | slide_effect: str | None, |
| 701 | ) -> tuple[tuple[int | None, str, dict[str, Any]], ...]: |
| 702 | """Return active legacy or multi-effect rows with stable write locations.""" |
| 703 | group_path = ( |
| 704 | f'slides[{json.dumps(slide_name, ensure_ascii=False)}]' |
| 705 | f'.groups[{json.dumps(group_id, ensure_ascii=False)}]' |
| 706 | ) |
| 707 | effect_entries = animation_group_effect_entries( |
| 708 | group_cfg, |
| 709 | path=group_path, |
| 710 | ) |
| 711 | is_multi_effect = 'effects' in group_cfg |
| 712 | active: list[tuple[int | None, str, dict[str, Any]]] = [] |
| 713 | for index, (effect_path, effect_cfg) in enumerate(effect_entries): |
| 714 | effect = normalize_animation_effect( |
| 715 | effect_cfg.get("effect", slide_effect) |
| 716 | ) |
| 717 | if effect is None: |
| 718 | continue |
| 719 | active.append(( |
| 720 | index if is_multi_effect else None, |
| 721 | effect_path, |
| 722 | effect_cfg, |
| 723 | )) |
| 724 | return tuple(active) |
| 725 | |
| 726 | |
| 727 | def _group_is_animated( |
| 728 | slide_name: str, |
| 729 | group_id: str, |
| 730 | group_cfg: dict[str, Any], |
| 731 | slide_effect: str | None, |
| 732 | ) -> bool: |
| 733 | return bool( |
| 734 | _active_group_effect_entries( |
| 735 | slide_name, |
| 736 | group_id, |
| 737 | group_cfg, |
| 738 | slide_effect, |
| 739 | ) |
| 740 | ) |
| 741 | |
| 742 | |
| 743 | def _active_group_order_requires_svg( |
| 744 | slide_name: str, |
| 745 | groups_cfg: dict[str, Any], |
| 746 | group_ids: list[str], |
| 747 | slide_effect: str | None, |
| 748 | ) -> bool: |
| 749 | """Return whether SVG group order is needed to break sequence ambiguity.""" |
| 750 | if len(group_ids) <= 1: |
| 751 | return False |
| 752 | order_owners: dict[int, str] = {} |
| 753 | for group_id in group_ids: |
| 754 | for _effect_index, _effect_path, effect_cfg in _active_group_effect_entries( |
| 755 | slide_name, |
| 756 | group_id, |
| 757 | groups_cfg[group_id], |
| 758 | slide_effect, |
| 759 | ): |
| 760 | order = effect_cfg.get("order") |
| 761 | if order is None: |
| 762 | return True |
| 763 | previous_group = order_owners.setdefault(order, group_id) |
| 764 | if previous_group != group_id: |
| 765 | return True |
| 766 | return False |
| 767 | |
| 768 | |
| 769 | def _needs_svg_group_resolution( |
| 770 | slide_name: str, |
| 771 | settings: SlideAnimationSettings, |
| 772 | groups_cfg: dict[str, Any], |
| 773 | plan_entries: list[TimingPlanEntry] | None, |
| 774 | ) -> bool: |
| 775 | """Return whether JSON alone cannot prove the effective group sequence.""" |
| 776 | active_explicit = [ |
| 777 | group_id |
| 778 | for group_id, group_cfg in groups_cfg.items() |
| 779 | if isinstance(group_cfg, dict) |
| 780 | and _group_is_animated( |
| 781 | slide_name, |
| 782 | group_id, |
| 783 | group_cfg, |
| 784 | settings.effect, |
| 785 | ) |
| 786 | ] |
| 787 | if plan_entries is None: |
| 788 | if settings.effect is not None: |
| 789 | return True |
| 790 | return _active_group_order_requires_svg( |
| 791 | slide_name, |
| 792 | groups_cfg, |
| 793 | active_explicit, |
| 794 | settings.effect, |
| 795 | ) |
| 796 | |
| 797 | candidate_ids = list( |
| 798 | dict.fromkeys( |
| 799 | [ |
| 800 | *(entry.group_id for entry in plan_entries), |
| 801 | *active_explicit, |
| 802 | ] |
| 803 | ) |
| 804 | ) |
| 805 | active_candidates = [ |
| 806 | group_id |
| 807 | for group_id in candidate_ids |
| 808 | if group_id in groups_cfg |
| 809 | and isinstance(groups_cfg[group_id], dict) |
| 810 | and _group_is_animated( |
| 811 | slide_name, |
| 812 | group_id, |
| 813 | groups_cfg[group_id], |
| 814 | settings.effect, |
| 815 | ) |
| 816 | ] |
| 817 | return _active_group_order_requires_svg( |
| 818 | slide_name, |
| 819 | groups_cfg, |
| 820 | active_candidates, |
| 821 | settings.effect, |
| 822 | ) |
| 823 | |
| 824 | |
| 825 | def _resolve_animation_groups( |
| 826 | project_path: Path, |
| 827 | slide_name: str, |
| 828 | slide_cfg: dict[str, Any], |
| 829 | settings: SlideAnimationSettings, |
| 830 | plan_entries: list[TimingPlanEntry] | None, |
| 831 | ) -> tuple[list[AnimationGroupState], bool]: |
| 832 | """Resolve one effective sequence, parsing only an ambiguous page SVG.""" |
| 833 | groups_value = slide_cfg.get("groups", {}) |
| 834 | if not isinstance(groups_value, dict): |
| 835 | raise ValueError( |
| 836 | f'Canonical animations.json slide "{slide_name}" groups must be an object' |
| 837 | ) |
| 838 | groups_cfg: dict[str, dict[str, Any]] = {} |
| 839 | for group_id, group_cfg in groups_value.items(): |
| 840 | if not isinstance(group_cfg, dict): |
| 841 | raise ValueError( |
| 842 | f'Canonical animations.json group "{slide_name}/{group_id}" ' |
| 843 | "must be an object" |
| 844 | ) |
| 845 | groups_cfg[group_id] = group_cfg |
| 846 | use_svg = _needs_svg_group_resolution( |
| 847 | slide_name, |
| 848 | settings, |
| 849 | groups_cfg, |
| 850 | plan_entries, |
| 851 | ) |
| 852 | candidate_ids: list[str] |
| 853 | if use_svg: |
| 854 | svg_path = project_path / "svg_output" / f"{slide_name}.svg" |
| 855 | if not svg_path.is_file(): |
| 856 | raise FileNotFoundError( |
| 857 | f"Animation group mapping is ambiguous and requires the page SVG: " |
| 858 | f"{svg_path}" |
| 859 | ) |
| 860 | targets, anonymous_groups = scan_svg_targets(svg_path) |
| 861 | duplicate_ids = sorted( |
| 862 | group_id |
| 863 | for group_id in {target.group_id for target in targets} |
| 864 | if sum(target.group_id == group_id for target in targets) > 1 |
| 865 | ) |
| 866 | if duplicate_ids: |
| 867 | raise ValueError( |
| 868 | f'SVG slide "{slide_name}" has duplicate top-level group id(s): ' |
| 869 | + ", ".join(duplicate_ids) |
| 870 | ) |
| 871 | for item in anonymous_groups: |
| 872 | print( |
| 873 | f"Warning: {item} has no id and cannot participate in narration timing", |
| 874 | file=sys.stderr, |
| 875 | ) |
| 876 | |
| 877 | targets_by_id = {target.group_id: target for target in targets} |
| 878 | referenced_ids = set(groups_cfg) |
| 879 | if plan_entries is not None: |
| 880 | referenced_ids.update(entry.group_id for entry in plan_entries) |
| 881 | missing_ids = sorted(referenced_ids - set(targets_by_id)) |
| 882 | if missing_ids: |
| 883 | raise ValueError( |
| 884 | f'Animation mapping for slide "{slide_name}" references missing ' |
| 885 | f"top-level group(s): {', '.join(missing_ids)}" |
| 886 | ) |
| 887 | structural_ids = sorted( |
| 888 | group_id |
| 889 | for group_id in referenced_ids |
| 890 | if targets_by_id[group_id].structurally_static |
| 891 | and ( |
| 892 | group_id not in groups_cfg |
| 893 | or _group_is_animated( |
| 894 | slide_name, |
| 895 | group_id, |
| 896 | groups_cfg[group_id], |
| 897 | settings.effect, |
| 898 | ) |
| 899 | ) |
| 900 | ) |
| 901 | if structural_ids: |
| 902 | raise ValueError( |
| 903 | f'Animation mapping for slide "{slide_name}" targets structural ' |
| 904 | f"group(s): {', '.join(structural_ids)}" |
| 905 | ) |
| 906 | |
| 907 | candidate_ids = [] |
| 908 | for target in targets: |
| 909 | if target.structurally_static: |
| 910 | continue |
| 911 | group_cfg = groups_cfg.get(target.group_id, {}) |
| 912 | explicitly_animated = ( |
| 913 | target.group_id in groups_cfg |
| 914 | and _group_is_animated( |
| 915 | slide_name, |
| 916 | target.group_id, |
| 917 | group_cfg, |
| 918 | settings.effect, |
| 919 | ) |
| 920 | ) |
| 921 | if target.chrome and not explicitly_animated: |
| 922 | continue |
| 923 | candidate_ids.append(target.group_id) |
| 924 | elif plan_entries is not None: |
| 925 | candidate_ids = list( |
| 926 | dict.fromkeys( |
| 927 | [ |
| 928 | *(entry.group_id for entry in plan_entries), |
| 929 | *groups_cfg, |
| 930 | ] |
| 931 | ) |
| 932 | ) |
| 933 | else: |
| 934 | candidate_ids = list(groups_cfg) |
| 935 | |
| 936 | preliminaries: list[ |
| 937 | tuple[ |
| 938 | int, |
| 939 | int, |
| 940 | int, |
| 941 | str, |
| 942 | int | None, |
| 943 | str, |
| 944 | dict[str, Any], |
| 945 | str, |
| 946 | ] |
| 947 | ] = [] |
| 948 | for source_index, group_id in enumerate(candidate_ids): |
| 949 | group_cfg = groups_cfg.get(group_id, {}) |
| 950 | effect_entries = _active_group_effect_entries( |
| 951 | slide_name, |
| 952 | group_id, |
| 953 | group_cfg, |
| 954 | settings.effect, |
| 955 | ) |
| 956 | for effect_position, ( |
| 957 | effect_index, |
| 958 | effect_path, |
| 959 | effect_cfg, |
| 960 | ) in enumerate(effect_entries): |
| 961 | order = effect_cfg.get("order", source_index + 1) |
| 962 | if isinstance(order, bool) or not isinstance(order, int) or order <= 0: |
| 963 | raise ValueError( |
| 964 | f'Canonical animation order for "{effect_path}" ' |
| 965 | f"must be a positive integer: {order!r}" |
| 966 | ) |
| 967 | if effect_cfg.get("trigger_shape") is not None: |
| 968 | raise ValueError( |
| 969 | f'Recorded narration cannot synchronize trigger-shape ' |
| 970 | f'animation "{effect_path}" on slide "{slide_name}"' |
| 971 | ) |
| 972 | effect_trigger = normalize_animation_trigger( |
| 973 | effect_cfg.get("trigger", settings.trigger) |
| 974 | ) |
| 975 | if effect_trigger == "on-click": |
| 976 | raise ValueError( |
| 977 | f'Recorded narration cannot synchronize on-click animation ' |
| 978 | f'"{effect_path}" on slide "{slide_name}"' |
| 979 | ) |
| 980 | preliminaries.append(( |
| 981 | order, |
| 982 | source_index, |
| 983 | effect_position, |
| 984 | group_id, |
| 985 | effect_index, |
| 986 | effect_path, |
| 987 | effect_cfg, |
| 988 | effect_trigger, |
| 989 | )) |
| 990 | preliminaries.sort(key=lambda item: (item[0], item[1], item[2])) |
| 991 | |
| 992 | states: list[AnimationGroupState] = [] |
| 993 | for sequence_index, ( |
| 994 | order, |
| 995 | source_index, |
| 996 | _effect_position, |
| 997 | group_id, |
| 998 | effect_index, |
| 999 | effect_path, |
| 1000 | effect_cfg, |
| 1001 | effect_trigger, |
| 1002 | ) in enumerate(preliminaries): |
| 1003 | duration_ms = animation_seconds_to_milliseconds( |
| 1004 | effect_cfg.get("duration", settings.duration_ms / 1000), |
| 1005 | f'canonical animation duration for "{effect_path}"', |
| 1006 | allow_zero=False, |
| 1007 | ) |
| 1008 | timing_options = dict(settings.timing_options) |
| 1009 | timing_options.update( |
| 1010 | { |
| 1011 | field: effect_cfg[field] |
| 1012 | for field in ANIMATION_TIMING_OPTION_FIELDS |
| 1013 | if field in effect_cfg |
| 1014 | } |
| 1015 | ) |
| 1016 | playback_duration_ms = _animation_playback_duration_ms( |
| 1017 | duration_ms, |
| 1018 | timing_options, |
| 1019 | label=f'canonical animation for "{effect_path}"', |
| 1020 | ) |
| 1021 | default_delay = ( |
| 1022 | settings.stagger_ms / 1000 |
| 1023 | if effect_trigger == "after-previous" and sequence_index > 0 |
| 1024 | else 0 |
| 1025 | ) |
| 1026 | original_delay_ms = animation_seconds_to_milliseconds( |
| 1027 | effect_cfg.get("delay", default_delay), |
| 1028 | f'canonical animation delay for "{effect_path}"', |
| 1029 | allow_zero=True, |
| 1030 | ) |
| 1031 | states.append( |
| 1032 | AnimationGroupState( |
| 1033 | group_id=group_id, |
| 1034 | effect_index=effect_index, |
| 1035 | order=order, |
| 1036 | source_index=source_index, |
| 1037 | duration_ms=playback_duration_ms, |
| 1038 | original_delay_ms=original_delay_ms, |
| 1039 | trigger=effect_trigger, |
| 1040 | ) |
| 1041 | ) |
| 1042 | return states, use_svg |
| 1043 | |
| 1044 | |
| 1045 | def _find_audio(audio_dir: Path, slide_name: str) -> Path: |
| 1046 | matches = [ |
| 1047 | path |
| 1048 | for path in audio_dir.iterdir() |
| 1049 | if path.is_file() |
| 1050 | and path.stem == slide_name |
| 1051 | and path.suffix.lower() in NARRATION_EXTENSIONS |
| 1052 | ] |
| 1053 | if not matches: |
| 1054 | raise FileNotFoundError(f"Missing narration audio for slide: {slide_name}") |
| 1055 | if len(matches) > 1: |
| 1056 | rendered = ", ".join(str(path) for path in matches) |
| 1057 | raise ValueError(f"Multiple narration audio files match {slide_name}: {rendered}") |
| 1058 | return matches[0] |
| 1059 | |
| 1060 | |
| 1061 | def _seconds_from_ms(value: int) -> float: |
| 1062 | return round(value / 1000, 3) |
| 1063 | |
| 1064 | |
| 1065 | def rebuild_animations( |
| 1066 | project_path: Path, |
| 1067 | *, |
| 1068 | canonical_path: Path, |
| 1069 | plan_path: Path, |
| 1070 | subtitle_dir: Path, |
| 1071 | audio_dir: Path, |
| 1072 | output_path: Path, |
| 1073 | narration_padding: float, |
| 1074 | force: bool, |
| 1075 | narration_start_floor: float = DEFAULT_NARRATION_START_FLOOR, |
| 1076 | ) -> AnimationBuildResult: |
| 1077 | """Derive narration timing without modifying the canonical animation file.""" |
| 1078 | narration_padding = _finite_non_negative_seconds( |
| 1079 | narration_padding, |
| 1080 | "Narration padding", |
| 1081 | ) |
| 1082 | narration_start_floor = _finite_non_negative_seconds( |
| 1083 | narration_start_floor, |
| 1084 | "Narration start floor", |
| 1085 | ) |
| 1086 | _reject_output_alias( |
| 1087 | output_path, |
| 1088 | [canonical_path, plan_path], |
| 1089 | label="Narration animation config", |
| 1090 | ) |
| 1091 | canonical = _load_canonical_animation_config(canonical_path) |
| 1092 | slide_names = _animation_slide_names(project_path, canonical, subtitle_dir) |
| 1093 | subtitle_paths = [subtitle_dir / f"{slide_name}.srt" for slide_name in slide_names] |
| 1094 | _reject_output_alias( |
| 1095 | output_path, |
| 1096 | subtitle_paths, |
| 1097 | label="Narration animation config", |
| 1098 | ) |
| 1099 | |
| 1100 | timing_plan: dict[str, list[TimingPlanEntry]] | None = None |
| 1101 | if plan_path.is_file(): |
| 1102 | ( |
| 1103 | expected_srt_sha256, |
| 1104 | planned_padding, |
| 1105 | planned_start_floor, |
| 1106 | loaded_plan, |
| 1107 | ) = _load_timing_plan(plan_path) |
| 1108 | if not math.isclose( |
| 1109 | narration_padding, |
| 1110 | planned_padding, |
| 1111 | rel_tol=0, |
| 1112 | abs_tol=1e-9, |
| 1113 | ): |
| 1114 | raise ValueError( |
| 1115 | "Narration padding differs from the timing plan: " |
| 1116 | f"plan={planned_padding}, command={narration_padding}" |
| 1117 | ) |
| 1118 | if planned_start_floor is not None and not math.isclose( |
| 1119 | narration_start_floor, |
| 1120 | planned_start_floor, |
| 1121 | rel_tol=0, |
| 1122 | abs_tol=1e-9, |
| 1123 | ): |
| 1124 | raise ValueError( |
| 1125 | "Narration start floor differs from the timing plan: " |
| 1126 | f"plan={planned_start_floor}, command={narration_start_floor}" |
| 1127 | ) |
| 1128 | current_srt_sha256 = _subtitle_fingerprint(slide_names, subtitle_dir) |
| 1129 | if current_srt_sha256 != expected_srt_sha256: |
| 1130 | raise ValueError( |
| 1131 | "Narration timing plan was authored for a different SRT set: " |
| 1132 | f"plan={expected_srt_sha256}, current={current_srt_sha256}" |
| 1133 | ) |
| 1134 | if set(loaded_plan) != set(slide_names): |
| 1135 | raise ValueError( |
| 1136 | "Narration timing plan slides do not match animations.json" |
| 1137 | ) |
| 1138 | timing_plan = loaded_plan |
| 1139 | |
| 1140 | derived = copy.deepcopy(canonical) |
| 1141 | canonical_slides = canonical.get("slides", {}) |
| 1142 | derived_slides = derived.setdefault("slides", {}) |
| 1143 | if not isinstance(canonical_slides, dict) or not isinstance( |
| 1144 | derived_slides, |
| 1145 | dict, |
| 1146 | ): |
| 1147 | raise ValueError('Canonical animations.json field "slides" must be an object') |
| 1148 | |
| 1149 | anchored_count = 0 |
| 1150 | fallback_count = 0 |
| 1151 | ignored_cue_count = 0 |
| 1152 | svg_fallback_slide_count = 0 |
| 1153 | drift_warnings: list[str] = [] |
| 1154 | positional_slides: list[tuple[str, int, int]] = [] |
| 1155 | audio_paths: list[Path] = [] |
| 1156 | |
| 1157 | for slide_name in slide_names: |
| 1158 | cues = _parse_srt(subtitle_dir / f"{slide_name}.srt") |
| 1159 | audio_path = _find_audio(audio_dir, slide_name) |
| 1160 | audio_paths.append(audio_path) |
| 1161 | audio_duration = probe_audio_duration(audio_path) |
| 1162 | if audio_duration is None: |
| 1163 | raise RuntimeError( |
| 1164 | f"Unable to read narration duration with ffprobe: {audio_path}" |
| 1165 | ) |
| 1166 | |
| 1167 | canonical_slide = canonical_slides.get(slide_name, {}) |
| 1168 | if not isinstance(canonical_slide, dict): |
| 1169 | raise ValueError( |
| 1170 | f'Canonical animations.json slide "{slide_name}" must be an object' |
| 1171 | ) |
| 1172 | derived_slide = derived_slides.setdefault( |
| 1173 | slide_name, |
| 1174 | copy.deepcopy(canonical_slide), |
| 1175 | ) |
| 1176 | if not isinstance(derived_slide, dict): |
| 1177 | raise ValueError( |
| 1178 | f'Derived animation slide "{slide_name}" must be an object' |
| 1179 | ) |
| 1180 | settings = _effective_slide_animation(canonical, canonical_slide) |
| 1181 | transition_duration_ms = _effective_transition_duration_ms( |
| 1182 | canonical, |
| 1183 | canonical_slide, |
| 1184 | ) |
| 1185 | narration_lead_in_ms = round( |
| 1186 | narration_lead_in_seconds( |
| 1187 | transition_duration_ms / 1000, |
| 1188 | start_floor=narration_start_floor, |
| 1189 | ) |
| 1190 | * 1000 |
| 1191 | ) |
| 1192 | plan_entries = timing_plan.get(slide_name) if timing_plan else None |
| 1193 | states, used_svg = _resolve_animation_groups( |
| 1194 | project_path, |
| 1195 | slide_name, |
| 1196 | canonical_slide, |
| 1197 | settings, |
| 1198 | plan_entries, |
| 1199 | ) |
| 1200 | if used_svg: |
| 1201 | svg_fallback_slide_count += 1 |
| 1202 | if timing_plan is None and states: |
| 1203 | positional_slides.append(( |
| 1204 | slide_name, |
| 1205 | len({state.group_id for state in states}), |
| 1206 | len(cues), |
| 1207 | )) |
| 1208 | |
| 1209 | state_ids = {state.group_id for state in states} |
| 1210 | cue_by_group: dict[str, int | None] = {} |
| 1211 | if plan_entries is not None: |
| 1212 | for entry in plan_entries: |
| 1213 | if entry.group_id not in state_ids: |
| 1214 | raise ValueError( |
| 1215 | f'Narration timing plan references a non-animated group: ' |
| 1216 | f"{slide_name}/{entry.group_id}" |
| 1217 | ) |
| 1218 | if entry.cue_number is not None and entry.cue_number > len(cues): |
| 1219 | raise ValueError( |
| 1220 | f'Narration timing plan "{slide_name}/{entry.group_id}" ' |
| 1221 | f"references cue {entry.cue_number}, but the SRT has " |
| 1222 | f"{len(cues)} cues" |
| 1223 | ) |
| 1224 | cue_by_group[entry.group_id] = entry.cue_number |
| 1225 | else: |
| 1226 | ordered_group_ids = list( |
| 1227 | dict.fromkeys(state.group_id for state in states) |
| 1228 | ) |
| 1229 | cue_by_group = { |
| 1230 | group_id: index + 1 if index < len(cues) else None |
| 1231 | for index, group_id in enumerate(ordered_group_ids) |
| 1232 | } |
| 1233 | |
| 1234 | animation_value = derived_slide.setdefault("animation", {}) |
| 1235 | if not isinstance(animation_value, dict): |
| 1236 | raise ValueError( |
| 1237 | f'Derived animation slide "{slide_name}" animation must be an object' |
| 1238 | ) |
| 1239 | animation_value["trigger"] = "after-previous" |
| 1240 | groups_value = derived_slide.setdefault("groups", {}) |
| 1241 | if not isinstance(groups_value, dict): |
| 1242 | raise ValueError( |
| 1243 | f'Derived animation slide "{slide_name}" groups must be an object' |
| 1244 | ) |
| 1245 | |
| 1246 | # Start modes are row-relative; slide completion spans overlapping rows. |
| 1247 | previous_start_ms = 0 |
| 1248 | previous_row_end_ms = 0 |
| 1249 | timeline_end_ms = 0 |
| 1250 | has_previous_row = False |
| 1251 | referenced_cues: set[int] = set() |
| 1252 | seen_groups: set[str] = set() |
| 1253 | for state in states: |
| 1254 | first_group_effect = state.group_id not in seen_groups |
| 1255 | seen_groups.add(state.group_id) |
| 1256 | cue_number = ( |
| 1257 | cue_by_group.get(state.group_id) |
| 1258 | if first_group_effect |
| 1259 | else None |
| 1260 | ) |
| 1261 | if not has_previous_row: |
| 1262 | sequence_base_ms = 0 |
| 1263 | elif state.trigger == "with-previous": |
| 1264 | sequence_base_ms = previous_start_ms |
| 1265 | else: |
| 1266 | sequence_base_ms = previous_row_end_ms |
| 1267 | if cue_number is None: |
| 1268 | if first_group_effect: |
| 1269 | fallback_count += 1 |
| 1270 | delay_ms = state.original_delay_ms |
| 1271 | actual_start_ms = sequence_base_ms + delay_ms |
| 1272 | else: |
| 1273 | anchored_count += 1 |
| 1274 | referenced_cues.add(cue_number) |
| 1275 | cue_start_ms = cues[cue_number - 1].start_ms |
| 1276 | desired_start_ms = narration_lead_in_ms + cue_start_ms |
| 1277 | actual_start_ms = max(desired_start_ms, sequence_base_ms) |
| 1278 | delay_ms = actual_start_ms - sequence_base_ms |
| 1279 | drift_ms = actual_start_ms - desired_start_ms |
| 1280 | if drift_ms > 500: |
| 1281 | drift_warnings.append( |
| 1282 | f"{slide_name}/{state.group_id}: cue {cue_number} " |
| 1283 | f"starts at {_seconds_from_ms(cue_start_ms):.3f}s after " |
| 1284 | f"a {_seconds_from_ms(narration_lead_in_ms):.3f}s lead-in; " |
| 1285 | f"animation starts at {_seconds_from_ms(actual_start_ms):.3f}s " |
| 1286 | f"({state.trigger} drift " |
| 1287 | f"{_seconds_from_ms(drift_ms):.3f}s)" |
| 1288 | ) |
| 1289 | |
| 1290 | group_value = groups_value.setdefault(state.group_id, {}) |
| 1291 | if not isinstance(group_value, dict): |
| 1292 | raise ValueError( |
| 1293 | f'Derived animation group "{slide_name}/{state.group_id}" ' |
| 1294 | "must be an object" |
| 1295 | ) |
| 1296 | if state.effect_index is None: |
| 1297 | effect_value = group_value |
| 1298 | else: |
| 1299 | group_path = ( |
| 1300 | f'slides[{json.dumps(slide_name, ensure_ascii=False)}]' |
| 1301 | f'.groups[{json.dumps(state.group_id, ensure_ascii=False)}]' |
| 1302 | ) |
| 1303 | derived_effect_entries = animation_group_effect_entries( |
| 1304 | group_value, |
| 1305 | path=group_path, |
| 1306 | ) |
| 1307 | effect_value = derived_effect_entries[state.effect_index][1] |
| 1308 | effect_value["order"] = state.order |
| 1309 | effect_value["delay"] = _seconds_from_ms(delay_ms) |
| 1310 | effect_value["trigger"] = state.trigger |
| 1311 | previous_start_ms = actual_start_ms |
| 1312 | previous_row_end_ms = actual_start_ms + state.duration_ms |
| 1313 | timeline_end_ms = max(timeline_end_ms, previous_row_end_ms) |
| 1314 | has_previous_row = True |
| 1315 | |
| 1316 | ignored_cue_count += len(cues) - len(referenced_cues) |
| 1317 | |
| 1318 | advance_ms = round( |
| 1319 | ( |
| 1320 | audio_duration |
| 1321 | + narration_padding |
| 1322 | + narration_lead_in_ms / 1000 |
| 1323 | ) |
| 1324 | * 1000 |
| 1325 | ) |
| 1326 | if timeline_end_ms > advance_ms: |
| 1327 | raise ValueError( |
| 1328 | f'Animations on slide "{slide_name}" end at ' |
| 1329 | f"{_seconds_from_ms(timeline_end_ms):.3f}s, after the recorded " |
| 1330 | f"slide advance at {_seconds_from_ms(advance_ms):.3f}s" |
| 1331 | ) |
| 1332 | |
| 1333 | _reject_output_alias( |
| 1334 | output_path, |
| 1335 | audio_paths, |
| 1336 | label="Narration animation config", |
| 1337 | ) |
| 1338 | _require_replaceable(output_path, force) |
| 1339 | derived_errors = list( |
| 1340 | dict.fromkeys( |
| 1341 | [ |
| 1342 | *validate_transition_config(derived), |
| 1343 | *validate_animation_config_errors(derived), |
| 1344 | ] |
| 1345 | ) |
| 1346 | ) |
| 1347 | if derived_errors: |
| 1348 | details = "\n".join(f"- {error}" for error in derived_errors) |
| 1349 | raise ValueError(f"Derived narration animation config is invalid:\n{details}") |
| 1350 | _atomic_write_text( |
| 1351 | output_path, |
| 1352 | json.dumps(derived, ensure_ascii=False, indent=2) + "\n", |
| 1353 | ) |
| 1354 | |
| 1355 | for warning in drift_warnings: |
| 1356 | print(f"Warning: {warning}", file=sys.stderr) |
| 1357 | if positional_slides: |
| 1358 | print( |
| 1359 | "Warning: no narration_timing.json found — object reveals were mapped " |
| 1360 | "positionally (group N -> subtitle cue N). This mistimes any page whose " |
| 1361 | "narration is longer than its object count: later objects reveal early, " |
| 1362 | "while the narrator is still on an earlier point. Author " |
| 1363 | f"{plan_path} mapping each SVG group to the subtitle cue that speaks " |
| 1364 | "about it (omit a group's cue to keep its canonical delay), then re-run.", |
| 1365 | file=sys.stderr, |
| 1366 | ) |
| 1367 | risky = [ |
| 1368 | (name, group_count, cue_count) |
| 1369 | for name, group_count, cue_count in positional_slides |
| 1370 | if group_count > 1 and cue_count > group_count + 1 |
| 1371 | ] |
| 1372 | for name, group_count, cue_count in risky: |
| 1373 | print( |
| 1374 | f"Warning: {name}: {group_count} object(s) but {cue_count} " |
| 1375 | "subtitle cue(s) — later objects likely reveal too early", |
| 1376 | file=sys.stderr, |
| 1377 | ) |
| 1378 | return AnimationBuildResult( |
| 1379 | slide_count=len(slide_names), |
| 1380 | group_count=anchored_count + fallback_count, |
| 1381 | anchored_count=anchored_count, |
| 1382 | fallback_count=fallback_count, |
| 1383 | ignored_cue_count=ignored_cue_count, |
| 1384 | svg_fallback_slide_count=svg_fallback_slide_count, |
| 1385 | timing_plan_written=False, |
| 1386 | ) |
| 1387 | |
| 1388 | |
| 1389 | def presentation_slide_members(package: zipfile.ZipFile) -> list[str]: |
| 1390 | """Return slide package members in presentation order.""" |
| 1391 | try: |
| 1392 | presentation_root = ET.fromstring(package.read("ppt/presentation.xml")) |
| 1393 | relationships_root = ET.fromstring( |
| 1394 | package.read("ppt/_rels/presentation.xml.rels") |
| 1395 | ) |
| 1396 | except KeyError as exc: |
| 1397 | raise ValueError( |
| 1398 | f"Narrated PPTX is missing presentation ordering data: {exc}" |
| 1399 | ) from exc |
| 1400 | |
| 1401 | relationship_targets: dict[str, str] = {} |
| 1402 | for relationship in relationships_root.iter(f"{{{_REL_NS}}}Relationship"): |
| 1403 | relationship_id = relationship.get("Id") |
| 1404 | target = relationship.get("Target") |
| 1405 | if ( |
| 1406 | relationship_id |
| 1407 | and target |
| 1408 | and relationship.get("TargetMode", "Internal") != "External" |
| 1409 | ): |
| 1410 | relationship_targets[relationship_id] = target.replace("\\", "/") |
| 1411 | |
| 1412 | slide_list = presentation_root.find(f"{{{_PML_NS}}}sldIdLst") |
| 1413 | if slide_list is None: |
| 1414 | raise ValueError("Narrated PPTX presentation has no slide order") |
| 1415 | |
| 1416 | members: list[str] = [] |
| 1417 | for slide_id in slide_list.findall(f"{{{_PML_NS}}}sldId"): |
| 1418 | relationship_id = slide_id.get(f"{{{_DOC_REL_NS}}}id") |
| 1419 | target = relationship_targets.get(relationship_id or "") |
| 1420 | if target is None: |
| 1421 | raise ValueError( |
| 1422 | "Narrated PPTX slide order references a missing relationship: " |
| 1423 | f"{relationship_id!r}" |
| 1424 | ) |
| 1425 | if target.startswith("/"): |
| 1426 | member = posixpath.normpath(target.lstrip("/")) |
| 1427 | else: |
| 1428 | member = posixpath.normpath(posixpath.join("ppt", target)) |
| 1429 | if member not in package.namelist(): |
| 1430 | raise ValueError( |
| 1431 | f"Narrated PPTX slide relationship target is missing: {member}" |
| 1432 | ) |
| 1433 | members.append(member) |
| 1434 | return members |
| 1435 | |
| 1436 | |
| 1437 | def _read_powerpoint_timings( |
| 1438 | pptx_path: Path, |
| 1439 | slide_count: int, |
| 1440 | ) -> list[PowerPointTiming]: |
| 1441 | timings: list[PowerPointTiming] = [] |
| 1442 | with zipfile.ZipFile(pptx_path) as package: |
| 1443 | slide_members = presentation_slide_members(package) |
| 1444 | if len(slide_members) != slide_count: |
| 1445 | raise ValueError( |
| 1446 | f"Narrated PPTX has {len(slide_members)} slides, " |
| 1447 | f"but the project has {slide_count}" |
| 1448 | ) |
| 1449 | for slide_index, member in enumerate(slide_members, 1): |
| 1450 | slide_xml = package.read(member) |
| 1451 | summary = read_slide_transition_xml(slide_xml) |
| 1452 | if summary.logical_count != 1: |
| 1453 | raise ValueError( |
| 1454 | f"Narrated PPTX slide {slide_index} has " |
| 1455 | f"{summary.logical_count} logical transition carriers" |
| 1456 | ) |
| 1457 | advance_ms = summary.advance_after_ms |
| 1458 | if advance_ms is None: |
| 1459 | raise ValueError( |
| 1460 | f"Narrated PPTX slide {slide_index} has no recorded advance time" |
| 1461 | ) |
| 1462 | transition_ms = summary.duration_ms or 0 |
| 1463 | narration_delay_ms = read_narration_start_delay_xml( |
| 1464 | slide_xml.decode("utf-8") |
| 1465 | ) |
| 1466 | if ( |
| 1467 | advance_ms <= 0 |
| 1468 | or transition_ms < 0 |
| 1469 | or narration_delay_ms < 0 |
| 1470 | or narration_delay_ms >= advance_ms |
| 1471 | ): |
| 1472 | raise ValueError( |
| 1473 | f"Narrated PPTX slide {slide_index} has invalid timing values" |
| 1474 | ) |
| 1475 | timings.append( |
| 1476 | PowerPointTiming( |
| 1477 | transition_ms=transition_ms, |
| 1478 | narration_delay_ms=narration_delay_ms, |
| 1479 | advance_ms=advance_ms, |
| 1480 | ) |
| 1481 | ) |
| 1482 | return timings |
| 1483 | |
| 1484 | |
| 1485 | def _powerpoint_audio_starts( |
| 1486 | timings: list[PowerPointTiming], |
| 1487 | ) -> tuple[list[int], int]: |
| 1488 | """Return theoretical narration starts and the complete PPTX timeline.""" |
| 1489 | audio_starts: list[int] = [] |
| 1490 | timeline_ms = 0 |
| 1491 | for timing in timings: |
| 1492 | slide_start_ms = timeline_ms |
| 1493 | audio_start_ms = ( |
| 1494 | slide_start_ms |
| 1495 | + timing.transition_ms |
| 1496 | + timing.narration_delay_ms |
| 1497 | ) |
| 1498 | audio_starts.append(audio_start_ms) |
| 1499 | timeline_ms = ( |
| 1500 | slide_start_ms + timing.transition_ms + timing.advance_ms |
| 1501 | ) |
| 1502 | return audio_starts, timeline_ms |
| 1503 | |
| 1504 | |
| 1505 | def _require_numpy() -> Any: |
| 1506 | try: |
| 1507 | import numpy as np |
| 1508 | except ImportError as exc: |
| 1509 | raise RuntimeError( |
| 1510 | "Exported-video timeline calibration requires numpy. " |
| 1511 | "Install it with: python3 -m pip install numpy" |
| 1512 | ) from exc |
| 1513 | return np |
| 1514 | |
| 1515 | |
| 1516 | def _decode_audio_envelopes(path: Path, ffmpeg_path: str) -> tuple[Any, Any]: |
| 1517 | """Decode the first audio stream and return 1 ms and 10 ms RMS envelopes.""" |
| 1518 | np = _require_numpy() |
| 1519 | command = [ |
| 1520 | ffmpeg_path, |
| 1521 | "-v", |
| 1522 | "error", |
| 1523 | "-i", |
| 1524 | str(path), |
| 1525 | "-map", |
| 1526 | "0:a:0", |
| 1527 | "-vn", |
| 1528 | "-ac", |
| 1529 | "1", |
| 1530 | "-ar", |
| 1531 | str(_ALIGNMENT_SAMPLE_RATE), |
| 1532 | "-f", |
| 1533 | "s16le", |
| 1534 | "-", |
| 1535 | ] |
| 1536 | result = subprocess.run(command, capture_output=True, check=False) |
| 1537 | if result.returncode != 0: |
| 1538 | details = result.stderr.decode("utf-8", errors="replace").strip() |
| 1539 | raise RuntimeError(f"Unable to decode audio with ffmpeg: {path}\n{details}") |
| 1540 | |
| 1541 | samples = np.frombuffer(result.stdout, dtype="<i2") |
| 1542 | samples_per_frame = _ALIGNMENT_SAMPLE_RATE // _ALIGNMENT_FINE_HZ |
| 1543 | usable_sample_count = len(samples) // samples_per_frame * samples_per_frame |
| 1544 | if usable_sample_count < samples_per_frame * 500: |
| 1545 | raise ValueError(f"Audio is too short for subtitle calibration: {path}") |
| 1546 | |
| 1547 | frames = ( |
| 1548 | samples[:usable_sample_count] |
| 1549 | .astype(np.float32) |
| 1550 | .reshape(-1, samples_per_frame) |
| 1551 | ) |
| 1552 | fine = np.sqrt(np.mean(frames * frames, axis=1) + 1e-12) |
| 1553 | fine = np.log1p(120 * fine) |
| 1554 | smooth_width = max(1, _ALIGNMENT_FINE_HZ // 200) |
| 1555 | fine = np.convolve( |
| 1556 | fine, |
| 1557 | np.ones(smooth_width, dtype=np.float64) / smooth_width, |
| 1558 | mode="same", |
| 1559 | ).astype(np.float64) |
| 1560 | |
| 1561 | coarse_count = len(fine) // _ALIGNMENT_COARSE_FACTOR |
| 1562 | coarse = fine[ |
| 1563 | :coarse_count * _ALIGNMENT_COARSE_FACTOR |
| 1564 | ].reshape(coarse_count, _ALIGNMENT_COARSE_FACTOR).mean(axis=1) |
| 1565 | return fine, coarse |
| 1566 | |
| 1567 | |
| 1568 | def _best_correlation(search: Any, template: Any) -> tuple[int, float]: |
| 1569 | """Return the best normalized-correlation index for one search window.""" |
| 1570 | np = _require_numpy() |
| 1571 | if len(search) < len(template): |
| 1572 | raise ValueError("Video alignment search window is shorter than its template") |
| 1573 | |
| 1574 | centered_template = template - template.mean() |
| 1575 | template_norm = float(np.linalg.norm(centered_template)) |
| 1576 | if template_norm <= 1e-9: |
| 1577 | raise ValueError("Narration audio has no usable variation for video alignment") |
| 1578 | |
| 1579 | numerators = np.correlate(search, centered_template, mode="valid") |
| 1580 | prefix = np.concatenate(([0.0], np.cumsum(search, dtype=np.float64))) |
| 1581 | squared_prefix = np.concatenate( |
| 1582 | ([0.0], np.cumsum(search * search, dtype=np.float64)) |
| 1583 | ) |
| 1584 | width = len(template) |
| 1585 | window_sums = prefix[width:] - prefix[:-width] |
| 1586 | window_squared_sums = squared_prefix[width:] - squared_prefix[:-width] |
| 1587 | window_variances = np.maximum( |
| 1588 | window_squared_sums - (window_sums * window_sums / width), |
| 1589 | 1e-18, |
| 1590 | ) |
| 1591 | scores = numerators / (np.sqrt(window_variances) * template_norm) |
| 1592 | best_index = int(np.argmax(scores)) |
| 1593 | return best_index, float(scores[best_index]) |
| 1594 | |
| 1595 | |
| 1596 | def _alignment_template_bounds( |
| 1597 | fine_envelope: Any, |
| 1598 | cues: list[SubtitleCue] | None, |
| 1599 | ) -> tuple[int, int]: |
| 1600 | duration_ms = len(fine_envelope) |
| 1601 | if cues: |
| 1602 | start_ms = min(cues[0].start_ms, max(0, duration_ms - 500)) |
| 1603 | cue_end_ms = min(cues[-1].end_ms, duration_ms) |
| 1604 | else: |
| 1605 | start_ms = 0 |
| 1606 | cue_end_ms = duration_ms |
| 1607 | end_ms = min(duration_ms, start_ms + _ALIGNMENT_TEMPLATE_MAX_MS) |
| 1608 | end_ms = min(end_ms, max(start_ms + 1000, cue_end_ms)) |
| 1609 | if end_ms - start_ms < 500: |
| 1610 | raise ValueError("Narration cue range is too short for video alignment") |
| 1611 | return start_ms, end_ms |
| 1612 | |
| 1613 | |
| 1614 | def _locate_audio_start( |
| 1615 | video_fine: Any, |
| 1616 | video_coarse: Any, |
| 1617 | audio_fine: Any, |
| 1618 | audio_coarse: Any, |
| 1619 | cues: list[SubtitleCue] | None, |
| 1620 | predicted_start_ms: int, |
| 1621 | ) -> tuple[int, float]: |
| 1622 | """Locate one page narration near its predicted exported-video position.""" |
| 1623 | start_ms, end_ms = _alignment_template_bounds(audio_fine, cues) |
| 1624 | coarse_start = start_ms // _ALIGNMENT_COARSE_FACTOR |
| 1625 | coarse_end = max( |
| 1626 | coarse_start + 50, |
| 1627 | end_ms // _ALIGNMENT_COARSE_FACTOR, |
| 1628 | ) |
| 1629 | coarse_template = audio_coarse[coarse_start:coarse_end] |
| 1630 | predicted_coarse = predicted_start_ms // _ALIGNMENT_COARSE_FACTOR |
| 1631 | coarse_window = _ALIGNMENT_SEARCH_WINDOW_MS // _ALIGNMENT_COARSE_FACTOR |
| 1632 | search_start = max( |
| 1633 | 0, |
| 1634 | predicted_coarse + coarse_start - coarse_window, |
| 1635 | ) |
| 1636 | search_end = min( |
| 1637 | len(video_coarse), |
| 1638 | predicted_coarse + coarse_start + coarse_window + len(coarse_template), |
| 1639 | ) |
| 1640 | coarse_index, _coarse_score = _best_correlation( |
| 1641 | video_coarse[search_start:search_end], |
| 1642 | coarse_template, |
| 1643 | ) |
| 1644 | coarse_audio_start_ms = ( |
| 1645 | search_start + coarse_index - coarse_start |
| 1646 | ) * _ALIGNMENT_COARSE_FACTOR |
| 1647 | |
| 1648 | fine_template = audio_fine[start_ms:end_ms] |
| 1649 | fine_search_start = max( |
| 1650 | 0, |
| 1651 | coarse_audio_start_ms + start_ms - _ALIGNMENT_REFINE_WINDOW_MS, |
| 1652 | ) |
| 1653 | fine_search_end = min( |
| 1654 | len(video_fine), |
| 1655 | coarse_audio_start_ms |
| 1656 | + start_ms |
| 1657 | + _ALIGNMENT_REFINE_WINDOW_MS |
| 1658 | + len(fine_template), |
| 1659 | ) |
| 1660 | fine_index, fine_score = _best_correlation( |
| 1661 | video_fine[fine_search_start:fine_search_end], |
| 1662 | fine_template, |
| 1663 | ) |
| 1664 | audio_start_ms = fine_search_start + fine_index - start_ms |
| 1665 | return audio_start_ms, fine_score |
| 1666 | |
| 1667 | |
| 1668 | def _align_audio_starts_to_video( |
| 1669 | *, |
| 1670 | slide_names: list[str], |
| 1671 | local_cues: dict[str, list[SubtitleCue] | None], |
| 1672 | theoretical_starts: list[int], |
| 1673 | audio_dir: Path, |
| 1674 | video_path: Path, |
| 1675 | ) -> tuple[list[int], list[float], list[Path]]: |
| 1676 | """Align every page narration to the audio track of an exported video.""" |
| 1677 | if not video_path.is_file(): |
| 1678 | raise FileNotFoundError(f"Exported video does not exist: {video_path}") |
| 1679 | ffmpeg_path = shutil.which("ffmpeg") |
| 1680 | if ffmpeg_path is None: |
| 1681 | raise RuntimeError( |
| 1682 | "Exported-video timeline calibration requires ffmpeg. " |
| 1683 | "Install ffmpeg and make it available on PATH." |
| 1684 | ) |
| 1685 | |
| 1686 | video_fine, video_coarse = _decode_audio_envelopes(video_path, ffmpeg_path) |
| 1687 | aligned_starts: list[int] = [] |
| 1688 | correlations: list[float] = [] |
| 1689 | audio_paths: list[Path] = [] |
| 1690 | |
| 1691 | for index, slide_name in enumerate(slide_names): |
| 1692 | audio_path = _find_audio(audio_dir, slide_name) |
| 1693 | audio_paths.append(audio_path) |
| 1694 | audio_fine, audio_coarse = _decode_audio_envelopes(audio_path, ffmpeg_path) |
| 1695 | slide_cues = local_cues.get(slide_name) |
| 1696 | if ( |
| 1697 | slide_cues |
| 1698 | and slide_cues[-1].end_ms |
| 1699 | > len(audio_fine) + _ALIGNMENT_END_TOLERANCE_MS |
| 1700 | ): |
| 1701 | raise ValueError( |
| 1702 | f"{slide_name}.srt ends after its narration audio: " |
| 1703 | f"cue end={_seconds_from_ms(slide_cues[-1].end_ms):.3f}s, " |
| 1704 | f"decoded audio={_seconds_from_ms(len(audio_fine)):.3f}s" |
| 1705 | ) |
| 1706 | if index == 0: |
| 1707 | predicted_start_ms = theoretical_starts[index] |
| 1708 | else: |
| 1709 | predicted_start_ms = ( |
| 1710 | aligned_starts[index - 1] |
| 1711 | + theoretical_starts[index] |
| 1712 | - theoretical_starts[index - 1] |
| 1713 | ) |
| 1714 | aligned_start_ms, correlation = _locate_audio_start( |
| 1715 | video_fine, |
| 1716 | video_coarse, |
| 1717 | audio_fine, |
| 1718 | audio_coarse, |
| 1719 | slide_cues, |
| 1720 | predicted_start_ms, |
| 1721 | ) |
| 1722 | if correlation < _ALIGNMENT_MIN_CORRELATION: |
| 1723 | raise ValueError( |
| 1724 | f"Exported-video audio match is unreliable for {slide_name}: " |
| 1725 | f"correlation={correlation:.3f}, " |
| 1726 | f"required>={_ALIGNMENT_MIN_CORRELATION:.2f}" |
| 1727 | ) |
| 1728 | if aligned_starts and aligned_start_ms <= aligned_starts[-1]: |
| 1729 | raise ValueError( |
| 1730 | f"Exported-video audio order is invalid at slide {slide_name}" |
| 1731 | ) |
| 1732 | final_audio_end_ms = aligned_start_ms + len(audio_fine) |
| 1733 | if ( |
| 1734 | final_audio_end_ms |
| 1735 | > len(video_fine) + _ALIGNMENT_END_TOLERANCE_MS |
| 1736 | ): |
| 1737 | raise ValueError( |
| 1738 | f"Exported video ends before the narration on slide {slide_name}: " |
| 1739 | f"audio end={_seconds_from_ms(final_audio_end_ms):.3f}s, " |
| 1740 | f"decoded video audio={_seconds_from_ms(len(video_fine)):.3f}s" |
| 1741 | ) |
| 1742 | aligned_starts.append(aligned_start_ms) |
| 1743 | correlations.append(correlation) |
| 1744 | |
| 1745 | return aligned_starts, correlations, audio_paths |
| 1746 | |
| 1747 | |
| 1748 | def calibrate_video_timeline( |
| 1749 | *, |
| 1750 | slide_names: list[str], |
| 1751 | pptx_path: Path, |
| 1752 | audio_dir: Path, |
| 1753 | video_path: Path, |
| 1754 | subtitle_dir: Path | None = None, |
| 1755 | ) -> VideoTimelineCalibration: |
| 1756 | """Calibrate PPTX slide starts against narration in an exported video. |
| 1757 | |
| 1758 | Page-local SRT improves the correlation template when available. Audio-only |
| 1759 | narration remains supported by matching each complete page track. |
| 1760 | """ |
| 1761 | if not slide_names: |
| 1762 | raise ValueError("Video timeline calibration requires at least one slide") |
| 1763 | if len(set(slide_names)) != len(slide_names): |
| 1764 | raise ValueError("Video timeline calibration slide names must be unique") |
| 1765 | |
| 1766 | timings = _read_powerpoint_timings(pptx_path, len(slide_names)) |
| 1767 | theoretical_starts, timeline_ms = _powerpoint_audio_starts(timings) |
| 1768 | local_cues: dict[str, list[SubtitleCue] | None] = {} |
| 1769 | for slide_name in slide_names: |
| 1770 | subtitle_path = ( |
| 1771 | subtitle_dir / f"{slide_name}.srt" |
| 1772 | if subtitle_dir is not None |
| 1773 | else None |
| 1774 | ) |
| 1775 | local_cues[slide_name] = ( |
| 1776 | _parse_srt(subtitle_path) |
| 1777 | if subtitle_path is not None and subtitle_path.is_file() |
| 1778 | else None |
| 1779 | ) |
| 1780 | |
| 1781 | aligned_starts, correlations, audio_paths = _align_audio_starts_to_video( |
| 1782 | slide_names=slide_names, |
| 1783 | local_cues=local_cues, |
| 1784 | theoretical_starts=theoretical_starts, |
| 1785 | audio_dir=audio_dir, |
| 1786 | video_path=video_path, |
| 1787 | ) |
| 1788 | |
| 1789 | slides: list[VideoSlideTiming] = [] |
| 1790 | powerpoint_slide_start_ms = 0 |
| 1791 | previous_video_slide_start_ms = -1 |
| 1792 | for ( |
| 1793 | slide_name, |
| 1794 | timing, |
| 1795 | powerpoint_narration_start_ms, |
| 1796 | video_narration_start_ms, |
| 1797 | correlation, |
| 1798 | audio_path, |
| 1799 | ) in zip( |
| 1800 | slide_names, |
| 1801 | timings, |
| 1802 | theoretical_starts, |
| 1803 | aligned_starts, |
| 1804 | correlations, |
| 1805 | audio_paths, |
| 1806 | ): |
| 1807 | raw_video_slide_start_ms = ( |
| 1808 | video_narration_start_ms |
| 1809 | - timing.transition_ms |
| 1810 | - timing.narration_delay_ms |
| 1811 | ) |
| 1812 | if raw_video_slide_start_ms < -_ALIGNMENT_END_TOLERANCE_MS: |
| 1813 | raise ValueError( |
| 1814 | f"Exported-video calibration places slide {slide_name} before " |
| 1815 | f"the video start: {raw_video_slide_start_ms}ms" |
| 1816 | ) |
| 1817 | video_slide_start_ms = max(0, raw_video_slide_start_ms) |
| 1818 | if video_slide_start_ms <= previous_video_slide_start_ms: |
| 1819 | raise ValueError( |
| 1820 | f"Exported-video slide order is invalid at {slide_name}" |
| 1821 | ) |
| 1822 | slides.append( |
| 1823 | VideoSlideTiming( |
| 1824 | slide_name=slide_name, |
| 1825 | transition_ms=timing.transition_ms, |
| 1826 | narration_delay_ms=timing.narration_delay_ms, |
| 1827 | advance_ms=timing.advance_ms, |
| 1828 | powerpoint_slide_start_ms=powerpoint_slide_start_ms, |
| 1829 | powerpoint_narration_start_ms=powerpoint_narration_start_ms, |
| 1830 | video_slide_start_ms=video_slide_start_ms, |
| 1831 | video_narration_start_ms=video_narration_start_ms, |
| 1832 | adjustment_ms=( |
| 1833 | video_narration_start_ms - powerpoint_narration_start_ms |
| 1834 | ), |
| 1835 | correlation=correlation, |
| 1836 | audio_path=audio_path, |
| 1837 | ) |
| 1838 | ) |
| 1839 | previous_video_slide_start_ms = video_slide_start_ms |
| 1840 | powerpoint_slide_start_ms += timing.transition_ms + timing.advance_ms |
| 1841 | |
| 1842 | return VideoTimelineCalibration( |
| 1843 | slides=tuple(slides), |
| 1844 | powerpoint_timeline_ms=timeline_ms, |
| 1845 | ) |
| 1846 | |
| 1847 | |
| 1848 | def _merge_subtitles_result( |
| 1849 | project_path: Path, |
| 1850 | *, |
| 1851 | pptx_path: Path, |
| 1852 | subtitle_dir: Path, |
| 1853 | output_path: Path, |
| 1854 | force: bool, |
| 1855 | audio_dir: Path | None = None, |
| 1856 | video_path: Path | None = None, |
| 1857 | ) -> SubtitleMergeResult: |
| 1858 | """Merge local SRT files on the PPTX or exported-video timeline.""" |
| 1859 | targets_by_slide, _anonymous_groups = scan_project_targets(project_path) |
| 1860 | slide_names = list(targets_by_slide) |
| 1861 | if not slide_names: |
| 1862 | raise ValueError(f"No SVG slides found under: {project_path / 'svg_output'}") |
| 1863 | |
| 1864 | local_subtitle_paths = [ |
| 1865 | subtitle_dir / f"{slide_name}.srt" |
| 1866 | for slide_name in slide_names |
| 1867 | ] |
| 1868 | _reject_output_alias( |
| 1869 | output_path, |
| 1870 | [pptx_path, *local_subtitle_paths, *([video_path] if video_path else [])], |
| 1871 | label="Merged subtitle", |
| 1872 | ) |
| 1873 | _require_replaceable(output_path, force) |
| 1874 | timings = _read_powerpoint_timings(pptx_path, len(slide_names)) |
| 1875 | theoretical_starts, timeline_ms = _powerpoint_audio_starts(timings) |
| 1876 | local_cues = { |
| 1877 | slide_name: _parse_srt(subtitle_dir / f"{slide_name}.srt") |
| 1878 | for slide_name in slide_names |
| 1879 | } |
| 1880 | video_adjustments: list[int] = [] |
| 1881 | correlations: list[float] = [] |
| 1882 | |
| 1883 | if video_path is None: |
| 1884 | audio_starts = theoretical_starts |
| 1885 | else: |
| 1886 | resolved_audio_dir = audio_dir or project_path / "audio" |
| 1887 | calibration = calibrate_video_timeline( |
| 1888 | slide_names=slide_names, |
| 1889 | pptx_path=pptx_path, |
| 1890 | audio_dir=resolved_audio_dir, |
| 1891 | video_path=video_path, |
| 1892 | subtitle_dir=subtitle_dir, |
| 1893 | ) |
| 1894 | if calibration.powerpoint_timeline_ms != timeline_ms: |
| 1895 | raise ValueError("Video calibration and subtitle timelines differ") |
| 1896 | audio_starts = [ |
| 1897 | slide.video_narration_start_ms |
| 1898 | for slide in calibration.slides |
| 1899 | ] |
| 1900 | correlations = [slide.correlation for slide in calibration.slides] |
| 1901 | audio_paths = [slide.audio_path for slide in calibration.slides] |
| 1902 | _reject_output_alias( |
| 1903 | output_path, |
| 1904 | audio_paths, |
| 1905 | label="Merged subtitle", |
| 1906 | ) |
| 1907 | video_adjustments = [ |
| 1908 | actual - theoretical |
| 1909 | for actual, theoretical in zip(audio_starts, theoretical_starts) |
| 1910 | ] |
| 1911 | |
| 1912 | merged_cues: list[SubtitleCue] = [] |
| 1913 | |
| 1914 | for slide_name, audio_start_ms, timing in zip( |
| 1915 | slide_names, |
| 1916 | audio_starts, |
| 1917 | timings, |
| 1918 | ): |
| 1919 | slide_cues = local_cues[slide_name] |
| 1920 | narration_window_ms = timing.advance_ms - timing.narration_delay_ms |
| 1921 | if slide_cues[-1].end_ms > narration_window_ms: |
| 1922 | raise ValueError( |
| 1923 | f"{slide_name}.srt ends at " |
| 1924 | f"{_seconds_from_ms(slide_cues[-1].end_ms):.3f}s, after the " |
| 1925 | "available narration window before PowerPoint advances at " |
| 1926 | f"{_seconds_from_ms(narration_window_ms):.3f}s" |
| 1927 | ) |
| 1928 | for cue in slide_cues: |
| 1929 | merged_cue = SubtitleCue( |
| 1930 | cue.start_ms + audio_start_ms, |
| 1931 | cue.end_ms + audio_start_ms, |
| 1932 | cue.text, |
| 1933 | ) |
| 1934 | if merged_cues and merged_cue.start_ms < merged_cues[-1].end_ms: |
| 1935 | raise ValueError( |
| 1936 | f"Video-calibrated subtitle overlap before slide {slide_name}" |
| 1937 | ) |
| 1938 | merged_cues.append(merged_cue) |
| 1939 | |
| 1940 | _atomic_write_text(output_path, _format_srt(merged_cues)) |
| 1941 | return SubtitleMergeResult( |
| 1942 | slide_count=len(slide_names), |
| 1943 | cue_count=len(merged_cues), |
| 1944 | powerpoint_timeline_ms=timeline_ms, |
| 1945 | minimum_video_adjustment_ms=( |
| 1946 | min(video_adjustments) if video_adjustments else None |
| 1947 | ), |
| 1948 | maximum_video_adjustment_ms=( |
| 1949 | max(video_adjustments) if video_adjustments else None |
| 1950 | ), |
| 1951 | minimum_video_correlation=( |
| 1952 | min(correlations) if correlations else None |
| 1953 | ), |
| 1954 | ) |
| 1955 | |
| 1956 | |
| 1957 | def merge_subtitles( |
| 1958 | project_path: Path, |
| 1959 | *, |
| 1960 | pptx_path: Path, |
| 1961 | subtitle_dir: Path, |
| 1962 | output_path: Path, |
| 1963 | force: bool, |
| 1964 | ) -> tuple[int, int, int]: |
| 1965 | """Merge local SRT files using timing values read from the final PPTX.""" |
| 1966 | result = _merge_subtitles_result( |
| 1967 | project_path, |
| 1968 | pptx_path=pptx_path, |
| 1969 | subtitle_dir=subtitle_dir, |
| 1970 | output_path=output_path, |
| 1971 | force=force, |
| 1972 | ) |
| 1973 | return ( |
| 1974 | result.slide_count, |
| 1975 | result.cue_count, |
| 1976 | result.powerpoint_timeline_ms, |
| 1977 | ) |
| 1978 | |
| 1979 | |
| 1980 | def merge_subtitles_to_video( |
| 1981 | project_path: Path, |
| 1982 | *, |
| 1983 | pptx_path: Path, |
| 1984 | subtitle_dir: Path, |
| 1985 | audio_dir: Path, |
| 1986 | video_path: Path, |
| 1987 | output_path: Path, |
| 1988 | force: bool, |
| 1989 | ) -> SubtitleMergeResult: |
| 1990 | """Merge local SRT files after calibrating page starts to an exported video.""" |
| 1991 | return _merge_subtitles_result( |
| 1992 | project_path, |
| 1993 | pptx_path=pptx_path, |
| 1994 | subtitle_dir=subtitle_dir, |
| 1995 | output_path=output_path, |
| 1996 | force=force, |
| 1997 | audio_dir=audio_dir, |
| 1998 | video_path=video_path, |
| 1999 | ) |
| 2000 | |
| 2001 | |
| 2002 | def build_parser() -> argparse.ArgumentParser: |
| 2003 | parser = argparse.ArgumentParser( |
| 2004 | description=( |
| 2005 | "Derive narrated object animation timings and merge page-local " |
| 2006 | "SRT files on PowerPoint's final timeline." |
| 2007 | ), |
| 2008 | formatter_class=argparse.RawDescriptionHelpFormatter, |
| 2009 | ) |
| 2010 | subparsers = parser.add_subparsers(dest="command", required=True) |
| 2011 | |
| 2012 | fingerprint = subparsers.add_parser( |
| 2013 | "fingerprint", |
| 2014 | help="print the SHA-256 used to bind a timing plan to page-local SRT files", |
| 2015 | ) |
| 2016 | fingerprint.add_argument("project_path", help="Project directory") |
| 2017 | fingerprint.add_argument( |
| 2018 | "--subtitle-dir", |
| 2019 | default=None, |
| 2020 | help="Page-local SRT directory; default: <project>/audio", |
| 2021 | ) |
| 2022 | |
| 2023 | animations = subparsers.add_parser( |
| 2024 | "animations", |
| 2025 | help="derive narration_animations.json from animations.json and page-local SRT", |
| 2026 | ) |
| 2027 | animations.add_argument("project_path", help="Project directory") |
| 2028 | animations.add_argument( |
| 2029 | "--animation-config", |
| 2030 | default=None, |
| 2031 | help="Read-only canonical config; default: <project>/animations.json", |
| 2032 | ) |
| 2033 | animations.add_argument( |
| 2034 | "--plan", |
| 2035 | default=None, |
| 2036 | help="Timing plan; default: <project>/narration_timing.json", |
| 2037 | ) |
| 2038 | animations.add_argument( |
| 2039 | "--subtitle-dir", |
| 2040 | default=None, |
| 2041 | help="Page-local SRT directory; default: <project>/audio", |
| 2042 | ) |
| 2043 | animations.add_argument( |
| 2044 | "--audio-dir", |
| 2045 | default=None, |
| 2046 | help="Narration audio directory; default: <project>/audio", |
| 2047 | ) |
| 2048 | animations.add_argument( |
| 2049 | "-o", |
| 2050 | "--output", |
| 2051 | default=None, |
| 2052 | help=( |
| 2053 | "Narration animation output; " |
| 2054 | "default: <project>/narration_animations.json" |
| 2055 | ), |
| 2056 | ) |
| 2057 | animations.add_argument( |
| 2058 | "--narration-padding", |
| 2059 | type=float, |
| 2060 | default=0.5, |
| 2061 | help="Seconds added after each narration before slide advance (default: 0.5)", |
| 2062 | ) |
| 2063 | animations.add_argument( |
| 2064 | "--narration-start-floor", |
| 2065 | type=float, |
| 2066 | default=DEFAULT_NARRATION_START_FLOOR, |
| 2067 | help=( |
| 2068 | "Minimum seconds from transition start to narration start; " |
| 2069 | "0 waits only for transition completion " |
| 2070 | f"(default: {DEFAULT_NARRATION_START_FLOOR:g})" |
| 2071 | ), |
| 2072 | ) |
| 2073 | animations.add_argument( |
| 2074 | "--force", |
| 2075 | action="store_true", |
| 2076 | help="Replace an existing narration_animations.json", |
| 2077 | ) |
| 2078 | |
| 2079 | subtitles = subparsers.add_parser( |
| 2080 | "subtitles", |
| 2081 | help="merge page-local SRT using timings read from a narrated PPTX", |
| 2082 | ) |
| 2083 | subtitles.add_argument("project_path", help="Project directory") |
| 2084 | subtitles.add_argument( |
| 2085 | "--pptx", |
| 2086 | required=True, |
| 2087 | help=( |
| 2088 | "Final narrated PPTX whose recorded timing values define the " |
| 2089 | "timeline; relative paths are resolved under the project" |
| 2090 | ), |
| 2091 | ) |
| 2092 | subtitles.add_argument( |
| 2093 | "--subtitle-dir", |
| 2094 | default=None, |
| 2095 | help="Page-local SRT directory; default: <project>/audio", |
| 2096 | ) |
| 2097 | subtitles.add_argument( |
| 2098 | "--video", |
| 2099 | default=None, |
| 2100 | help=( |
| 2101 | "PowerPoint-exported video whose audio track calibrates page starts; " |
| 2102 | "relative paths are resolved under the project" |
| 2103 | ), |
| 2104 | ) |
| 2105 | subtitles.add_argument( |
| 2106 | "--audio-dir", |
| 2107 | default=None, |
| 2108 | help=( |
| 2109 | "Page-local narration audio used with --video; " |
| 2110 | "default: <project>/audio" |
| 2111 | ), |
| 2112 | ) |
| 2113 | subtitles.add_argument( |
| 2114 | "-o", |
| 2115 | "--output", |
| 2116 | default=None, |
| 2117 | help="Merged SRT output; default: <project>/audio/total.srt", |
| 2118 | ) |
| 2119 | subtitles.add_argument( |
| 2120 | "--force", |
| 2121 | action="store_true", |
| 2122 | help="Replace an existing merged SRT", |
| 2123 | ) |
| 2124 | return parser |
| 2125 | |
| 2126 | |
| 2127 | def main(argv: list[str] | None = None) -> int: |
| 2128 | parser = build_parser() |
| 2129 | args = parser.parse_args(argv) |
| 2130 | project_path = Path(args.project_path).resolve() |
| 2131 | if not project_path.is_dir(): |
| 2132 | parser.error(f"Project path does not exist: {project_path}") |
| 2133 | |
| 2134 | try: |
| 2135 | if args.command == "fingerprint": |
| 2136 | subtitle_dir = _project_path( |
| 2137 | project_path, |
| 2138 | args.subtitle_dir, |
| 2139 | Path("audio"), |
| 2140 | ) |
| 2141 | slide_names = _page_subtitle_names(subtitle_dir) |
| 2142 | for slide_name in slide_names: |
| 2143 | _parse_srt(subtitle_dir / f"{slide_name}.srt") |
| 2144 | print(_subtitle_fingerprint(slide_names, subtitle_dir)) |
| 2145 | return 0 |
| 2146 | |
| 2147 | if args.command == "animations": |
| 2148 | canonical_path = _project_path( |
| 2149 | project_path, |
| 2150 | args.animation_config, |
| 2151 | Path("animations.json"), |
| 2152 | ) |
| 2153 | plan_path = _project_path( |
| 2154 | project_path, |
| 2155 | args.plan, |
| 2156 | Path("narration_timing.json"), |
| 2157 | ) |
| 2158 | subtitle_dir = _project_path( |
| 2159 | project_path, |
| 2160 | args.subtitle_dir, |
| 2161 | Path("audio"), |
| 2162 | ) |
| 2163 | audio_dir = _project_path( |
| 2164 | project_path, |
| 2165 | args.audio_dir, |
| 2166 | Path("audio"), |
| 2167 | ) |
| 2168 | output_path = _project_path( |
| 2169 | project_path, |
| 2170 | args.output, |
| 2171 | Path("narration_animations.json"), |
| 2172 | ) |
| 2173 | result = rebuild_animations( |
| 2174 | project_path, |
| 2175 | canonical_path=canonical_path, |
| 2176 | plan_path=plan_path, |
| 2177 | subtitle_dir=subtitle_dir, |
| 2178 | audio_dir=audio_dir, |
| 2179 | output_path=output_path, |
| 2180 | narration_padding=args.narration_padding, |
| 2181 | narration_start_floor=args.narration_start_floor, |
| 2182 | force=args.force, |
| 2183 | ) |
| 2184 | print(f"Narration animation config written: {output_path}") |
| 2185 | print( |
| 2186 | f"Slides: {result.slide_count}; groups: {result.group_count}; " |
| 2187 | f"SRT-anchored: {result.anchored_count}; " |
| 2188 | f"canonical-delay fallback: {result.fallback_count}; " |
| 2189 | f"SVG fallback slides: {result.svg_fallback_slide_count}; " |
| 2190 | f"unused cues: {result.ignored_cue_count}" |
| 2191 | ) |
| 2192 | return 0 |
| 2193 | |
| 2194 | pptx_path = _project_input_path(project_path, args.pptx).resolve() |
| 2195 | subtitle_dir = _project_path( |
| 2196 | project_path, |
| 2197 | args.subtitle_dir, |
| 2198 | Path("audio"), |
| 2199 | ) |
| 2200 | audio_dir = _project_path( |
| 2201 | project_path, |
| 2202 | args.audio_dir, |
| 2203 | Path("audio"), |
| 2204 | ) |
| 2205 | video_path = ( |
| 2206 | _project_input_path(project_path, args.video).resolve() |
| 2207 | if args.video |
| 2208 | else None |
| 2209 | ) |
| 2210 | output_path = _project_path( |
| 2211 | project_path, |
| 2212 | args.output, |
| 2213 | Path("audio/total.srt"), |
| 2214 | ) |
| 2215 | result = _merge_subtitles_result( |
| 2216 | project_path, |
| 2217 | pptx_path=pptx_path, |
| 2218 | subtitle_dir=subtitle_dir, |
| 2219 | audio_dir=audio_dir, |
| 2220 | video_path=video_path, |
| 2221 | output_path=output_path, |
| 2222 | force=args.force, |
| 2223 | ) |
| 2224 | print(f"Merged subtitle written: {output_path}") |
| 2225 | print( |
| 2226 | f"Slides: {result.slide_count}; cues: {result.cue_count}; " |
| 2227 | "PowerPoint timeline: " |
| 2228 | f"{_seconds_from_ms(result.powerpoint_timeline_ms):.3f}s" |
| 2229 | ) |
| 2230 | if result.minimum_video_correlation is not None: |
| 2231 | print( |
| 2232 | "Exported-video calibration: page adjustment " |
| 2233 | f"{_seconds_from_ms(result.minimum_video_adjustment_ms or 0):+.3f}s " |
| 2234 | "to " |
| 2235 | f"{_seconds_from_ms(result.maximum_video_adjustment_ms or 0):+.3f}s; " |
| 2236 | f"minimum correlation: {result.minimum_video_correlation:.3f}" |
| 2237 | ) |
| 2238 | return 0 |
| 2239 | except ( |
| 2240 | ET.ParseError, |
| 2241 | OSError, |
| 2242 | OverflowError, |
| 2243 | ValueError, |
| 2244 | RuntimeError, |
| 2245 | zipfile.BadZipFile, |
| 2246 | ) as exc: |
| 2247 | print(f"Error: {exc}", file=sys.stderr) |
| 2248 | return 1 |
| 2249 | |
| 2250 | |
| 2251 | if __name__ == "__main__": |
| 2252 | raise SystemExit(main()) |
| 2253 |