Skip to content

AI turns documents or topics into real, native PowerPoint decks—with native shapes, transitions and animations, data-backed charts and tables on demand, audio narration from speaker notes, and support for your own .pptx templates. · by Hugo He

fix: round-trip quality gate reads a validation/ report bound to the authoring-SVG fingerprint; checker enforces the exporter's one-text semantic-shape contract early; --no-notes removes every notes part; negative edge-tts values parse with a space

hugohe373bc51a38007
diff --git a/skills/ppt-master/scripts/authoring_roundtrip.py b/skills/ppt-master/scripts/authoring_roundtrip.py
index c4db4c80..33f3142f 100644
--- a/skills/ppt-master/scripts/authoring_roundtrip.py
+++ b/skills/ppt-master/scripts/authoring_roundtrip.py
@@ -37,6 +37,7 @@ from extract_svg_assets import (
)
from hyperlink_contract import ADOPTED_SOURCE_LINK_ATTR, SHAPE_HYPERLINK_ATTR, SOURCE_HREF_ATTR
from pptx_workspace import (
+ AUTHORING_SVG_FLAT_DIR,
ROUNDTRIP_MANIFEST_PATH,
ROUNDTRIP_PAGE_PLAN_PATH,
)
@@ -88,6 +89,27 @@ class AuthoringRoundtripError(RuntimeError):
"""Reject stale, ambiguous, or incomplete authoring round-trip input."""
+def roundtrip_source_fingerprint(project_path: Path) -> dict[str, object]:
+ """Bind validation and export to the authoring SVGs and optional page plan."""
+ paths = list((project_path / AUTHORING_SVG_FLAT_DIR).glob('*.svg'))
+ plan_path = project_path / ROUNDTRIP_PAGE_PLAN_PATH
+ if plan_path.is_file():
+ paths.append(plan_path)
+ files: list[dict[str, str]] = []
+ aggregate = hashlib.sha256()
+ for path in sorted(paths, key=lambda item: item.relative_to(project_path).as_posix()):
+ name = path.relative_to(project_path).as_posix()
+ digest = hashlib.sha256(path.read_bytes()).hexdigest()
+ files.append({'file': name, 'sha256': digest})
+ aggregate.update(name.encode('utf-8') + b'\0' + digest.encode('ascii') + b'\n')
+ return {
+ 'algorithm': 'sha256',
+ 'digest': aggregate.hexdigest(),
+ 'file_count': len(files),
+ 'files': files,
+ }
+
+
@dataclass(frozen=True)
class SourceRefRecord:
source_path: tuple[int, ...]
diff --git a/skills/ppt-master/scripts/docs/narration.md b/skills/ppt-master/scripts/docs/narration.md
index 7950c70e..3eecfd16 100644
--- a/skills/ppt-master/scripts/docs/narration.md
+++ b/skills/ppt-master/scripts/docs/narration.md
@@ -13,7 +13,7 @@ Tool behavior behind the [`generate-audio`](../../workflows/stages/generate-audi
```bash
python3 skills/ppt-master/scripts/notes_to_audio.py --list-voices --locale <locale> # edge
python3 skills/ppt-master/scripts/notes_to_audio.py --provider <elevenlabs|minimax|qwen|cosyvoice> --list-voices
-python3 skills/ppt-master/scripts/notes_to_audio.py <project_path> --voice <ShortName> --rate=<rate> [--concurrency <N>]
+python3 skills/ppt-master/scripts/notes_to_audio.py <project_path> --voice <ShortName> --rate=<rate> [--pitch=<pitch>] [--volume=<volume>] [--concurrency <N>]
python3 skills/ppt-master/scripts/notes_to_audio.py <project_path> --provider elevenlabs --voice-id <id> --elevenlabs-model eleven_multilingual_v2
python3 skills/ppt-master/scripts/notes_to_audio.py <project_path> --provider minimax --voice-id <id> --minimax-model speech-2.8-hd
python3 skills/ppt-master/scripts/notes_to_audio.py <project_path> --provider qwen --voice-id <voice> --qwen-model qwen3-tts-flash --qwen-language-type Chinese
@@ -26,6 +26,8 @@ python3 skills/ppt-master/scripts/notes_to_audio.py <project_path> --provider co
- **Edge SRT**: MP3 and page SRT come from the same `edge-tts` stream using `WordBoundary` timing. Sentence-ending punctuation always closes a cue; text over the 20-visible-character default (`--subtitle-max-chars`) splits at commas, semicolons, or colons, then at the nearest word boundary. Adjacent overlap up to 100 ms moves the later cue start to the previous end; larger overlap fails. Each SRT uses a page-local timeline starting at `00:00:00,000` including leading silence. Default concurrency is three slide pairs (`--concurrency 1` for serial troubleshooting); cloud providers stay serial.
- **Provider timing**: MiniMax reads word timing from its synchronous subtitle file (a normalized number such as `42` comes back as one row per spoken syllable sharing one original-text span; the adapter merges rows by that span, never by text alone); ElevenLabs uses `/with-timestamps` with original-text character alignment; CosyVoice enables HTTP streaming plus `word_timestamp_enabled` and uses the final audio URL and word timing — unsupported model/voice pairs fail without replacing the prior pair unless `--cosyvoice-audio-only` was explicit (model and voice families must match; cloned voices need a supported v3.5/v3/v2 model or a timestamp-supported system voice); Qwen exposes no timing and never gets estimated SRT. Provider-timed paths share punctuation-first, `--subtitle-max-chars`-bounded regrouping, exact-text validation, and rollback-safe pair publication.
+Signed Edge values accept both equals and space forms: `--rate=-5%` / `--rate -5%`, `--pitch=-2Hz` / `--pitch -2Hz`, and `--volume=-10%` / `--volume -10%`. The examples use equals for clarity; omitted pitch and volume remain `+0Hz` and `+0%`. These three options are ignored by cloud providers.
+
## `narration_sync.py`
```bash
diff --git a/skills/ppt-master/scripts/docs/svg-pipeline.md b/skills/ppt-master/scripts/docs/svg-pipeline.md
index 5566c5de..5edf9ec9 100644
--- a/skills/ppt-master/scripts/docs/svg-pipeline.md
+++ b/skills/ppt-master/scripts/docs/svg-pipeline.md
@@ -185,7 +185,7 @@ notes and an absent file keeps them. Deleting inherited source notes only on a
copy is not supported in v1. The same output-stem rule applies to narration
audio.
-When a round-trip recorded-narration export omits `--animation-config`, it uses the workspace `animations.json` when present and otherwise applies no sidecar while preserving source motion.
+When a round-trip recorded-narration export omits `--animation-config`, it reads the workspace `animations.json` the importer wrote; that sidecar is part of the workspace contract, so deleting it fails export rather than restoring source motion.
Narration audio is keyed by the output SVG stem. A copied output page uses its
own stem-keyed notes when present and otherwise inherits the declared source
@@ -210,7 +210,7 @@ while order, notes, or motion may change; `rebuilt` means visible authoring or
one of its referenced resources changed.
Without `-o`, round-trip export names the deck `<workspace-directory-name>_<timestamp>[<flavor-suffix>].pptx` under `exports/`.
-Before export, run `python3 scripts/svg_quality_checker.py <workspace> --roundtrip`
+Before export, run `python3 scripts/svg_quality_checker.py <workspace> --roundtrip --json`
as the round-trip text-capacity gate. It resolves the output roster from
`authoring-svg-flat/` and optional `page_plan.json`, then applies the shared
font-family, font-size, text-width, and canvas metrics only to new text or
@@ -222,6 +222,24 @@ leaving the page canvas remain blocking. Other advisories remain non-blocking.
Unchanged source refs, source proxies, and generated-project-only spec,
template, canonical-authoring, and resource-manifest checks are excluded.
+The checker also enforces the exporter's semantic shape text contract on
+converted content: a `data-pptx-semantic-object="shape"` group may contain at
+most one direct `<text>`, with `data-pptx-frame` on its owner. Multiple lines
+belong in that component's `<tspan>` paragraphs. Ordinary groups are unaffected;
+unchanged native source objects remain excluded. Generate checks the same
+contract.
+
+`--json` writes `validation/svg_quality_report.json`. Both tools fingerprint
+`authoring-svg-flat/*.svg` and the optional `page_plan.json`; changing either
+after validation makes the export receipt `quality_gate=stale`. A current
+report with no blocking errors yields `quality_gate=passed`. Without a report,
+export remains allowed and reports `quality_gate=not-provided`.
+
+`--roundtrip --no-notes` removes all speaker notes slide/master parts, their
+relationships, and Content-Type overrides, including on the identity export
+without a page plan. `pptx_delivery_check.py` reports unreferenced notes parts
+as an advisory.
+
Regenerate the summary after direct edits that do not pass through one of the
in-place normalization tools:
diff --git a/skills/ppt-master/scripts/notes_to_audio.py b/skills/ppt-master/scripts/notes_to_audio.py
index bc07147b..d105d0fa 100644
--- a/skills/ppt-master/scripts/notes_to_audio.py
+++ b/skills/ppt-master/scripts/notes_to_audio.py
@@ -283,6 +283,8 @@ def _provider_manifest_details(
if backend.provider == "edge":
return "edge-tts", {
"rate": args.rate,
+ "pitch": args.pitch,
+ "volume": args.volume,
}
if backend.provider == "elevenlabs":
return args.elevenlabs_model, {
@@ -380,6 +382,8 @@ async def _generate_edge_jobs(
*,
voice: str,
rate: str,
+ pitch: str = "+0Hz",
+ volume: str = "+0%",
subtitle_max_chars: int,
concurrency: int,
) -> list[BaseException | None]:
@@ -393,6 +397,8 @@ async def _generate_edge_jobs(
job.output_path,
voice=voice,
rate=rate,
+ pitch=pitch,
+ volume=volume,
subtitle_path=subtitle_dir / f"{job.output_path.stem}.srt",
subtitle_max_chars=subtitle_max_chars,
)
@@ -407,7 +413,26 @@ async def _generate_edge_jobs(
]
-def main() -> int:
+def _normalize_voice_option_args(argv: list[str]) -> list[str]:
+ """Keep signed voice values from being interpreted as argparse options."""
+ normalized: list[str] = []
+ index = 0
+ while index < len(argv):
+ option = argv[index]
+ if (
+ option in {"--rate", "--pitch", "--volume"}
+ and index + 1 < len(argv)
+ and re.fullmatch(r"-\d+(?:\.\d+)?(?:%|Hz)?", argv[index + 1])
+ ):
+ normalized.append(f"{option}={argv[index + 1]}")
+ index += 2
+ else:
+ normalized.append(option)
+ index += 1
+ return normalized
+
+
+def main(argv: list[str] | None = None) -> int:
_load_tts_env_file()
parser = argparse.ArgumentParser(
@@ -437,6 +462,14 @@ def main() -> int:
default="+0%",
help='edge-tts speaking rate, e.g. "+0%%", "-10%%", "+15%%" (default: +0%%). Ignored by cloud providers.',
)
+ parser.add_argument(
+ "--pitch", default="+0Hz",
+ help="edge-tts pitch adjustment (default: +0Hz). Ignored by cloud providers.",
+ )
+ parser.add_argument(
+ "--volume", default="+0%",
+ help="edge-tts volume adjustment (default: +0%%). Ignored by cloud providers.",
+ )
parser.add_argument(
"--concurrency",
type=int,
@@ -560,12 +593,9 @@ def main() -> int:
parser.add_argument("--list-common-voices", action="store_true", help="print a curated voice list and exit")
parser.add_argument("--list-voices", action="store_true", help="query provider voices and exit")
parser.add_argument("--locale", default=None, help='filter --list-voices by locale, e.g. "zh-CN"')
- raw_args = list(sys.argv[1:])
- # argparse otherwise interprets a separate negative percentage as an option.
- for index in range(len(raw_args) - 2, -1, -1):
- if raw_args[index] == "--rate" and re.fullmatch(r"-\d+(?:\.\d+)?%", raw_args[index + 1]):
- raw_args[index:index + 2] = [f"--rate={raw_args[index + 1]}"]
- args = parser.parse_args(raw_args)
+ args = parser.parse_args(_normalize_voice_option_args(
+ sys.argv[1:] if argv is None else argv
+ ))
if args.list_common_voices:
backend_edge.print_common_voices()
@@ -747,6 +777,8 @@ def main() -> int:
subtitle_dir,
voice=args.voice,
rate=args.rate,
+ pitch=args.pitch,
+ volume=args.volume,
subtitle_max_chars=args.subtitle_max_chars,
concurrency=args.concurrency,
))
diff --git a/skills/ppt-master/scripts/pptx_delivery_check.py b/skills/ppt-master/scripts/pptx_delivery_check.py
index 5a628167..76e0dd50 100644
--- a/skills/ppt-master/scripts/pptx_delivery_check.py
+++ b/skills/ppt-master/scripts/pptx_delivery_check.py
@@ -883,6 +883,20 @@ def audit_pptx_delivery(path: str | Path) -> dict[str, object]:
reference_counts, relationship_types, external_media = (
_relationship_inventory(archive, part_names, errors)
)
+ orphan_notes = sorted(
+ name for name in part_names
+ if _content_type_for_part(name, defaults, overrides) in {
+ 'application/vnd.openxmlformats-officedocument.presentationml.notesSlide+xml',
+ 'application/vnd.openxmlformats-officedocument.presentationml.notesMaster+xml',
+ }
+ and reference_counts.get(canonical_opc_part_path(name) or '', 0) == 0
+ )
+ if orphan_notes:
+ advisories.append(_issue(
+ "orphan_notes_parts",
+ "Unreferenced speaker notes parts remain in the package; remove them if no longer needed.",
+ parts=orphan_notes,
+ ))
media_infos = [
info_by_name[name]
diff --git a/skills/ppt-master/scripts/svg_authoring_view.py b/skills/ppt-master/scripts/svg_authoring_view.py
index 0b0e61b0..fff8771c 100644
--- a/skills/ppt-master/scripts/svg_authoring_view.py
+++ b/skills/ppt-master/scripts/svg_authoring_view.py
@@ -202,6 +202,31 @@ def _local_name(name: object) -> str:
return name.rsplit("}", 1)[-1] if isinstance(name, str) else ""
+def semantic_shape_text_component(
+ shape: ET.Element,
+ *,
+ included_text_ids: set[int] | None = None,
+) -> ET.Element | None:
+ """Validate text compiled into one semantic shape, excluding native restores."""
+ def included(element: ET.Element) -> bool:
+ return (
+ element.tag.replace(f'{{{SVG_NS}}}', '') == 'text'
+ and (included_text_ids is None or id(element) in included_text_ids)
+ )
+
+ texts = [child for child in shape if included(child)]
+ nested_texts = [child for child in shape.iter() if included(child)]
+ if nested_texts != texts:
+ raise ValueError('Semantic shape text must be one direct SVG text component')
+ if not texts:
+ return None
+ if len(texts) != 1:
+ raise ValueError('Semantic shape requires at most one paragraph-based text component')
+ if shape.get('data-pptx-frame') is None:
+ raise ValueError('Semantic shape text requires data-pptx-frame on its owner')
+ return texts[0]
+
+
@dataclass
class SourceReference:
source_ref: str
diff --git a/skills/ppt-master/scripts/svg_quality/checker.py b/skills/ppt-master/scripts/svg_quality/checker.py
index 21e38e21..658b7e88 100644
--- a/skills/ppt-master/scripts/svg_quality/checker.py
+++ b/skills/ppt-master/scripts/svg_quality/checker.py
@@ -33,6 +33,11 @@ from pptx_workspace import (
)
from slide_roster import discover_slide_svgs
from svg_authoring_contract import canonical_authoring_errors
+from svg_authoring_view import (
+ SEMANTIC_OBJECT_ATTRIBUTE,
+ SEMANTIC_SHAPE_KIND,
+ semantic_shape_text_component,
+)
from . import svg_contracts
from .xml_support import (
@@ -1175,6 +1180,7 @@ class SVGQualityChecker:
self.quick_generate = quick_generate
self.canonical_authoring = canonical_authoring
self.results = []
+ self._roundtrip_source_fingerprint: Dict[str, object] | None = None
self.summary = {
'total': 0,
'passed': 0,
@@ -1399,6 +1405,7 @@ class SVGQualityChecker:
# 5. Check text wrapping methods
self._check_text_elements(content, root, result)
+ self._check_semantic_shape_text(root, result)
# 5b. Validate native hyperlink targets and carrier structure.
self._check_hyperlinks(root, result)
@@ -1548,6 +1555,7 @@ class SVGQualityChecker:
_load_documents,
_load_page_plan,
_parse_svg,
+ roundtrip_source_fingerprint,
)
project_path = Path(workspace).resolve()
@@ -1569,6 +1577,7 @@ class SVGQualityChecker:
return []
try:
+ self._roundtrip_source_fingerprint = roundtrip_source_fingerprint(project_path)
source_root, source_proxy_dir, documents, _, _ = _load_documents(
project_path,
authoring_dir,
@@ -1676,6 +1685,9 @@ class SVGQualityChecker:
)
)
result['info']['edited_text_elements'] = len(included_text_ids)
+ self._check_semantic_shape_text(
+ root, result, included_text_ids=included_text_ids,
+ )
if included_text_ids:
scoped_content = self._roundtrip_text_scope_content(
root,
@@ -1716,6 +1728,27 @@ class SVGQualityChecker:
result['passed'] = False
return self._record_result(result)
+ @staticmethod
+ def _check_semantic_shape_text(
+ root: ET.Element,
+ result: Dict,
+ *,
+ included_text_ids: set[int] | None = None,
+ ) -> None:
+ """Report the exporter's semantic-shape text contract before conversion."""
+ parents = {id(child): parent for parent in root.iter() for child in parent}
+ for shape in root.iter(f'{{{SVG_NS}}}g'):
+ if shape.get(SEMANTIC_OBJECT_ATTRIBUTE) != SEMANTIC_SHAPE_KIND:
+ continue
+ if _svg_hidden_reason is not None and _svg_hidden_reason(
+ shape, parents, preserve_native_carriers=True,
+ ) is not None:
+ continue
+ try:
+ semantic_shape_text_component(shape, included_text_ids=included_text_ids)
+ except ValueError as exc:
+ result['errors'].append(f"{_element_label(shape)}: {exc}")
+
@staticmethod
def _roundtrip_text_diff_ids(
root: ET.Element,
@@ -9848,7 +9881,11 @@ class SVGQualityChecker:
'schema': 'ppt-master.svg-quality-report.v1',
'stage': stage,
'target': str(Path(target).resolve()),
- 'source_fingerprint': _quality_source_fingerprint(self.results),
+ 'source_fingerprint': (
+ self._roundtrip_source_fingerprint
+ if self._roundtrip_source_fingerprint is not None
+ else _quality_source_fingerprint(self.results)
+ ),
'summary': dict(self.summary),
'issue_types': dict(sorted(self.issue_types.items())),
'categories': {
diff --git a/skills/ppt-master/scripts/svg_quality/cli.py b/skills/ppt-master/scripts/svg_quality/cli.py
index 2d6aec33..fc661777 100644
--- a/skills/ppt-master/scripts/svg_quality/cli.py
+++ b/skills/ppt-master/scripts/svg_quality/cli.py
@@ -90,6 +90,8 @@ def _default_json_report_path(
checker: SVGQualityChecker,
target: str,
stage: str,
+ *,
+ roundtrip: bool = False,
) -> Path:
"""Choose a stage-specific report path without overwriting the final gate."""
target_path = Path(target)
@@ -100,6 +102,8 @@ def _default_json_report_path(
"early": "svg_quality_early_report.json",
"page": "svg_quality_page_report.json",
}[stage]
+ if roundtrip:
+ return target_path / "validation" / report_name
if (
(project_path / "svg_output").is_dir()
or (project_path / "design_spec.md").is_file()
@@ -334,7 +338,9 @@ def main() -> None:
sys.exit(1)
json_output = Path(sys.argv[idx + 1])
else:
- json_output = _default_json_report_path(checker, target, stage)
+ json_output = _default_json_report_path(
+ checker, target, stage, roundtrip=roundtrip,
+ )
checker.export_json_report(
str(json_output),
target=target,
diff --git a/skills/ppt-master/scripts/svg_to_pptx/drawingml/converter.py b/skills/ppt-master/scripts/svg_to_pptx/drawingml/converter.py
index 7d9e3f6d..4fe85259 100644
--- a/skills/ppt-master/scripts/svg_to_pptx/drawingml/converter.py
+++ b/skills/ppt-master/scripts/svg_to_pptx/drawingml/converter.py
@@ -36,6 +36,7 @@ from resource_paths import icon_dir_for_project
from svg_authoring_view import (
SEMANTIC_OBJECT_ATTRIBUTE,
SEMANTIC_SHAPE_KIND,
+ semantic_shape_text_component,
)
from svg_compatibility import normalize_single_child_group_filters
@@ -751,33 +752,15 @@ def _semantic_shape_text_body(
preserved = _decode_unchanged_txbody(shape, metadata)
if preserved is not None:
return preserved[0]
- texts = [
- child
- for child in shape
- if child.tag.replace(f'{{{SVG_NS}}}', '') == 'text'
- ]
- nested_texts = [
- child
- for child in shape.iter()
- if child.tag.replace(f'{{{SVG_NS}}}', '') == 'text'
- ]
- if nested_texts != texts:
- raise SvgNativeConversionError(
- 'Semantic shape text must be one direct SVG text component'
- )
- if not texts:
+ try:
+ component = semantic_shape_text_component(shape)
+ except ValueError as exc:
+ raise SvgNativeConversionError(str(exc)) from exc
+ if component is None:
return None
- if len(texts) != 1:
- raise SvgNativeConversionError(
- 'Semantic shape requires at most one paragraph-based text component'
- )
frame = shape.get('data-pptx-frame')
- if frame is None:
- raise SvgNativeConversionError(
- 'Semantic shape text requires data-pptx-frame on its owner'
- )
- text = copy.deepcopy(texts[0])
+ text = copy.deepcopy(component)
text.set('data-pptx-frame', frame)
for name in (
'data-pptx-shape-id',
diff --git a/skills/ppt-master/scripts/svg_to_pptx/pptx_package/builder.py b/skills/ppt-master/scripts/svg_to_pptx/pptx_package/builder.py
index 4a9f1153..776ecaee 100644
--- a/skills/ppt-master/scripts/svg_to_pptx/pptx_package/builder.py
+++ b/skills/ppt-master/scripts/svg_to_pptx/pptx_package/builder.py
@@ -1896,6 +1896,71 @@ def _unused_part_path(directory: Path, filename: str) -> Path:
return candidate
+def _remove_all_notes_parts(extract_dir: Path) -> None:
+ """Remove notes slides, masters, their relationships, and type overrides."""
+ content_types_path = extract_dir / '[Content_Types].xml'
+ content_tree = ET.parse(content_types_path)
+ content_root = content_tree.getroot()
+ notes_content_types = {
+ 'application/vnd.openxmlformats-officedocument.presentationml.notesSlide+xml',
+ _NOTES_MASTER_CONTENT_TYPE,
+ }
+ notes_parts = {
+ _canonical_opc_part_path(node.get('PartName', '').lstrip('/'))
+ for node in content_root
+ if node.get('ContentType') in notes_content_types
+ } - {None}
+ notes_files = [
+ path for path in extract_dir.rglob('*')
+ if path.is_file() and (
+ _canonical_opc_part_path(path.relative_to(extract_dir).as_posix()) in notes_parts
+ or path.relative_to(extract_dir).as_posix().startswith((
+ 'ppt/notesSlides/', 'ppt/notesMasters/',
+ ))
+ )
+ ]
+ notes_parts.update(
+ _canonical_opc_part_path(path.relative_to(extract_dir).as_posix())
+ for path in notes_files
+ )
+ for rels_path in extract_dir.rglob('*.rels'):
+ tree = ET.parse(rels_path)
+ root = tree.getroot()
+ removed = False
+ rels_name = rels_path.relative_to(extract_dir).as_posix()
+ for relationship in list(root):
+ if (
+ relationship.get('Type') in {NOTES_SLIDE_REL_TYPE, _NOTES_MASTER_REL_TYPE}
+ or (
+ relationship.get('TargetMode', '').lower() != 'external'
+ and _resolve_internal_opc_target(
+ rels_name, relationship.get('Target', ''),
+ ) in notes_parts
+ )
+ ):
+ root.remove(relationship)
+ removed = True
+ if removed:
+ _write_xml_tree(rels_path, tree)
+ presentation = extract_dir / 'ppt/presentation.xml'
+ tree = ET.parse(presentation)
+ notes_masters = tree.getroot().find(f'{{{PML_NS}}}notesMasterIdLst')
+ if notes_masters is not None:
+ tree.getroot().remove(notes_masters)
+ _write_xml_tree(presentation, tree)
+ for path in notes_files:
+ rels_path = _relationships_path_for_part(extract_dir, path.relative_to(extract_dir).as_posix())
+ rels_path.unlink(missing_ok=True)
+ path.unlink(missing_ok=True)
+ removed = False
+ for node in list(content_root):
+ if _canonical_opc_part_path(node.get('PartName', '').lstrip('/')) in notes_parts:
+ content_root.remove(node)
+ removed = True
+ if removed:
+ _write_xml_tree(content_types_path, content_tree)
+
+
def _apply_slide_notes(
extract_dir: Path,
rels_path: Path,
@@ -8371,6 +8436,9 @@ def create_pptx_with_native_svg(
f"{reinjected_resources} source package part(s)"
)
+ if not enable_notes:
+ _remove_all_notes_parts(extract_dir)
+
rels_problems = verify_internal_relationships(extract_dir)
if rels_problems:
details = '\n'.join(f' - {p}' for p in rels_problems)
diff --git a/skills/ppt-master/scripts/svg_to_pptx/pptx_package/cli.py b/skills/ppt-master/scripts/svg_to_pptx/pptx_package/cli.py
index c34bc45d..22112c38 100644
--- a/skills/ppt-master/scripts/svg_to_pptx/pptx_package/cli.py
+++ b/skills/ppt-master/scripts/svg_to_pptx/pptx_package/cli.py
@@ -31,6 +31,7 @@ from authoring_roundtrip import ( # noqa: E402
RoundtripPage,
is_flat_authoring_bundle,
materialize_flat_authoring_roundtrip,
+ roundtrip_source_fingerprint,
)
from console_encoding import configure_utf8_stdio # noqa: E402
from language_tags import ( # noqa: E402
@@ -1421,7 +1422,11 @@ def _write_postflight_report(
f"{package['slides']} != {len(svg_files)}"
)
source_audit = _source_resource_audit(svg_files)
- source_fingerprint = _svg_source_fingerprint(svg_files)
+ source_fingerprint = (
+ roundtrip_source_fingerprint(project_path)
+ if authoring_roundtrip is not None
+ else _svg_source_fingerprint(svg_files)
+ )
quality = _quality_report_context(project_path, source_fingerprint)
quality_gate, introduced_warning_count = _quality_gate_status(quality)
unresolved_tokens = source_audit['unresolved_template_tokens']
@@ -1573,9 +1578,14 @@ def _load_deck_motion_handoff(
if report.get('status') not in {'passed', 'passed-with-warnings'}:
raise ValueError('deck-motion handoff report is not a successful export')
source = _as_dict(report.get('source'))
- if source.get('fingerprint') != _svg_source_fingerprint(svg_files):
+ current_fingerprint = (
+ roundtrip_source_fingerprint(project_path)
+ if source.get('authoring_roundtrip') is not None
+ else _svg_source_fingerprint(svg_files)
+ )
+ if source.get('fingerprint') != current_fingerprint:
raise ValueError(
- 'deck-motion handoff does not match the current svg_output; '
+ 'deck-motion handoff does not match the current source inputs; '
'run the base export again'
)
motion = report.get('deck_motion')
diff --git a/skills/ppt-master/scripts/tests/test_edit_native_batch_c.py b/skills/ppt-master/scripts/tests/test_edit_native_batch_c.py
new file mode 100644
index 00000000..9e3bb3e5
--- /dev/null
+++ b/skills/ppt-master/scripts/tests/test_edit_native_batch_c.py
@@ -0,0 +1,270 @@
+#!/usr/bin/env python3
+"""Regress CLI arguments, round-trip quality receipts, shape text, and notes.
+
+Usage:
+ python3 -m unittest tests.test_edit_native_batch_c
+
+Dependencies:
+ Standard library and local PPT Master conversion modules.
+"""
+
+from __future__ import annotations
+
+import asyncio
+import contextlib
+import copy
+import io
+import json
+import sys
+import tempfile
+import unittest
+import zipfile
+from pathlib import Path
+from types import SimpleNamespace
+from unittest.mock import AsyncMock, Mock, patch
+from xml.etree import ElementTree as ET
+
+SCRIPTS_DIR = Path(__file__).resolve().parents[1]
+if str(SCRIPTS_DIR) not in sys.path:
+ sys.path.insert(0, str(SCRIPTS_DIR))
+
+import notes_to_audio # noqa: E402
+from authoring_roundtrip import SourceRefRecord # noqa: E402
+from pptx_delivery_check import audit_pptx_delivery # noqa: E402
+from pptx_to_svg.converter import ConvertOptions, convert_pptx_to_svg # noqa: E402
+from svg_quality import cli as quality_cli # noqa: E402
+from svg_quality.checker import SVGQualityChecker # noqa: E402
+from svg_to_pptx.drawingml.converter import convert_svg_to_slide_shapes # noqa: E402
+from svg_to_pptx.pptx_package import cli as export_cli # noqa: E402
+from svg_to_pptx.pptx_package.builder import create_pptx_with_native_svg # noqa: E402
+
+SVG_NS = "http://www.w3.org/2000/svg"
+SVG = (
+ f'<svg xmlns="{SVG_NS}" viewBox="0 0 1280 720" font-family="Arial">'
+ '<rect x="20" y="20" width="500" height="100" fill="#123456"/>'
+ '<text x="40" y="80" font-size="24" fill="#FFFFFF">Example</text></svg>'
+)
+
+
+def _invoke(main, *args: str) -> tuple[int, str]:
+ output = io.StringIO()
+ with patch.object(sys, "argv", ["batch-c", *map(str, args)]), \
+ contextlib.redirect_stdout(output), contextlib.redirect_stderr(output):
+ try:
+ code = main()
+ except SystemExit as exc:
+ code = exc.code
+ return code or 0, output.getvalue()
+
+
+def _workspace(root: Path, *, notes: bool = False) -> Path:
+ pages = []
+ for index in (1, 2):
+ svg = root / f"slide_{index:02d}.svg"
+ svg.write_text(SVG, encoding="utf-8")
+ pages.append(svg)
+ source = root / "source.pptx"
+ with contextlib.redirect_stdout(io.StringIO()):
+ success = create_pptx_with_native_svg(
+ pages, source, resource_root=root, pptx_structure="flat",
+ use_compat_mode=False, transition=None, workers=1, verbose=False,
+ notes={"slide_01": "Private speaker note."} if notes else None,
+ )
+ if not success:
+ raise AssertionError("Could not create the minimal source deck")
+ workspace = root / "workspace"
+ convert_pptx_to_svg(
+ source, workspace, ConvertOptions(inheritance_mode="both", roundtrip=True),
+ )
+ return workspace
+
+
+class EditNativeBatchCTests(unittest.TestCase):
+ def test_c1_negative_voice_options_accept_space_and_equals(self) -> None:
+ for option, value in (("--rate", "-5%"), ("--pitch", "-2Hz"), ("--volume", "-10%")):
+ for arguments in ([option, value], [f"{option}={value}"]):
+ with self.subTest(arguments=arguments):
+ code, output = _invoke(
+ notes_to_audio.main, *arguments, "--list-common-voices",
+ )
+ self.assertEqual(code, 0, output)
+ for arguments in (["--rate"], ["--rate", "--unknown"], ["--unknown"]):
+ with self.subTest(invalid=arguments):
+ code, _ = _invoke(notes_to_audio.main, *arguments, "--list-common-voices")
+ self.assertEqual(code, 2)
+ job = notes_to_audio.AudioJob(Path("note.md"), "Example", Path("page.mp3"))
+ with patch.object(notes_to_audio.backend_edge, "generate", new_callable=AsyncMock) as generate:
+ results = asyncio.run(notes_to_audio._generate_edge_jobs(
+ [job], Path("subtitles"), voice="en-US-AriaNeural",
+ rate="-5%", pitch="-2Hz", volume="-10%",
+ subtitle_max_chars=20, concurrency=1,
+ ))
+ self.assertEqual(results, [None])
+ self.assertEqual(generate.await_args.kwargs["rate"], "-5%")
+ self.assertEqual(generate.await_args.kwargs["pitch"], "-2Hz")
+ self.assertEqual(generate.await_args.kwargs["volume"], "-10%")
+ communicate = Mock(return_value=SimpleNamespace(save=AsyncMock()))
+ with patch.dict(sys.modules, {"edge_tts": SimpleNamespace(Communicate=communicate)}):
+ asyncio.run(notes_to_audio.backend_edge.generate(
+ "Example", Path("page.mp3"), voice="en-US-AriaNeural",
+ rate="-5%", pitch="-2Hz", volume="-10%",
+ ))
+ self.assertEqual(communicate.call_args.kwargs["pitch"], "-2Hz")
+ self.assertEqual(communicate.call_args.kwargs["volume"], "-10%")
+ with patch.object(notes_to_audio.backend_edge, "_generate_with_subtitles", new_callable=AsyncMock) as generate:
+ asyncio.run(notes_to_audio.backend_edge.generate(
+ "Example", Path("page.mp3"), voice="en-US-AriaNeural",
+ rate="-5%", pitch="-2Hz", volume="-10%", subtitle_path=Path("page.srt"),
+ ))
+ self.assertEqual(generate.await_args.kwargs["pitch"], "-2Hz")
+ self.assertEqual(generate.await_args.kwargs["volume"], "-10%")
+
+ def test_c2_roundtrip_quality_receipt_tracks_authoring_and_page_plan(self) -> None:
+ with tempfile.TemporaryDirectory() as tmp:
+ root = Path(tmp)
+ workspace = _workspace(root)
+ output = workspace / "exports/result.pptx"
+ code, log = _invoke(export_cli.main, workspace, "--roundtrip", "-o", output)
+ self.assertEqual(code, 0, log)
+ self.assertIn("quality_gate=not-provided", log)
+ code, log = _invoke(quality_cli.main, workspace, "--roundtrip", "--json")
+ self.assertEqual(code, 0, log)
+ report_path = workspace / "validation/svg_quality_report.json"
+ self.assertTrue(report_path.is_file(), log)
+ code, log = _invoke(export_cli.main, workspace, "--roundtrip", "-o", output)
+ self.assertEqual(code, 0, log)
+ self.assertIn("quality_gate=passed", log)
+ postflight = workspace / "validation/result.report.json"
+ self.assertEqual(
+ export_cli._load_deck_motion_handoff(workspace, str(postflight), []),
+ json.loads(postflight.read_text(encoding="utf-8"))["deck_motion"],
+ )
+ plan = workspace / "page_plan.json"
+ plan.write_text(json.dumps({
+ "schema": "ppt-master.roundtrip-page-plan.v1",
+ "pages": [{"source_slide": 2}, {"source_slide": 1}],
+ }), encoding="utf-8")
+ code, log = _invoke(export_cli.main, workspace, "--roundtrip", "-o", output)
+ self.assertEqual(code, 0, log)
+ self.assertIn("quality_gate=stale", log)
+ code, log = _invoke(quality_cli.main, workspace, "--roundtrip", "--json")
+ self.assertEqual(code, 0, log)
+ code, log = _invoke(export_cli.main, workspace, "--roundtrip", "-o", output)
+ self.assertEqual(code, 0, log)
+ self.assertIn("quality_gate=passed", log)
+ plan.write_text(json.dumps({
+ "schema": "ppt-master.roundtrip-page-plan.v1",
+ "pages": [{"source_slide": 1}, {"source_slide": 2}],
+ }), encoding="utf-8")
+ code, log = _invoke(export_cli.main, workspace, "--roundtrip", "-o", output)
+ self.assertEqual(code, 0, log)
+ self.assertIn("quality_gate=stale", log)
+ code, log = _invoke(quality_cli.main, workspace, "--roundtrip", "--json")
+ self.assertEqual(code, 0, log)
+ svg = workspace / "authoring-svg-flat/slide_01.svg"
+ svg.write_text(svg.read_text(encoding="utf-8") + "\n", encoding="utf-8")
+ code, log = _invoke(export_cli.main, workspace, "--roundtrip", "-o", output)
+ self.assertEqual(code, 0, log)
+ self.assertIn("quality_gate=stale", log)
+
+ def test_c3_semantic_text_contract_matches_export_and_skips_unchanged(self) -> None:
+ first = '<text x="40" y="60" font-size="20">First</text>'
+ second = '<text x="40" y="90" font-size="20">Second</text>'
+ cases = (
+ (first + second, "Semantic shape requires at most one paragraph-based text component"),
+ (f"<g>{first}</g>", "Semantic shape text must be one direct SVG text component"),
+ (first, None),
+ ('<text x="40" y="60" font-size="20">First<tspan x="40" dy="30">Second</tspan></text>', None),
+ ("", None),
+ )
+ with tempfile.TemporaryDirectory() as tmp:
+ svg = Path(tmp) / "slide.svg"
+ for body, error in cases:
+ with self.subTest(body=body):
+ content = (
+ f'<svg xmlns="{SVG_NS}" viewBox="0 0 1280 720" font-family="Arial">'
+ '<g id="shape" data-pptx-semantic-object="shape" data-pptx-frame="20 20 500 100">'
+ '<rect data-pptx-part="geometry" x="20" y="20" width="500" height="100"/>'
+ f'{body}</g></svg>'
+ )
+ svg.write_text(content, encoding="utf-8")
+ root = ET.fromstring(content)
+ if error:
+ with self.assertRaisesRegex(RuntimeError, error):
+ convert_svg_to_slide_shapes(str(svg), resource_root=svg.parent)
+ else:
+ convert_svg_to_slide_shapes(str(svg), resource_root=svg.parent)
+ checker = SVGQualityChecker()
+ with contextlib.redirect_stdout(io.StringIO()):
+ generated = checker.check_file(str(svg))
+ roundtrip = SVGQualityChecker()._check_roundtrip_file(
+ svg, SimpleNamespace(source_refs={}), root, {}, {},
+ output_index=1, source_slide=1,
+ )
+ for result in (generated, roundtrip):
+ errors = "\n".join(result["errors"])
+ if error:
+ self.assertIn(error, errors)
+ else:
+ self.assertNotIn("Semantic shape", errors)
+ group = root.find(f"{{{SVG_NS}}}g")
+ group.set("data-pptx-source-ref", "slide:2")
+ svg.write_bytes(ET.tostring(root))
+ source = SimpleNamespace(source_refs={
+ "slide:2": SourceRefRecord((), "", "semantic-shape"),
+ })
+ unchanged = SVGQualityChecker()._check_roundtrip_file(
+ svg, source, copy.deepcopy(root), {}, {},
+ output_index=1, source_slide=1,
+ )
+ self.assertFalse(unchanged["errors"], unchanged)
+ ordinary = (
+ f'<svg xmlns="{SVG_NS}" viewBox="0 0 1280 720" font-family="Arial">'
+ f'<g id="ordinary">{first}{second}</g></svg>'
+ )
+ svg.write_text(ordinary, encoding="utf-8")
+ convert_svg_to_slide_shapes(str(svg), resource_root=svg.parent)
+ with contextlib.redirect_stdout(io.StringIO()):
+ result = SVGQualityChecker().check_file(str(svg))
+ self.assertNotIn("Semantic shape", "\n".join(result["errors"]))
+
+ def test_c4_no_notes_removes_parts_overrides_and_relationships(self) -> None:
+ with tempfile.TemporaryDirectory() as tmp:
+ workspace = _workspace(Path(tmp), notes=True)
+ self.assertFalse((workspace / "page_plan.json").exists())
+ output = workspace / "exports/no_notes.pptx"
+ code, log = _invoke(export_cli.main, workspace, "--roundtrip", "--no-notes", "-o", output)
+ self.assertEqual(code, 0, log)
+ with zipfile.ZipFile(output) as archive:
+ self.assertFalse([name for name in archive.namelist() if name.startswith((
+ "ppt/notesSlides/", "ppt/notesMasters/",
+ ))])
+ for name in archive.namelist():
+ if name.endswith(".rels") or name == "[Content_Types].xml":
+ self.assertNotIn(b"notesSlide", archive.read(name), name)
+ self.assertNotIn(b"notesMaster", archive.read(name), name)
+ report = audit_pptx_delivery(output)
+ self.assertFalse(report["errors"], report)
+ orphan = workspace / "exports/orphan.pptx"
+ with zipfile.ZipFile(output) as source, zipfile.ZipFile(orphan, "w") as target:
+ for name in source.namelist():
+ payload = source.read(name)
+ if name == "[Content_Types].xml":
+ tree = ET.fromstring(payload)
+ ET.SubElement(tree, f"{tree.tag.removesuffix('Types')}Override", {
+ "PartName": "/ppt/notesSlides/notesSlide1_tf1.xml",
+ "ContentType": "application/vnd.openxmlformats-officedocument.presentationml.notesSlide+xml",
+ })
+ payload = ET.tostring(tree)
+ target.writestr(name, payload)
+ target.writestr(
+ "ppt/notesSlides/notesSlide1_tf1.xml",
+ '<p:notes xmlns:p="http://schemas.openxmlformats.org/presentationml/2006/main"/>',
+ )
+ report = audit_pptx_delivery(orphan)
+ self.assertFalse(report["errors"], report)
+ advisory = next((
+ item for item in report["advisories"] if item["code"] == "orphan_notes_parts"
+ ), None)
+ self.assertIsNotNone(advisory, report)
+ self.assertEqual(advisory["parts"], ["ppt/notesSlides/notesSlide1_tf1.xml"])
diff --git a/skills/ppt-master/scripts/tts_backends/backend_edge.py b/skills/ppt-master/scripts/tts_backends/backend_edge.py
index 77506f04..2f5dff5d 100644
--- a/skills/ppt-master/scripts/tts_backends/backend_edge.py
+++ b/skills/ppt-master/scripts/tts_backends/backend_edge.py
@@ -80,6 +80,8 @@ async def generate(
*,
voice: str,
rate: str,
+ pitch: str = "+0Hz",
+ volume: str = "+0%",
subtitle_path: Path | None = None,
subtitle_max_chars: int = DEFAULT_SUBTITLE_MAX_CHARS,
) -> None:
@@ -91,6 +93,8 @@ async def generate(
subtitle_path,
voice=voice,
rate=rate,
+ pitch=pitch,
+ volume=volume,
max_chars=subtitle_max_chars,
)
return
@@ -103,7 +107,9 @@ async def generate(
"python3 -m pip install edge-tts"
) from exc
- communicate = edge_tts.Communicate(text, voice=voice, rate=normalize_rate(rate))
+ communicate = edge_tts.Communicate(
+ text, voice=voice, rate=normalize_rate(rate), pitch=pitch, volume=volume,
+ )
await communicate.save(str(output_path))
@@ -439,6 +445,8 @@ async def _generate_with_subtitles(
*,
voice: str,
rate: str,
+ pitch: str = "+0Hz",
+ volume: str = "+0%",
max_chars: int,
) -> None:
"""Generate one MP3 and compact SRT from the same Edge word-timing stream."""
@@ -454,6 +462,8 @@ async def _generate_with_subtitles(
text,
voice=voice,
rate=normalize_rate(rate),
+ pitch=pitch,
+ volume=volume,
boundary="WordBoundary",
)
audio_descriptor = -1