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(text-layout): restore heuristic paragraph merging
hugohe3e8c70107ada6
diff --git a/skills/ppt-master/scripts/finalize_svg.py b/skills/ppt-master/scripts/finalize_svg.pyindex 947afc57..05781a47 100644--- a/skills/ppt-master/scripts/finalize_svg.py+++ b/skills/ppt-master/scripts/finalize_svg.py@@ -57,7 +57,6 @@ from svg_to_pptx.geometry_properties import (GeometryStyleError,materialize_inline_geometry_in_file,)-from svg_to_pptx.text_contract import RUNTIME_ATTRS, TextContractErrorfrom svg_to_pptx.use_expander import (UseExpansionError,expand_local_use_references_in_file,@@ -90,22 +89,13 @@ def process_flatten_text(svg_file: Path, verbose: bool = False) -> bool:from xml.etree import ElementTree as ETtree = ET.parse(str(svg_file))- changed = flatten_text_with_tspans(- tree,- enforce_native_carrier=False,- source_name=svg_file.name,- )+ changed = flatten_text_with_tspans(tree)if changed:- for element in tree.getroot().iter():- for attribute in RUNTIME_ATTRS:- element.attrib.pop(attribute, None)tree.write(str(svg_file), encoding='unicode', xml_declaration=False)if verbose:safe_print(f" [OK] {svg_file.name}: text flattened")return changed- except TextContractError:- raiseexcept Exception as e:if verbose:safe_print(f" [ERROR] {svg_file.name}: {e}")@@ -371,19 +361,10 @@ Aliases (still accepted):safe_print("[ERROR] --image-scale must be >= 1")sys.exit(1)- try:- success = finalize_project(- args.project_dir,- options,- args.dry_run,- args.quiet,- compress=args.compress,- max_dimension=args.max_dimension,- image_scale=args.image_scale,- )- except TextContractError as exc:- safe_print(f"[ERROR] {exc.format()}")- sys.exit(1)+ success = finalize_project(args.project_dir, options, args.dry_run, args.quiet,+ compress=args.compress,+ max_dimension=args.max_dimension,+ image_scale=args.image_scale)sys.exit(0 if success else 1)diff --git a/skills/ppt-master/scripts/pptx_shapes/semantic_hash.py b/skills/ppt-master/scripts/pptx_shapes/semantic_hash.pyindex 52471f82..88cea092 100644--- a/skills/ppt-master/scripts/pptx_shapes/semantic_hash.py+++ b/skills/ppt-master/scripts/pptx_shapes/semantic_hash.py@@ -39,13 +39,6 @@ _ROOT_TEXT_STYLE_ATTRS = frozenset({"word-spacing",})-_TEXT_CONTRACT_ATTRS = frozenset({- "data-pptx-text-mode",- "data-pptx-text-bounds",- "data-pptx-break",- "data-pptx-join",-})-NATIVE_FALLBACK_SHA256_ATTR = "data-pptx-fallback-sha256"_NATIVE_FALLBACK_IGNORED_TAGS = frozenset({"metadata", "title", "desc"})_NATIVE_FALLBACK_IGNORED_ATTRS = frozenset({@@ -246,7 +239,6 @@ def _preview_subtree(if name != "id"and name != "data-pptx-preview-sha256"and not name.startswith("data-pptx-runtime-")- and name not in _TEXT_CONTRACT_ATTRS),"children": children,}@@ -446,11 +438,7 @@ def _element_payload(element: ET.Element) -> dict:"attrs": sorted((name, value)for name, value in element.attrib.items()- if name != "id"- and (- not name.startswith("data-pptx-")- or name in _TEXT_CONTRACT_ATTRS- )+ if not name.startswith("data-pptx-") and name != "id"),"text": element.text or "","children": [diff --git a/skills/ppt-master/scripts/pptx_to_svg/slide_to_svg.py b/skills/ppt-master/scripts/pptx_to_svg/slide_to_svg.pyindex 01468011..2cbb5270 100644--- a/skills/ppt-master/scripts/pptx_to_svg/slide_to_svg.py+++ b/skills/ppt-master/scripts/pptx_to_svg/slide_to_svg.py@@ -40,7 +40,7 @@ from .color_resolver import ColorPalette, find_color_elem, resolve_colorfrom .chart_to_svg import CHART_URI, CHARTEX_URI, extract_native_chart_payloadfrom .custgeom_to_svg import convert_custom_geomfrom .effect_to_svg import convert_effects-from .emu_units import NS, Xfrm, fmt_num, ooxml_bool+from .emu_units import NS, Xfrm, fmt_numfrom .fill_to_svg import resolve_fillfrom .ln_to_svg import resolve_strokefrom .ooxml_loader import OoxmlPackage, PartRef, SlideRef@@ -61,9 +61,6 @@ from .txbody_to_svg import ()-_TEXT_CARRIER_VERSION = "1"--# ---------------------------------------------------------------------------# AssemblyContext# ---------------------------------------------------------------------------@@ -288,33 +285,17 @@ def assemble_part_solo(# Per-node dispatch# ----------------------------------------------------------------------------def _convert_node(- node: ShapeNode,- ctx: AssemblyContext,- *,- top_level: bool,- transformed_ancestor: bool = False,-) -> str:+def _convert_node(node: ShapeNode, ctx: AssemblyContext, *, top_level: bool) -> str:if node.hidden and not ctx.keep_hidden:return ""if node.kind == SHAPE:- return _convert_shape(- node,- ctx,- top_level=top_level,- transformed_ancestor=transformed_ancestor,- )+ return _convert_shape(node, ctx, top_level=top_level)if node.kind == PICTURE:return _convert_picture(node, ctx, top_level=top_level)if node.kind == CONNECTOR:return _convert_connector(node, ctx, top_level=top_level)if node.kind == GROUP:- return _convert_group(- node,- ctx,- top_level=top_level,- transformed_ancestor=transformed_ancestor,- )+ return _convert_group(node, ctx, top_level=top_level)if node.kind == GRAPHIC:return _convert_graphic_fallback(node, ctx, top_level=top_level)return ""@@ -324,13 +305,7 @@ def _convert_node(# Shape (<p:sp>)# ----------------------------------------------------------------------------def _convert_shape(- node: ShapeNode,- ctx: AssemblyContext,- *,- top_level: bool,- transformed_ancestor: bool,-) -> str:+def _convert_shape(node: ShapeNode, ctx: AssemblyContext, *, top_level: bool) -> str:sp_pr = node.xml.find("p:spPr", NS)# Check for blipFill (image-filled shape, e.g. Canva exports where images@@ -414,16 +389,6 @@ def _convert_shape(text_result.svg,))- inner_parts.append(- _text_carrier_metadata(- node,- sp_pr,- tx_body,- is_vertical=is_vertical,- has_blip_fill=blip_fill_elem is not None,- transformed_ancestor=transformed_ancestor,- )- )if text_result.svg:inner_parts.append(text_result.svg)inner = "\n".join(inner_parts) if inner_parts else ""@@ -457,148 +422,6 @@ def _txbody_metadata()-def _style_ref_is_visible(node: ShapeNode, ref_name: str) -> bool:- """Return whether a shape-style reference selects a visible style entry."""- style = node.xml.find("p:style", NS)- ref = style.find(f"a:{ref_name}", NS) if style is not None else None- if ref is None:- return False- raw_idx = ref.get("idx")- if raw_idx is None:- return True- try:- return int(raw_idx) > 0- except ValueError:- return True---def _has_visible_shape_fill(node: ShapeNode, sp_pr: ET.Element | None) -> bool:- """Return whether the source shape owns a visible non-image fill."""- if sp_pr is not None:- if sp_pr.find("a:noFill", NS) is not None:- return False- if any(- sp_pr.find(f"a:{tag}", NS) is not None- for tag in ("solidFill", "gradFill", "pattFill", "grpFill")- ):- return True- return _style_ref_is_visible(node, "fillRef")---def _has_visible_shape_line(node: ShapeNode, sp_pr: ET.Element | None) -> bool:- """Return whether the source shape owns a visible outline."""- line = sp_pr.find("a:ln", NS) if sp_pr is not None else None- if line is not None:- return line.find("a:noFill", NS) is None- return _style_ref_is_visible(node, "lnRef")---def _has_visible_shape_effect(node: ShapeNode, sp_pr: ET.Element | None) -> bool:- """Return whether the source shape owns a visible effect or 3D treatment."""- if sp_pr is not None:- for tag in ("effectLst", "effectDag"):- effect = sp_pr.find(f"a:{tag}", NS)- if effect is not None and len(effect):- return True- if any(- sp_pr.find(f"a:{tag}", NS) is not None- for tag in ("scene3d", "sp3d")- ):- return True- return _style_ref_is_visible(node, "effectRef")---def _text_carrier_ineligibility_reasons(- node: ShapeNode,- sp_pr: ET.Element | None,- tx_body: ET.Element,- nv_sp_pr: ET.Element | None,- *,- is_vertical: bool,- has_blip_fill: bool,- transformed_ancestor: bool,-) -> list[str]:- """Return stable reasons why a native text carrier is unsafe to rewrite."""- reasons: list[str] = []- c_nv_sp_pr = (- nv_sp_pr.find("p:cNvSpPr", NS)- if nv_sp_pr is not None- else None- )- owns_text_frame = (- node.placeholder is not None- or (- c_nv_sp_pr is not None- and ooxml_bool(c_nv_sp_pr.get("txBox"))- )- )-- if is_vertical:- reasons.append("vertical-text")- if (- transformed_ancestor- or abs(node.xfrm.rot) > 1e-9- or node.xfrm.flip_h- or node.xfrm.flip_v- ):- reasons.append("rotated-or-flipped")- if not owns_text_frame:- reasons.append("missing-frame-owner")- if has_blip_fill:- reasons.append("blip-fill")- if _has_visible_shape_fill(node, sp_pr):- reasons.append("visible-fill")- if _has_visible_shape_line(node, sp_pr):- reasons.append("visible-line")- if _has_visible_shape_effect(node, sp_pr):- reasons.append("visible-effect")- if has_relationship_attributes(tx_body):- reasons.append("txbody-relationships")- if nv_sp_pr is None:- reasons.append("missing-nvsppr")- elif has_relationship_attributes(nv_sp_pr):- reasons.append("nvsppr-relationships")- return reasons---def _text_carrier_metadata(- node: ShapeNode,- sp_pr: ET.Element | None,- tx_body: ET.Element,- *,- is_vertical: bool,- has_blip_fill: bool,- transformed_ancestor: bool,-) -> str:- """Describe whether a native text frame can accept a rebuilt txBody."""- nv_sp_pr = node.xml.find("p:nvSpPr", NS)- reasons = _text_carrier_ineligibility_reasons(- node,- sp_pr,- tx_body,- nv_sp_pr,- is_vertical=is_vertical,- has_blip_fill=has_blip_fill,- transformed_ancestor=transformed_ancestor,- )- eligible = not reasons- reason = ",".join(reasons) if reasons else "eligible"- attrs = [- 'data-pptx-part="text-carrier"',- f'data-pptx-text-carrier-version="{_TEXT_CARRIER_VERSION}"',- f'data-pptx-text-carrier-eligible="{1 if eligible else 0}"',- f'data-pptx-text-carrier-reason="{reason}"',- ]-- if nv_sp_pr is None or has_relationship_attributes(nv_sp_pr):- return f'<metadata {" ".join(attrs)}/>'-- raw = ET.tostring(nv_sp_pr, encoding="utf-8")- encoded = base64.b64encode(raw).decode("ascii")- attrs.append('data-pptx-encoding="base64"')- return f'<metadata {" ".join(attrs)}>{encoded}</metadata>'--def _resolve_geometry(node: ShapeNode, sp_pr: ET.Element | None) -> GeomResult | None:"""Resolve a DrawingML shape geometry into an absolute SVG geometry model."""prst_geom = sp_pr.find("a:prstGeom", NS) if sp_pr is not None else None@@ -857,28 +680,11 @@ def _convert_connector(node: ShapeNode, ctx: AssemblyContext, *, top_level: bool# Group (<p:grpSp>)# ----------------------------------------------------------------------------def _convert_group(- node: ShapeNode,- ctx: AssemblyContext,- *,- top_level: bool,- transformed_ancestor: bool,-) -> str:+def _convert_group(node: ShapeNode, ctx: AssemblyContext, *, top_level: bool) -> str:"""Render group contents flat (children already remapped to slide space)."""inner_parts: list[str] = []- child_has_transformed_ancestor = (- transformed_ancestor- or abs(node.xfrm.rot) > 1e-9- or node.xfrm.flip_h- or node.xfrm.flip_v- )for child in node.children:- chunk = _convert_node(- child,- ctx,- top_level=False,- transformed_ancestor=child_has_transformed_ancestor,- )+ chunk = _convert_node(child, ctx, top_level=False)if chunk:inner_parts.append(chunk)if not inner_parts:diff --git a/skills/ppt-master/scripts/svg_authoring_view.py b/skills/ppt-master/scripts/svg_authoring_view.pyindex 0ba8a601..c3030963 100644--- a/skills/ppt-master/scripts/svg_authoring_view.py+++ b/skills/ppt-master/scripts/svg_authoring_view.py@@ -72,7 +72,6 @@ def _local_name(name: object) -> str:@dataclassclass ProjectionStats:txbody_metadata: int = 0- text_carrier_metadata: int = 0hidden_geometry_carriers: int = 0geometry_preview_wrappers: int = 0geometry_detail_markers: int = 0@@ -82,7 +81,6 @@ class ProjectionStats:def as_dict(self) -> dict[str, object]:return {"txbody_metadata": self.txbody_metadata,- "text_carrier_metadata": self.text_carrier_metadata,"hidden_geometry_carriers": self.hidden_geometry_carriers,"geometry_preview_wrappers": self.geometry_preview_wrappers,"geometry_detail_markers": self.geometry_detail_markers,@@ -92,7 +90,6 @@ class ProjectionStats:def merge(self, other: "ProjectionStats") -> None:self.txbody_metadata += other.txbody_metadata- self.text_carrier_metadata += other.text_carrier_metadataself.hidden_geometry_carriers += other.hidden_geometry_carriersself.geometry_preview_wrappers += other.geometry_preview_wrappersself.geometry_detail_markers += other.geometry_detail_markers@@ -190,11 +187,8 @@ def _project_subtree(parent: ET.Element, stats: ProjectionStats) -> None:part = child.get("data-pptx-part")tag = _local_name(child.tag)- if tag == "metadata" and part in {"txbody", "text-carrier"}:- if part == "txbody":- stats.txbody_metadata += 1- else:- stats.text_carrier_metadata += 1+ if tag == "metadata" and part == "txbody":+ stats.txbody_metadata += 1_remove_child(parent, child)continuediff --git a/skills/ppt-master/scripts/svg_editor/annotations.py b/skills/ppt-master/scripts/svg_editor/annotations.pyindex e43ea892..7a02353c 100644--- a/skills/ppt-master/scripts/svg_editor/annotations.py+++ b/skills/ppt-master/scripts/svg_editor/annotations.py@@ -158,8 +158,6 @@ def promote_tspan_to_text(return False, 'not-found'if _local_name(target) != 'tspan' or _local_name(text_el) != 'text':return False, 'not-tspan'- if text_el.get('data-pptx-text-mode') == 'paragraph':- return False, 'explicit-paragraph-line-move'grandparent: Optional[ET.Element] = Nonefor candidate in root.iter():diff --git a/skills/ppt-master/scripts/svg_editor/server.py b/skills/ppt-master/scripts/svg_editor/server.pyindex 9ddfc4d8..d1d74a74 100644--- a/skills/ppt-master/scripts/svg_editor/server.py+++ b/skills/ppt-master/scripts/svg_editor/server.py@@ -242,10 +242,6 @@ _UNSAFE_COLOR_RE = re.compile(r'[;:@\\]|url\s*\(', re.IGNORECASE)_SAFE_ATTR_NAME_RE = re.compile(r'^[A-Za-z_][A-Za-z0-9_.:-]*$')_MAX_ATTR_VALUE_LEN = 256_MAX_EDIT_TEXT_LEN = 5000-_PURE_TRANSLATE_RE = re.compile(- r'^\s*(?:translate\(\s*[-+]?\d+(?:\.\d+)?'- r'(?:(?:\s*,\s*|\s+)[-+]?\d+(?:\.\d+)?)?\s*\)\s*)+$'-)_ADDABLE_BATCH_ATTRS = frozenset({'fill', 'stroke', 'opacity','font-size', 'font-family', 'font-weight', 'text-anchor',@@ -318,42 +314,6 @@ def _find_by_id(root: ET.Element, element_id: str) -> Optional[ET.Element]:return None-def _explicit_paragraph_owner(- root: ET.Element,- target: ET.Element,-) -> Optional[ET.Element]:- """Return the explicit paragraph <text> owning target, if any."""- for candidate in root.iter():- tag = candidate.tag.split('}', 1)[-1]- if tag != 'text' or candidate.get('data-pptx-text-mode') != 'paragraph':- continue- if candidate is target or target in candidate.iter():- return candidate- return None---def _validate_explicit_paragraph_edit(- root: ET.Element,- target: ET.Element,- attrs: dict,-) -> Optional[str]:- """Protect paragraph frame/visual-line ownership in Live Preview."""- owner = _explicit_paragraph_owner(root, target)- if owner is None:- return None- if target is not owner and {'x', 'y', 'dy', 'transform'} & attrs.keys():- return 'explicit paragraph lines cannot be moved independently; switch to lines mode'- if target is owner and {'x', 'y'} & attrs.keys():- return 'move an explicit paragraph with a pure translate, not parent x/y'- if target is owner and 'data-pptx-text-bounds' in attrs:- return 'explicit paragraph bounds require a dedicated frame edit operation'- if target is owner and 'transform' in attrs:- transform = attrs.get('transform')- if transform is not None and not _PURE_TRANSLATE_RE.fullmatch(transform):- return 'explicit paragraphs only support pure translate transforms'- return None--def _apply_edit_record(root: ET.Element, record: dict) -> tuple[bool, Optional[str]]:element_id = record.get('element_id')if not isinstance(element_id, str):@@ -379,9 +339,6 @@ def _apply_edit_record(root: ET.Element, record: dict) -> tuple[bool, Optional[starget = _find_by_id(root, element_id)if target is None:return False, 'not-found'- paragraph_error = _validate_explicit_paragraph_edit(root, target, attrs)- if paragraph_error:- return False, paragraph_error_strip_edited_inline_geometry(target, attrs.keys())ok, reason = set_attributes(root, element_id, attrs)if not ok:@@ -831,9 +788,6 @@ def create_app(attr_err = _validate_edit_attrs(attrs, set(target.attrib.keys()))if attr_err:return jsonify({'error': attr_err}), 400- paragraph_error = _validate_explicit_paragraph_edit(root, target, attrs)- if paragraph_error:- return jsonify({'error': paragraph_error}), 400changes = []staged: dict = {'element_id': element_id}diff --git a/skills/ppt-master/scripts/svg_editor/static/app.js b/skills/ppt-master/scripts/svg_editor/static/app.jsindex 0bf04c88..6327536e 100644--- a/skills/ppt-master/scripts/svg_editor/static/app.js+++ b/skills/ppt-master/scripts/svg_editor/static/app.js@@ -1886,10 +1886,6 @@function renderEditableProps(el) {var tag = localName(el);var isGroup = tag === "g";- var isExplicitParagraph = tag === "text" &&- el.getAttribute("data-pptx-text-mode") === "paragraph";- var paragraphOwner = explicitParagraphOwner(el);- var isExplicitParagraphLine = paragraphOwner && paragraphOwner !== el;var html = '<div class="prop-caption">' +escapeHtml(isGroup ? t("label_group_edit") : t("label_direct_edit")) + '</div>';@@ -1901,26 +1897,22 @@'</button></td></tr></table>';}- if (!isExplicitParagraphLine) {- html += '<div class="prop-section">' + escapeHtml(t("section_geometry")) + '</div>';- html += '<table class="prop-table">';- // Geometry: position + size in render coords — editable, folded into- // the element's transform matrix on commit (works for any element type).- try {- var rb = computeRenderBox(el);- html += '<tr><td class="prop-key">position</td><td class="prop-val geo-cell">' +- '<input type="number" class="prop-edit prop-edit-geo" data-geo="x" value="' + Math.round(rb.x) + '">' +- '<input type="number" class="prop-edit prop-edit-geo" data-geo="y" value="' + Math.round(rb.y) + '">' +- '</td></tr>';- if (!isExplicitParagraph) {- html += '<tr><td class="prop-key">size</td><td class="prop-val geo-cell">' +- '<input type="number" class="prop-edit prop-edit-geo" data-geo="w" min="1" value="' + Math.round(rb.w) + '">' +- '<input type="number" class="prop-edit prop-edit-geo" data-geo="h" min="1" value="' + Math.round(rb.h) + '">' +- '</td></tr>';- }- } catch (e) { /* no geometry */ }- html += '</table>';- }+ html += '<div class="prop-section">' + escapeHtml(t("section_geometry")) + '</div>';+ html += '<table class="prop-table">';+ // Geometry: position + size in render coords — editable, folded into+ // the element's transform matrix on commit (works for any element type).+ try {+ var rb = computeRenderBox(el);+ html += '<tr><td class="prop-key">position</td><td class="prop-val geo-cell">' ++ '<input type="number" class="prop-edit prop-edit-geo" data-geo="x" value="' + Math.round(rb.x) + '">' ++ '<input type="number" class="prop-edit prop-edit-geo" data-geo="y" value="' + Math.round(rb.y) + '">' ++ '</td></tr>';+ html += '<tr><td class="prop-key">size</td><td class="prop-val geo-cell">' ++ '<input type="number" class="prop-edit prop-edit-geo" data-geo="w" min="1" value="' + Math.round(rb.w) + '">' ++ '<input type="number" class="prop-edit prop-edit-geo" data-geo="h" min="1" value="' + Math.round(rb.h) + '">' ++ '</td></tr>';+ } catch (e) { /* no geometry */ }+ html += '</table>';var textStyleTargets = textStyleTargetsForSelection(el);@@ -1974,22 +1966,13 @@function ownAttributeSpecs(el) {var specs = [];- var paragraphOwner = explicitParagraphOwner(el);Array.from(el.attributes).forEach(function (attr) {var key = attr.name;if (key === "class" || key === "data-edit-target" || key === "data-edit-annotation") return;- var paragraphLocked = paragraphOwner && (- (paragraphOwner === el && [- "x", "y", "transform", "data-pptx-text-bounds"- ].indexOf(key) !== -1) ||- (paragraphOwner !== el && [- "x", "y", "dy", "transform"- ].indexOf(key) !== -1)- );specs.push({key: key,value: attr.value,- editable: isEditableAttributeName(key) && !paragraphLocked+ editable: isEditableAttributeName(key)});});return specs;@@ -2389,23 +2372,6 @@}function rawMoveAttrs(el, dx, dy) {- var paragraphOwner = explicitParagraphOwner(el);- if (paragraphOwner) {- if (paragraphOwner !== el) return null;- var paragraphOwn = ownMatrix(el);- var paragraphClose = function (a, b) { return Math.abs(a - b) < 1e-6; };- if (!paragraphClose(paragraphOwn.a, 1) || !paragraphClose(paragraphOwn.b, 0) ||- !paragraphClose(paragraphOwn.c, 0) || !paragraphClose(paragraphOwn.d, 1)) {- return null;- }- var paragraphDelta = parentDelta(dx, dy, el);- if (!paragraphDelta) return null;- var tx = paragraphOwn.e + paragraphDelta.x;- var ty = paragraphOwn.f + paragraphDelta.y;- return {- transform: "translate(" + formatCoord(tx) + " " + formatCoord(ty) + ")"- };- }if (localName(el) === "g" && el.hasAttribute("data-icon") &&el.hasAttribute("data-use-x") && el.hasAttribute("data-use-y")) {var iconOwn = ownMatrix(el);@@ -2496,8 +2462,6 @@}function attrsForMove(el, dx, dy) {- var paragraphOwner = explicitParagraphOwner(el);- if (paragraphOwner && paragraphOwner !== el) return null;var attrs = rawMoveAttrs(el, dx, dy);if (attrs) return attrs;var rb = computeRenderBox(el);@@ -2506,18 +2470,6 @@return tstr ? { transform: tstr } : null;}- function explicitParagraphOwner(el) {- var candidate = el;- while (candidate && candidate !== svgContent) {- if (localName(candidate) === "text" &&- candidate.getAttribute("data-pptx-text-mode") === "paragraph") {- return candidate;- }- candidate = candidate.parentElement;- }- return null;- }-function computeElementsBox(elements) {var minX = Infinity, minY = Infinity, maxX = -Infinity, maxY = -Infinity;elements.forEach(function (el) {diff --git a/skills/ppt-master/scripts/svg_finalize/flatten_tspan.py b/skills/ppt-master/scripts/svg_finalize/flatten_tspan.pyindex 3a88a4ee..1a09efc1 100644--- a/skills/ppt-master/scripts/svg_finalize/flatten_tspan.py+++ b/skills/ppt-master/scripts/svg_finalize/flatten_tspan.py@@ -2,7 +2,6 @@ import osimport sysimport reimport argparse-import weakreffrom pathlib import Pathfrom xml.etree import ElementTree as ET@@ -11,21 +10,6 @@ if str(_SCRIPTS_DIR) not in sys.path:sys.path.insert(0, str(_SCRIPTS_DIR))from console_encoding import configure_utf8_stdio # noqa: E402-from svg_to_pptx.text_contract import ( # noqa: E402- RUNTIME_CONTRACT_ATTR,- RUNTIME_EFFECTIVE_MODE_ATTR,- RUNTIME_LINE_COUNT_ATTR,- RUNTIME_LINE_INDEX_ATTR,- RUNTIME_MODE_REASON_ATTR,- RUNTIME_REQUESTED_MODE_ATTR,- RUNTIME_SOURCE_ID_ATTR,- TEXT_BOUNDS_ATTR,- TEXT_BREAK_ATTR,- TEXT_JOIN_ATTR,- TEXT_MODE_ATTR,- TextBlockContract,- validate_text_contracts,-)configure_utf8_stdio()@@ -36,8 +20,6 @@ NSMAP = {"svg": SVG_NS}# Ensure pretty element names without ns0 prefix on writeET.register_namespace("", SVG_NS)-_NORMALIZED_ROOTS: weakref.WeakSet[ET.Element] = weakref.WeakSet()-TEXT_STYLE_ATTRS = {# common text styling@@ -202,10 +184,10 @@ def copy_text_attrs(PARAGRAPH_MARK_ATTR = "data-paragraph-line-height"PARAGRAPH_SPACE_BEFORE_ATTR = "data-paragraph-space-before"-PARAGRAPH_TOP_INSET_ATTR = "data-paragraph-top-inset"-# Internal line-break marker consumed by the DrawingML converter. Explicit-# soft joins remain represented solely by author-owned data-pptx-break="soft".-PARAGRAPH_LINE_BREAK_ATTR = "data-paragraph-line-break"+# Marks a line-break tspan as a SOFT break inside the current paragraph+# (SVG used dy to simulate text wrapping; the downstream converter should+# merge its runs into the previous <a:p> rather than start a new one).+PARAGRAPH_SOFT_BREAK_ATTR = "data-paragraph-soft-break"# Tolerance for detecting "base line-height" vs "paragraph gap": dy values# within ±DY_TOLERANCE_PX of each other are considered the same line-height.@@ -282,25 +264,19 @@ def _classify_paragraph_block(text_el: ET.Element,is_svg_tag,is_new_line_tspan,-) -> tuple[- float,- list[float],- list[bool],- list[list[ET.Element]],- ET.Element | None,-] | None:+) -> tuple[float, list[float], list[bool], list[list[ET.Element]], ET.Element | None] | None:"""Detect a mergeable paragraph block.Returns ``(base_line_height_px, extra_space_before_px_per_line,- is_line_break_per_line, line_groups, synthetic_first_line)`` if the children+ is_soft_break_per_line, line_groups, synthetic_first_line)`` if the childrenform a mergeable paragraph. Each list has one entry per direct-child tspan(line), including a synthetic first line when the source used leading text:- extra_space_before_px_per_line[i]: extra px above base line-height,used as <a:spcBef> on the downstream <a:p>. First entry is 0.- - is_line_break_per_line[i]: True if the line stays in the previous- <a:p> behind an explicit line break; False if it starts a fresh <a:p>.- First entry is always False (paragraph head).+ - is_soft_break_per_line[i]: True if this line should merge into the+ previous <a:p> (SVG dy was simulating word-wrap); False if it starts+ a fresh <a:p>. First entry is always False (paragraph head).Conditions (all must hold):- No direct text under <text>, except simple leading text that can be@@ -381,7 +357,7 @@ def _classify_paragraph_block(return Noneextras: list[float] = [0.0] # first line never has space-before- line_breaks: list[bool] = [False] # first line starts a paragraph+ soft_breaks: list[bool] = [False] # first line starts a paragraphfor idx, d in enumerate(dy_values[1:], start=1):if d + DY_TOLERANCE_PX < base:return None # below base — line overlap, not a paragraph@@ -390,24 +366,25 @@ def _classify_paragraph_block(extra = d - baseif extra < 0:extra = 0.0- # Equal-dy visual lines are authored line breaks, not permission for- # Office to reflow them. Larger gaps and list markers start a fresh- # paragraph so spacing and bullet semantics remain independent.- is_line = (+ # dy at the base line-height = soft break (SVG was simulating wrap);+ # dy strictly greater than base = hard paragraph break. List markers+ # also start a fresh paragraph so bullet/ordered items do not merge+ # into the previous item when exported to PowerPoint.+ is_soft = (abs(extra) <= DY_TOLERANCE_PXand not _starts_with_list_marker(line_groups[idx]))- extras.append(0.0 if is_line else extra)- line_breaks.append(is_line)+ extras.append(0.0 if is_soft else extra)+ soft_breaks.append(is_soft)- return base, extras, line_breaks, line_groups, synthetic_first+ return base, extras, soft_breaks, line_groups, synthetic_firstdef _emit_mergeable_paragraph(text_el: ET.Element,- base_dy: float | None,+ base_dy: float,extras: list[float],- line_breaks: list[bool],+ soft_breaks: list[bool],line_groups: list[list[ET.Element]],synthetic_first: ET.Element | None = None,) -> None:@@ -416,14 +393,12 @@ def _emit_mergeable_paragraph(The base line-height goes on the parent <text> via PARAGRAPH_MARK_ATTR.Each direct-child tspan is normalized: x/y/dy stripped; inline-runstyling and nested tspans are preserved. Per-tspan attrs:- - PARAGRAPH_LINE_BREAK_ATTR="1" on authored visual line breaks+ - PARAGRAPH_SOFT_BREAK_ATTR="1" on tspans that should be appended to+ the previous <a:p> downstream (SVG used dy to simulate wrap)- PARAGRAPH_SPACE_BEFORE_ATTR on tspans that open a new paragraphwith an extra gap (omitted when 0)"""- if base_dy is not None:- text_el.set(PARAGRAPH_MARK_ATTR, format_number(base_dy))- else:- text_el.attrib.pop(PARAGRAPH_MARK_ATTR, None)+ text_el.set(PARAGRAPH_MARK_ATTR, format_number(base_dy))if synthetic_first is not None:text_el.text = Nonetext_el.insert(0, synthetic_first)@@ -450,144 +425,37 @@ def _emit_mergeable_paragraph(text_el.remove(child)extras_iter = iter(extras)- line_break_iter = iter(line_breaks)+ soft_iter = iter(soft_breaks)for tspan in normalized_lines:for k in ("x", "y", "dy"):if k in tspan.attrib:del tspan.attrib[k]- tspan.attrib.pop(PARAGRAPH_LINE_BREAK_ATTR, None)- tspan.attrib.pop(PARAGRAPH_SPACE_BEFORE_ATTR, None)try:extra = next(extras_iter)- line_break = next(line_break_iter)+ soft = next(soft_iter)except StopIteration:extra = 0.0- line_break = False- if line_break:- tspan.set(PARAGRAPH_LINE_BREAK_ATTR, "1")+ soft = False+ if soft:+ tspan.set(PARAGRAPH_SOFT_BREAK_ATTR, "1")elif extra > 1e-6:tspan.set(PARAGRAPH_SPACE_BEFORE_ATTR, format_number(extra))-def _set_runtime_contract_attrs(- elem: ET.Element,- contract: TextBlockContract,- *,- effective_mode: str,- reason: str,- line_index: int | None = None,- line_count: int | None = None,-) -> None:- """Attach rendering-neutral runtime provenance to a normalized text node."""- if contract.preserve_space:- elem.set("{http://www.w3.org/XML/1998/namespace}space", "preserve")- elem.set(RUNTIME_CONTRACT_ATTR, "1")- elem.set(RUNTIME_SOURCE_ID_ATTR, contract.element_id)- elem.set(RUNTIME_REQUESTED_MODE_ATTR, contract.requested_mode)- elem.set(RUNTIME_EFFECTIVE_MODE_ATTR, effective_mode)- elem.set(RUNTIME_MODE_REASON_ATTR, reason)- if line_index is not None:- elem.set(RUNTIME_LINE_INDEX_ATTR, str(line_index))- if line_count is not None:- elem.set(RUNTIME_LINE_COUNT_ATTR, str(line_count))---def _strip_public_contract_from_split_line(elem: ET.Element) -> None:- """Remove block/break author metadata after a block becomes line frames."""- elem.attrib.pop(TEXT_MODE_ATTR, None)- elem.attrib.pop(TEXT_BOUNDS_ATTR, None)- for descendant in elem.iter():- descendant.attrib.pop(TEXT_BREAK_ATTR, None)- descendant.attrib.pop(TEXT_JOIN_ATTR, None)---def _mark_split_lines(- new_texts: list[ET.Element],- contract: TextBlockContract,- *,- reason: str,-) -> None:- """Record one explicit source block on every independent runtime line."""- line_count = len(new_texts)- for line_index, line in enumerate(new_texts, start=1):- _strip_public_contract_from_split_line(line)- _set_runtime_contract_attrs(- line,- contract,- effective_mode="lines",- reason=reason,- line_index=line_index,- line_count=line_count,- )---def _emit_explicit_paragraph(- text_el: ET.Element,- contract: TextBlockContract,-) -> None:- """Normalize a validated explicit paragraph into the existing internal IR."""- if text_el.get("x") is None and contract.anchor_x is not None:- text_el.set("x", format_number(contract.anchor_x))- if text_el.get("y") is None and contract.anchor_y is not None:- text_el.set("y", format_number(contract.anchor_y))-- line_groups = [list(line.elements) for line in contract.visual_lines]- extras = [line.space_before for line in contract.visual_lines]- line_breaks = [- line.break_kind == "line"- for line in contract.visual_lines- ]- _emit_mergeable_paragraph(- text_el,- contract.base_line_height,- extras,- line_breaks,- line_groups,- )-- if contract.top_inset is not None:- text_el.set(PARAGRAPH_TOP_INSET_ATTR, format_number(contract.top_inset))- _set_runtime_contract_attrs(- text_el,- contract,- effective_mode="paragraph",- reason="explicit-paragraph",- line_count=len(contract.visual_lines),- )- for line_index, line in enumerate(list(text_el), start=1):- if line.tag != f"{{{SVG_NS}}}tspan":- continue- line.set(RUNTIME_LINE_INDEX_ATTR, str(line_index))- line.set(RUNTIME_LINE_COUNT_ATTR, str(len(contract.visual_lines)))--def flatten_text_with_tspans(tree: ET.ElementTree,merge_paragraphs: bool = False,- *,- enforce_native_carrier: bool = True,- source_name: str | None = None,) -> bool:"""Flatten multi-line tspan text into independent text nodes when needed.When ``merge_paragraphs`` is True, mergeable paragraph blocks (same x,- dy clustered around one base line-height) remain one <text>. Downstream- conversion emits one editable PowerPoint text frame with authored hard- line breaks and paragraph gaps. Default False preserves the original- behavior: every positioned tspan becomes its own <text>, matching the- SVG's strict line-layout contract. ``enforce_native_carrier=False`` is- reserved for that on-disk visual preview; PPTX export keeps the native- carrier guard.+ dy clustered around one base line-height) are kept as a single <text>+ so downstream conversion emits one editable PowerPoint text frame+ with multiple <a:p>. Default False preserves the original behavior:+ every line-break tspan becomes its own <text>, matching the SVG's+ pixel-fidelity contract."""root = tree.getroot()- if root in _NORMALIZED_ROOTS:- return False- contracts = validate_text_contracts(- root,- merge_paragraphs=merge_paragraphs,- enforce_native_carrier=enforce_native_carrier,- source_name=source_name,- )parent_map = {c: p for p in root.iter() for c in p}changed = False@@ -615,23 +483,13 @@ def flatten_text_with_tspans(for el in root.iter():if is_svg_tag(el, "text"):has_tspan_child = any(is_svg_tag(c, "tspan") for c in list(el))- if has_tspan_child or el in contracts:+ if has_tspan_child:candidates.append(el)for text_el in candidates:parent = parent_map.get(text_el)if parent is None:continue- contract = contracts.get(text_el)-- if (- contract is not None- and contract.requested_mode == "paragraph"- and merge_paragraphs- ):- _emit_explicit_paragraph(text_el, contract)- changed = True- continue# First check whether any tspan needs flattening (dy != 0 or has its own y attribute)needs_flatten = False@@ -644,59 +502,27 @@ def flatten_text_with_tspans(# If no tspan needs a line break, skip the entire text elementif not needs_flatten:- if contract is not None:- reason = (- "cli-no-merge"- if not merge_paragraphs- else "explicit-lines"- if contract.requested_mode == "lines"- else "legacy-auto-single-line"- )- _set_runtime_contract_attrs(- text_el,- contract,- effective_mode="lines",- reason=reason,- line_index=1,- line_count=1,- )- changed = Truecontinue# Paragraph fast-path (opt-in via merge_paragraphs=True): if the# children form a mergeable paragraph (same x, dy clustered around# one base line-height with optional paragraph gaps, no nested- # positional tspans), keep one <text> and let the downstream converter- # preserve hard line/paragraph boundaries in one frame. When disabled,- # every tspan gets its own independent <text> so the SVG's exact line- # layout is preserved in PowerPoint.- if (- merge_paragraphs- and (contract is None or contract.requested_mode == "auto")- ):+ # positional tspans), keep as one <text> and let the downstream+ # converter emit multiple <a:p> runs. When disabled, every tspan+ # gets its own independent <text> so the SVG's exact line layout+ # is preserved in PowerPoint.+ if merge_paragraphs:paragraph = _classify_paragraph_block(text_el, is_svg_tag, is_new_line_tspan)if paragraph is not None:- base_dy, extras, line_breaks, line_groups, synthetic_first = paragraph+ base_dy, extras, soft_breaks, line_groups, synthetic_first = paragraph_emit_mergeable_paragraph(text_el,base_dy,extras,- line_breaks,+ soft_breaks,line_groups,synthetic_first=synthetic_first,)- if contract is not None:- _set_runtime_contract_attrs(- text_el,- contract,- effective_mode="paragraph",- reason="legacy-auto-merge",- line_count=len(line_groups),- )- for line_index, line in enumerate(list(text_el), start=1):- if is_svg_tag(line, "tspan"):- line.set(RUNTIME_LINE_INDEX_ATTR, str(line_index))- line.set(RUNTIME_LINE_COUNT_ATTR, str(len(line_groups)))changed = Truecontinue@@ -748,15 +574,6 @@ def flatten_text_with_tspans(new_texts.append(ne)if new_texts:- if contract is not None:- if not merge_paragraphs:- reason = "cli-no-merge"- elif contract.requested_mode == "lines":- reason = "explicit-lines"- else:- reason = "legacy-auto-split"- _mark_split_lines(new_texts, contract, reason=reason)-# Replace original <text> with the list of new <text> nodestry:idx = list(parent).index(text_el)@@ -774,8 +591,6 @@ def flatten_text_with_tspans(parent.remove(text_el)changed = True- if changed:- _NORMALIZED_ROOTS.add(root)return changed@@ -834,12 +649,6 @@ def _create_text_element_from_line(tspan = tspans[0]content = collect_text_content(tspan)- if text_el.get(TEXT_MODE_ATTR) is not None:- xml_space_attr = "{http://www.w3.org/XML/1998/namespace}space"- child_xml_space = tspan.get(xml_space_attr) or tspan.get("xml:space")- if child_xml_space is not None:- ne.set(xml_space_attr, child_xml_space)-# Merge stylemerged_style = merge_styles(text_el.get("style"), tspan.get("style"))if merged_style:@@ -882,11 +691,7 @@ def process_svg_file(print(f"[WARN] Failed to parse {src_path}: {e}")return False- changed = flatten_text_with_tspans(- tree,- merge_paragraphs=merge_paragraphs,- source_name=Path(src_path).name,- )+ changed = flatten_text_with_tspans(tree, merge_paragraphs=merge_paragraphs)# Ensure destination directory existsos.makedirs(os.path.dirname(dst_path), exist_ok=True)@@ -962,9 +767,8 @@ def main() -> None:help=("Opt-in: merge mergeable paragraph blocks (same x, dy clustered ""around one base line-height) into a single <text> annotated for "- "downstream hard-line and paragraph conversion. Default off — "- "every line-break tspan becomes its own <text>, preserving SVG "- "pixel fidelity."+ "downstream multi-<a:p> conversion. Default off — every line-break "+ "tspan becomes its own <text>, preserving SVG pixel fidelity."),)diff --git a/skills/ppt-master/scripts/svg_quality_checker.py b/skills/ppt-master/scripts/svg_quality_checker.pyindex 057fe9ab..4a9a15d3 100644--- a/skills/ppt-master/scripts/svg_quality_checker.py+++ b/skills/ppt-master/scripts/svg_quality_checker.py@@ -22,7 +22,6 @@ from collections import Counter, defaultdictfrom xml.etree import ElementTree as ETfrom console_encoding import configure_utf8_stdio-from svg_to_pptx.text_contract import TextContractError, validate_text_contractsconfigure_utf8_stdio()@@ -1369,17 +1368,6 @@ class SVGQualityChecker:def _check_text_elements(self, content: str, root: ET.Element, result: Dict):"""Check text elements and wrapping methods"""- try:- validate_text_contracts(- root,- merge_paragraphs=True,- source_name=result.get('file'),- )- except TextContractError as exc:- result['errors'].extend(- diagnostic.format() for diagnostic in exc.diagnostics- )-# Count text and tspan elementstext_count = content.count('<text')tspan_count = content.count('<tspan')@@ -2349,8 +2337,6 @@ class SVGQualityChecker:def _categorize_issue(self, error_msg: str) -> str:"""Categorize issue type"""- if error_msg.startswith('[TEXT_'):- return 'Text contract'if 'Invalid XML' in error_msg:return 'XML well-formedness'elif 'viewBox' in error_msg:diff --git a/skills/ppt-master/scripts/svg_to_pptx/drawingml/context.py b/skills/ppt-master/scripts/svg_to_pptx/drawingml/context.pyindex 66a84e10..4d677e87 100644--- a/skills/ppt-master/scripts/svg_to_pptx/drawingml/context.py+++ b/skills/ppt-master/scripts/svg_to_pptx/drawingml/context.py@@ -21,7 +21,6 @@ class ShapeResult:xml: strbounds_emu: tuple[int, int, int, int] | None = None- trace_metadata: dict[str, Any] = field(default_factory=dict)@dataclass@@ -69,8 +68,8 @@ class ConvertContext:# Explicit sidecar group ids may override the legacy chrome-name heuristic.# Explicit structural layer/role/placeholder markers remain non-animatable.animation_group_overrides: frozenset[str] = frozenset()- # Default-on flag: merge eligible dy-stacked lines into one editable text- # frame while preserving authored line and paragraph boundaries.+ # Default-on flag: merge mergeable paragraph blocks into one editable+ # text frame with multiple <a:p>. Disable it for strict line fidelity.merge_paragraphs: bool = True# Explicit opt-in: convert data-pptx-native table/chart marker groups to# native PowerPoint graphicFrames. Default stays off to preserve SVG output.diff --git a/skills/ppt-master/scripts/svg_to_pptx/drawingml/converter.py b/skills/ppt-master/scripts/svg_to_pptx/drawingml/converter.pyindex 69174209..1a969d47 100644--- a/skills/ppt-master/scripts/svg_to_pptx/drawingml/converter.py+++ b/skills/ppt-master/scripts/svg_to_pptx/drawingml/converter.py@@ -5,7 +5,6 @@ from __future__ import annotationsimport base64import binasciiimport hashlib-import jsonimport mathimport refrom pathlib import Path@@ -159,117 +158,6 @@ def _txbody_metadata(elem: ET.Element) -> ET.Element | None:return None-def _text_carrier_metadata(elem: ET.Element) -> ET.Element | None:- for child in elem:- if (- child.tag.replace(f'{{{SVG_NS}}}', '') == 'metadata'- and child.get('data-pptx-part') == 'text-carrier'- ):- return child- return None---def _decode_text_carrier_nvsppr(metadata: ET.Element) -> ET.Element:- """Decode the versioned native carrier identity metadata."""- if metadata.get('data-pptx-text-carrier-version') != '1':- raise SvgNativeConversionError(- '[TEXT_NATIVE_CARRIER_VERSION] Unsupported text-carrier metadata version'- )- if metadata.get('data-pptx-text-carrier-eligible') != '1':- reason = metadata.get('data-pptx-text-carrier-reason') or 'unknown'- raise SvgNativeConversionError(- f'[TEXT_NATIVE_CARRIER_INELIGIBLE] Native text carrier is not eligible: {reason}'- )- if metadata.get('data-pptx-encoding') != 'base64':- raise SvgNativeConversionError(- '[TEXT_NATIVE_CARRIER_ENCODING] Eligible text carrier requires base64 nvSpPr'- )- try:- raw = base64.b64decode((metadata.text or '').strip(), validate=True)- nv_sp_pr = ET.fromstring(raw)- except (ValueError, binascii.Error, ET.ParseError) as exc:- raise SvgNativeConversionError(- f'[TEXT_NATIVE_CARRIER_XML] Invalid text-carrier nvSpPr: {exc}'- ) from exc- p_ns = 'http://schemas.openxmlformats.org/presentationml/2006/main'- if nv_sp_pr.tag != f'{{{p_ns}}}nvSpPr':- raise SvgNativeConversionError(- '[TEXT_NATIVE_CARRIER_XML] text-carrier payload must be p:nvSpPr'- )- if has_relationship_attributes(nv_sp_pr):- raise SvgNativeConversionError(- '[TEXT_NATIVE_CARRIER_REL] text-carrier nvSpPr contains relationships'- )- return nv_sp_pr---def _replace_native_text_carrier(- geometry: ShapeResult,- compiled_text: ShapeResult,- carrier_metadata: ET.Element,-) -> ShapeResult:- """Attach a rebuilt paragraph txBody to its original native p:sp."""- shape_match = re.search(r'<p:cNvPr id="(\d+)"', geometry.xml)- if shape_match is None:- raise SvgNativeConversionError(- '[TEXT_NATIVE_CARRIER_SHAPE] Native geometry has no output shape id'- )- output_shape_id = int(shape_match.group(1))- nv_sp_pr = _decode_text_carrier_nvsppr(carrier_metadata)- p_ns = 'http://schemas.openxmlformats.org/presentationml/2006/main'- a_ns = 'http://schemas.openxmlformats.org/drawingml/2006/main'- c_nv_pr = nv_sp_pr.find(f'{{{p_ns}}}cNvPr')- if c_nv_pr is None:- raise SvgNativeConversionError(- '[TEXT_NATIVE_CARRIER_XML] text-carrier nvSpPr has no p:cNvPr'- )- c_nv_pr.set('id', str(output_shape_id))- ET.register_namespace('p', p_ns)- ET.register_namespace('a', a_ns)- nv_sp_pr_xml = ET.tostring(nv_sp_pr, encoding='unicode')-- txbody_match = re.search(r'<p:txBody>.*?</p:txBody>', compiled_text.xml, re.DOTALL)- xfrm_match = re.search(r'<a:xfrm(?:\s[^>]*)?>.*?</a:xfrm>', compiled_text.xml, re.DOTALL)- if txbody_match is None or xfrm_match is None:- raise SvgNativeConversionError(- '[TEXT_NATIVE_CARRIER_COMPILE] Rebuilt paragraph lacks txBody or xfrm'- )- combined_xml, nv_count = re.subn(- r'<p:nvSpPr>.*?</p:nvSpPr>',- lambda _match: nv_sp_pr_xml,- geometry.xml,- count=1,- flags=re.DOTALL,- )- if nv_count != 1:- raise SvgNativeConversionError(- '[TEXT_NATIVE_CARRIER_SHAPE] Native geometry has no replaceable p:nvSpPr'- )- combined_xml, xfrm_count = re.subn(- r'<a:xfrm(?:\s[^>]*)?>.*?</a:xfrm>',- lambda _match: xfrm_match.group(0),- combined_xml,- count=1,- flags=re.DOTALL,- )- if xfrm_count != 1:- raise SvgNativeConversionError(- '[TEXT_NATIVE_CARRIER_SHAPE] Native geometry has no replaceable a:xfrm'- )- metadata = {- **compiled_text.trace_metadata,- 'txbody_restore': 'rebuilt-same-shape',- }- return _append_shape_text(- ShapeResult(- xml=combined_xml,- bounds_emu=compiled_text.bounds_emu,- trace_metadata=metadata,- ),- txbody_match.group(0),- )--_TXBODY_UNCHANGED_ATTR = 'data-pptx-runtime-txbody-unchanged'_PREVIEW_UNCHANGED_ATTR = 'data-pptx-runtime-preview-unchanged'@@ -376,7 +264,6 @@ def _append_shape_text(+ shape.xml[closing:]),bounds_emu=shape.bounds_emu,- trace_metadata=shape.trace_metadata,)@@ -530,74 +417,6 @@ def convert_g(elem: ET.Element, ctx: ConvertContext) -> ShapeResult | None:and child.get('data-pptx-part') not in allowed_partsfor child in elem)- visible_texts = [- descendant- for descendant in elem.iter()- if descendant.tag.replace(f'{{{SVG_NS}}}', '') == 'text'- and ''.join(descendant.itertext()).strip()- ]- contracted_texts = [- text- for text in visible_texts- if text.get('data-pptx-runtime-text-contract') == '1'- ]- if contracted_texts:- if len(contracted_texts) != 1 or len(visible_texts) != 1:- raise SvgNativeConversionError(- '[TEXT_NATIVE_TEXT_COUNT] Native carrier requires exactly one '- 'non-empty explicit text block'- )- contract_text = contracted_texts[0]- requested_mode = contract_text.get('data-pptx-runtime-requested-mode')- effective_mode = contract_text.get('data-pptx-runtime-effective-mode')- if requested_mode in {'auto', 'lines'} or effective_mode != 'paragraph':- raise SvgNativeConversionError(- '[TEXT_NATIVE_MODE_UNSUPPORTED] Native text carriers only '- 'support explicit paragraph mode'- )- carrier_meta = _text_carrier_metadata(elem)- if carrier_meta is None:- raise SvgNativeConversionError(- '[TEXT_NATIVE_CARRIER_METADATA_REQUIRED] Re-import the source '- 'PPTX to obtain versioned text-carrier metadata'- )- if len(carrier_children) != 1 or has_foreign_visual:- raise SvgNativeConversionError(- '[TEXT_NATIVE_CARRIER_STRUCTURE] Native text carrier has '- 'unsupported visible geometry children'- )- geometry_ctx = child_ctx- if transform and not native_subtree_active:- geometry_ctx = ctx.child(- 0, 0, 1.0, 1.0,- transform_matrix=parse_transform_matrix(transform),- filter_id=filter_id,- style_overrides=style_overrides,- opacity_multiplier=local_opacity,- )- geometry_result = convert_element(carrier_children[0], geometry_ctx)- ctx.sync_from_child(geometry_ctx)- if geometry_result is None:- raise SvgNativeConversionError(- '[TEXT_NATIVE_CARRIER_SHAPE] Native carrier has no geometry shape'- )- compiled_text = convert_text(contract_text, child_ctx)- ctx.sync_from_child(child_ctx)- if compiled_text is None:- raise SvgNativeConversionError(- '[TEXT_NATIVE_CARRIER_COMPILE] Explicit paragraph compiled empty'- )- restored = _replace_native_text_carrier(- geometry_result,- compiled_text,- carrier_meta,- )- if should_animate_group and elem_id:- shape_match = re.search(r'<p:cNvPr id="(\d+)"', restored.xml)- if shape_match:- ctx.anim_targets.append((int(shape_match.group(1)), elem_id))- return restored-native_text = _decode_unchanged_txbody(elem, txbody_meta)if len(carrier_children) == 1 and not has_foreign_visual and native_text is not None:geometry_ctx = child_ctx@@ -658,11 +477,7 @@ def convert_g(elem: ET.Element, ctx: ConvertContext) -> ShapeResult | None:shape_match = re.search(r'<p:cNvPr id="(\d+)"', child_results[0].xml)if shape_match:ctx.anim_targets.append((int(shape_match.group(1)), elem_id))- child_result = child_results[0]- return ShapeResult(- xml=child_result.xml,- bounds_emu=child_result.bounds_emu,- )+ return child_results[0]# Multiple children, or a top-level semantic one-child group: wrap in# <p:grpSp> so PowerPoint can animate the group as one unit.@@ -1064,72 +879,6 @@ def _geometry_trace_metadata(elem: ET.Element, result: ShapeResult) -> dict[str,return {'output_geometry': 'unknown', 'fidelity': 'visual-only'}-def _aggregate_text_contract_trace(- events: list[dict[str, Any]],-) -> list[dict[str, Any]]:- """Aggregate per-shape text events into stable source-block mappings."""- grouped: dict[str, list[dict[str, Any]]] = {}- for event in events:- if not event.get('text_contract'):- continue- source_id = event.get('source_text_id')- if not isinstance(source_id, str) or not source_id:- continue- grouped.setdefault(source_id, []).append(event)-- blocks: list[dict[str, Any]] = []- for source_id, source_events in grouped.items():- ordered = sorted(- source_events,- key=lambda event: int(event.get('line_index') or 1),- )- shape_ids = [- int(event['shape_id'])- for event in ordered- if event.get('shape_id') is not None- ]- bounds = [- event['bounds_emu']- for event in ordered- if isinstance(event.get('bounds_emu'), list)- and len(event['bounds_emu']) == 4- ]- union_bounds = None- if bounds:- union_bounds = [- min(value[0] for value in bounds),- min(value[1] for value in bounds),- max(value[2] for value in bounds),- max(value[3] for value in bounds),- ]- content_hashes = [- str(event.get('content_sha256'))- for event in ordered- if event.get('content_sha256')- ]- if len(content_hashes) == 1:- content_sha256 = content_hashes[0]- else:- payload = json.dumps(content_hashes, separators=(',', ':'))- content_sha256 = hashlib.sha256(payload.encode('utf-8')).hexdigest()- blocks.append({- 'source': {'element_id': source_id},- 'requested_mode': ordered[0].get('requested_mode'),- 'effective_mode': ordered[0].get('effective_mode'),- 'resolution_reason': ordered[0].get('resolution_reason'),- 'visual_line_count': ordered[0].get('visual_line_count'),- 'output_shape_ids': shape_ids,- 'output_shape_count': len(shape_ids),- 'paragraph_count': sum(int(event.get('paragraph_count') or 0) for event in ordered),- 'break_count': sum(int(event.get('break_count') or 0) for event in ordered),- 'slide_bounds_emu': union_bounds,- 'content_sha256': content_sha256,- 'txbody_restore': ordered[0].get('txbody_restore', 'not-applicable'),- 'readback': {'status': 'passed'},- })- return blocks--def convert_element(elem: ET.Element, ctx: ConvertContext) -> ShapeResult | None:"""Dispatch an SVG element to the appropriate converter."""tag = elem.tag.replace(f'{{{SVG_NS}}}', '')@@ -1190,7 +939,6 @@ def convert_element(elem: ET.Element, ctx: ConvertContext) -> ShapeResult | Noneif result.bounds_emu is not None:metadata['bounds_emu'] = list(result.bounds_emu)metadata.update(_geometry_trace_metadata(elem, result))- metadata.update(result.trace_metadata)trace('native', **metadata)else:trace('skip', reason='empty-or-non-rendering')@@ -1284,9 +1032,10 @@ def convert_svg_to_slide_shapes(svg_path: Path to the SVG file.slide_num: Slide number (for naming).verbose: Print progress info.- merge_paragraphs: When True, eligible dy-stacked lines become one- editable text frame with authored line and paragraph boundaries.- Disable it to keep one textbox per visual line.+ merge_paragraphs: When True, mergeable paragraph blocks (same x,+ dy clustered around one base line-height) become a single+ editable text frame with multiple <a:p>. Disable it to preserve+ the SVG's exact line layout (one textbox per line).image_optimize: Downsample oversized raster images for PPTX export.image_max_dimension: Maximum optimized image dimension in pixels.image_sizing: ``cap`` to only cap source dimensions, ``display`` to@@ -1419,14 +1168,10 @@ def convert_svg_to_slide_shapes(# and an x-anchored tspan would render in the wrong column. finalize_svg# does the same flattening on disk; doing it here keeps native pptx output# correct when reading raw svg_output/.- # merge_paragraphs additionally keeps eligible blocks as one annotated- # <text> for downstream hard-line and paragraph conversion.+ # merge_paragraphs additionally folds mergeable paragraph blocks into a+ # single annotated <text> for downstream multi-<a:p> conversion.from ..tspan_flattener import flatten_positional_tspans- flattened = flatten_positional_tspans(- tree,- merge_paragraphs=merge_paragraphs,- source_name=svg_path.name,- )+ flattened = flatten_positional_tspans(tree, merge_paragraphs=merge_paragraphs)if flattened:trace_steps.append({'action': 'flatten-positional-tspans',@@ -1571,7 +1316,6 @@ def convert_svg_to_slide_shapes(},'preprocess': trace_steps,'events': trace_events or [],- 'text_blocks': _aggregate_text_contract_trace(trace_events or []),})shapes_xml = '\n'.join(shapes)diff --git a/skills/ppt-master/scripts/svg_to_pptx/drawingml/elements.py b/skills/ppt-master/scripts/svg_to_pptx/drawingml/elements.pyindex 623db9f7..a6a94801 100644--- a/skills/ppt-master/scripts/svg_to_pptx/drawingml/elements.py+++ b/skills/ppt-master/scripts/svg_to_pptx/drawingml/elements.py@@ -6,10 +6,8 @@ import base64import binasciiimport hashlibimport io-import jsonimport mathimport re-from decimal import Decimal, ROUND_HALF_EVENfrom pathlib import Pathfrom typing import Anyfrom urllib.parse import unquote_to_bytes@@ -38,7 +36,7 @@ from .utils import (rect_to_dml_xfrm,combine_opacity, parse_hex_color, parse_svg_color,resolve_url_id, get_effective_filter_id,- parse_inline_style, parse_font_family, estimate_text_width,+ parse_inline_style, parse_font_family, is_cjk_char, estimate_text_width,detect_text_lang, font_px_to_hpt, resolve_text_run_fonts,matrix_multiply, parse_transform_matrix, transform_point, _xml_escape,)@@ -1792,9 +1790,6 @@ def _strip_leading_chars_from_runs(stripped: list[dict[str, Any]] = []remaining = char_countfor run in runs:- if run.get('_break'):- stripped.append(run)- continuetext = str(run.get('text', ''))if remaining >= len(text):remaining -= len(text)@@ -2029,11 +2024,7 @@ def _collect_tspan_runs(if tspan.text:t = _normalize_text(tspan.text, preserve_space=child_preserve_space)if t:- runs.append({- **own_attrs,- '_preserve_space': child_preserve_space,- 'text': t,- })+ runs.append({**own_attrs, 'text': t})for child in tspan:child_tag = child.tag.replace(f'{{{SVG_NS}}}', '')@@ -2042,11 +2033,7 @@ def _collect_tspan_runs(if child.tail:t = _normalize_text(child.tail, preserve_space=child_preserve_space)if t:- runs.append({- **own_attrs,- '_preserve_space': child_preserve_space,- 'text': t,- })+ runs.append({**own_attrs, 'text': t})return runs@@ -2054,8 +2041,6 @@ def _collect_tspan_runs(def _build_text_runs(elem: ET.Element,parent_attrs: dict[str, Any],- *,- respect_preserved_edges: bool = False,) -> list[dict[str, Any]]:"""Build a list of text runs from a <text> element, handling <tspan> children.@@ -2069,11 +2054,7 @@ def _build_text_runs(if elem.text:t = _normalize_text(elem.text, preserve_space=preserve_space)if t:- runs.append({- **parent_attrs,- '_preserve_space': preserve_space,- 'text': t,- })+ runs.append({**parent_attrs, 'text': t})for child in elem:child_tag = child.tag.replace(f'{{{SVG_NS}}}', '')@@ -2082,19 +2063,13 @@ def _build_text_runs(if child.tail:t = _normalize_text(child.tail, preserve_space=preserve_space)if t:- runs.append({- **parent_attrs,- '_preserve_space': preserve_space,- 'text': t,- })+ runs.append({**parent_attrs, 'text': t})# Strip the paragraph's overall leading / trailing whitespace once unless# xml:space="preserve" asks us to keep source indentation.if runs and not preserve_space:- if not respect_preserved_edges or not runs[0].get('_preserve_space'):- runs[0]['text'] = runs[0]['text'].lstrip(' ')- if not respect_preserved_edges or not runs[-1].get('_preserve_space'):- runs[-1]['text'] = runs[-1]['text'].rstrip(' ')+ runs[0]['text'] = runs[0]['text'].lstrip(' ')+ runs[-1]['text'] = runs[-1]['text'].rstrip(' ')runs = [r for r in runs if r['text']]return runs@@ -2217,152 +2192,6 @@ def _build_run_xml(</a:r>'''-_TEXT_CONTRACT_MARK_ATTR = 'data-pptx-runtime-text-contract'-_TEXT_CONTRACT_SOURCE_ATTR = 'data-pptx-runtime-source-text-id'-_TEXT_CONTRACT_REQUESTED_ATTR = 'data-pptx-runtime-requested-mode'-_TEXT_CONTRACT_EFFECTIVE_ATTR = 'data-pptx-runtime-effective-mode'-_TEXT_CONTRACT_LINE_INDEX_ATTR = 'data-pptx-runtime-line-index'-_TEXT_CONTRACT_LINE_COUNT_ATTR = 'data-pptx-runtime-line-count'-_TEXT_CONTRACT_REASON_ATTR = 'data-pptx-runtime-mode-reason'-_TEXT_CONTRACT_TOP_INSET_ATTR = 'data-paragraph-top-inset'-_PARAGRAPH_LINE_BREAK_ATTR = 'data-paragraph-line-break'---def _contract_round(- value: float,- multiplier: int,- *factors: float,-) -> int:- """Apply decimal ties-to-even rounding for explicit contract metrics."""- scaled = Decimal(str(value)) * Decimal(multiplier)- for factor in factors:- scaled *= Decimal(str(factor))- return int(scaled.to_integral_value(rounding=ROUND_HALF_EVEN))---def _trim_explicit_paragraph_edges(- paragraphs: list[list[dict[str, Any]]],-) -> None:- """Trim only the block's outer whitespace, preserving join boundaries."""- text_runs = [- run- for paragraph in paragraphs- for run in paragraph- if not run.get('_break') and run.get('text')- ]- if not text_runs:- return- if not text_runs[0].get('_preserve_space'):- text_runs[0]['text'] = str(text_runs[0]['text']).lstrip(' ')- if not text_runs[-1].get('_preserve_space'):- text_runs[-1]['text'] = str(text_runs[-1]['text']).rstrip(' ')- for paragraph in paragraphs:- paragraph[:] = [- run for run in paragraph- if run.get('_break') or run.get('text')- ]---def _append_explicit_soft_join(- paragraph: list[dict[str, Any]],- next_runs: list[dict[str, Any]],- join_kind: str,-) -> None:- """Append an explicit soft-wrapped line using the declared join policy."""- if join_kind == 'space' and paragraph and next_runs:- previous_text = str(paragraph[-1].get('text', ''))- next_text = str(next_runs[0].get('text', ''))- if (- previous_text- and next_text- and not previous_text[-1].isspace()- and not next_text[0].isspace()- ):- paragraph[-1] = {**paragraph[-1], 'text': previous_text + ' '}- paragraph.extend(next_runs)---def _paragraph_content_sha256(- paragraphs: list[list[dict[str, Any]]],- bullets: list[dict[str, Any] | None] | None = None,-) -> str:- """Hash paragraph/run/break structure without copying text into traces."""- tokens: list[list[dict[str, str]]] = []- paragraph_bullets = bullets or [None] * len(paragraphs)- for paragraph, bullet in zip(paragraphs, paragraph_bullets):- paragraph_tokens: list[dict[str, str]] = []- if bullet:- paragraph_tokens.append({- 'kind': 'bullet',- 'text': str(bullet.get('char', '•')),- })- for item in paragraph:- if item.get('_break'):- paragraph_tokens.append({'kind': 'break'})- else:- paragraph_tokens.append({- 'kind': 'run',- 'text': str(item.get('text', '')),- })- tokens.append(paragraph_tokens)- payload = json.dumps(tokens, ensure_ascii=False, separators=(',', ':'))- return hashlib.sha256(payload.encode('utf-8')).hexdigest()---def _validate_compiled_text_shape(- shape_xml: str,- *,- expected_paragraphs: int,- expected_breaks: int,- expected_bounds: tuple[int, int, int, int],-) -> None:- """Reparse the generated shape and verify contract-owned structure."""- wrapper = (- '<root '- 'xmlns:a="http://schemas.openxmlformats.org/drawingml/2006/main" '- 'xmlns:p="http://schemas.openxmlformats.org/presentationml/2006/main">'- + shape_xml- + '</root>'- )- try:- root = ET.fromstring(wrapper)- except ET.ParseError as exc:- raise ValueError(f'[TEXT_READBACK_XML] Invalid generated text XML: {exc}') from exc- ns = {- 'a': 'http://schemas.openxmlformats.org/drawingml/2006/main',- 'p': 'http://schemas.openxmlformats.org/presentationml/2006/main',- }- shape = root.find('p:sp', ns)- if shape is None:- raise ValueError('[TEXT_READBACK_SHAPE] Generated text is not one p:sp')- paragraphs = shape.findall('./p:txBody/a:p', ns)- breaks = shape.findall('.//a:br', ns)- if len(paragraphs) != expected_paragraphs:- raise ValueError(- '[TEXT_READBACK_PARAGRAPHS] Generated paragraph count '- f'{len(paragraphs)} != {expected_paragraphs}'- )- if len(breaks) != expected_breaks:- raise ValueError(- '[TEXT_READBACK_BREAKS] Generated line-break count '- f'{len(breaks)} != {expected_breaks}'- )- off = shape.find('./p:spPr/a:xfrm/a:off', ns)- ext = shape.find('./p:spPr/a:xfrm/a:ext', ns)- if off is None or ext is None:- raise ValueError('[TEXT_READBACK_BOUNDS] Generated text has no complete a:xfrm')- actual = (- int(off.get('x', '0')),- int(off.get('y', '0')),- int(off.get('x', '0')) + int(ext.get('cx', '0')),- int(off.get('y', '0')) + int(ext.get('cy', '0')),- )- if actual != expected_bounds:- raise ValueError(- f'[TEXT_READBACK_BOUNDS] Generated bounds {actual} != {expected_bounds}'- )--def convert_text(elem: ET.Element, ctx: ConvertContext) -> ShapeResult | None:"""Convert SVG <text> to DrawingML text shape with multi-run support."""x = ctx_x(svg_length_x(elem.get('x'), ctx), ctx)@@ -2413,73 +2242,60 @@ def convert_text(elem: ET.Element, ctx: ConvertContext) -> ShapeResult | None:'stroke_opacity': stroke_opacity,}- # Legacy paragraphs continue to use the internal data-paragraph-* IR.- # Explicit contracts additionally preserve author-declared break/join- # semantics and may remain a paragraph even when they contain one line.- explicit_contract = elem.get(_TEXT_CONTRACT_MARK_ATTR) == '1'- requested_mode = elem.get(_TEXT_CONTRACT_REQUESTED_ATTR)- effective_mode = elem.get(_TEXT_CONTRACT_EFFECTIVE_ATTR)- explicit_paragraph = (- explicit_contract- and requested_mode == 'paragraph'- and effective_mode == 'paragraph'- )+ # Paragraph mode: flatten_tspan marks <text> with data-paragraph-line-height+ # when its direct-child tspans form a mergeable paragraph (same x, dy+ # clustered around one base line-height). Each direct tspan becomes one+ # <a:p> so the paragraph survives as a single editable text frame.+ # Per-line data-paragraph-space-before encodes paragraph gaps (extra dy+ # above the base line-height) for the corresponding <a:p>.+ # Paragraph mode is controlled by ctx.merge_paragraphs. When off, ignore+ # any data-paragraph-* markers and fall through to the original+ # one-text-per-tspan path so the SVG's pixel layout is preserved.line_height_attr = elem.get('data-paragraph-line-height') if ctx.merge_paragraphs else Noneline_height_px = _f(line_height_attr) if line_height_attr is not None else Noneparagraph_runs: list[list[dict[str, Any]]] | None = Noneparagraph_space_before: list[float] = []paragraph_bullets: list[dict[str, Any] | None] = []- single_bullet: dict[str, Any] | None = None# Per-tspan widths (visual lines as the deck author drew them) regardless# of how many merge into one <a:p>; used to size the textbox so PowerPoint# has room to wrap text to the SVG's original line widths.visual_line_widths: list[float] = []- if explicit_paragraph or (line_height_px is not None and line_height_px > 0):+ if line_height_px is not None and line_height_px > 0:preserve_space = _preserves_space(elem)paragraph_runs = []for child in elem:if child.tag != f'{{{SVG_NS}}}tspan':continueline_runs = _collect_tspan_runs(child, parent_attrs, preserve_space)- if line_runs and not preserve_space and not explicit_paragraph:+ if line_runs and not preserve_space:line_runs[0]['text'] = line_runs[0]['text'].lstrip(' ')line_runs[-1]['text'] = line_runs[-1]['text'].rstrip(' ')line_runs = [r for r in line_runs if r['text']]if not line_runs:continuevisual_line_widths.append(_estimate_bullet_line_width(line_runs))- if explicit_paragraph:- break_kind = child.get('data-pptx-break')- if not paragraph_runs:- paragraph_runs.append(line_runs)- paragraph_space_before.append(0.0)- continue- if break_kind == 'soft':- _append_explicit_soft_join(- paragraph_runs[-1],- line_runs,- child.get('data-pptx-join') or 'none',- )- continue- if break_kind == 'line':- paragraph_runs[-1].append({'_break': True})- paragraph_runs[-1].extend(line_runs)- continue+ soft_break = child.get('data-paragraph-soft-break') == '1'+ if soft_break and paragraph_runs:+ # Append to the previous paragraph. A Latin line-wrap needs a+ # space to keep two words apart (SVG used a dy break, not+ # punctuation); CJK wraps mid-sentence with no inter-character+ # space, so a joining space there is a spurious artifact.+ prev = paragraph_runs[-1]+ prev_text = prev[-1]['text'] if prev else ''+ next_text = line_runs[0]['text']+ boundary_is_cjk = (+ (prev_text and is_cjk_char(prev_text[-1]))+ or (next_text and is_cjk_char(next_text[0]))+ )+ if prev and not prev_text.endswith(' ') \+ and not next_text.startswith(' ') \+ and not boundary_is_cjk:+ prev[-1] = {**prev[-1], 'text': prev_text + ' '}+ prev.extend(line_runs)+ else:paragraph_runs.append(line_runs)sb_attr = child.get('data-paragraph-space-before')paragraph_space_before.append(_f(sb_attr) if sb_attr else 0.0)- continue-- line_break = child.get(_PARAGRAPH_LINE_BREAK_ATTR) == '1'- if line_break and paragraph_runs:- paragraph_runs[-1].append({'_break': True})- paragraph_runs[-1].extend(line_runs)- continue- paragraph_runs.append(line_runs)- sb_attr = child.get('data-paragraph-space-before')- paragraph_space_before.append(_f(sb_attr) if sb_attr else 0.0)- if explicit_paragraph and paragraph_runs and not preserve_space:- _trim_explicit_paragraph_edges(paragraph_runs)if not paragraph_runs:paragraph_runs = Noneparagraph_space_before = []@@ -2493,18 +2309,9 @@ def convert_text(elem: ET.Element, ctx: ConvertContext) -> ShapeResult | None:paragraph_runs = stripped_paragraphsif paragraph_runs is not None:- runs = [- run- for paragraph in paragraph_runs- for run in paragraph- if not run.get('_break')- ]+ runs = [r for line in paragraph_runs for r in line]else:- runs = _build_text_runs(- elem,- parent_attrs,- respect_preserved_edges=explicit_contract,- )+ runs = _build_text_runs(elem, parent_attrs)runs, single_bullet = _extract_text_bullet(runs)if not runs:@@ -2517,16 +2324,15 @@ def convert_text(elem: ET.Element, ctx: ConvertContext) -> ShapeResult | None:# Estimate text dimensionsif paragraph_runs is not None:# Use the WIDEST visual line (per-tspan as the deck author drew it),- # not the joined-up paragraph: explicit soft joins concatenate many- # visual lines into one <a:p>, and measuring the joined string would+ # not the joined-up paragraph: soft-broken paragraphs concatenate+ # many lines into one <a:p>, and measuring the joined string would# blow the textbox past the canvas.text_width = max(visual_line_widths) if visual_line_widths else 0.0# Total height assumes the visual line count from the SVG source;# if PowerPoint wraps to more or fewer lines after the user resizes,# the user resizes the height accordingly.- effective_line_height = line_height_px or 0.0text_height = (- effective_line_height * (len(visual_line_widths) - 1)+ line_height_px * (len(visual_line_widths) - 1)+ sum(paragraph_space_before)+ font_size * 1.5)@@ -2536,38 +2342,19 @@ def convert_text(elem: ET.Element, ctx: ConvertContext) -> ShapeResult | None:fs_px = float(runs[0].get('font_size', font_size)) if runs else font_sizetext_width += _bullet_margin_px(single_bullet, fs_px)text_height = font_size * 1.5- explicit_bounds: tuple[float, float, float, float] | None = None- if explicit_paragraph:- bounds_raw = elem.get('data-pptx-text-bounds')- try:- values = [float(value) for value in re.split(r'[\s,]+', bounds_raw or '') if value]- except ValueError as exc:- raise ValueError('[TEXT_BOUNDS_FORMAT] Invalid explicit text bounds') from exc- if len(values) != 4:- raise ValueError('[TEXT_BOUNDS_FORMAT] Explicit text bounds require x y width height')- explicit_bounds = (values[0], values[1], values[2], values[3])-- if explicit_bounds is not None:- raw_box_x, raw_box_y, raw_box_w, raw_box_h = explicit_bounds- box_x = ctx_x(raw_box_x, ctx)- box_y = ctx_y(raw_box_y, ctx)- box_w = ctx_w(raw_box_w, ctx)- box_h = ctx_h(raw_box_h, ctx)- padding = 0.0- else:- padding = _textbox_padding(font_size)+ padding = _textbox_padding(font_size)- # Adjust position based on text-anchor- if text_anchor == 'middle':- box_x = x - text_width / 2 - padding- elif text_anchor == 'end':- box_x = x - text_width - padding- else:- box_x = x - padding+ # Adjust position based on text-anchor+ if text_anchor == 'middle':+ box_x = x - text_width / 2 - padding+ elif text_anchor == 'end':+ box_x = x - text_width - padding+ else:+ box_x = x - padding- box_y = y - font_size * 0.85- box_w = text_width + padding * 2- box_h = text_height + padding+ box_y = y - font_size * 0.85+ box_w = text_width + padding * 2+ box_h = text_height + paddingtext_transform = elem.get('transform', '')if text_transform and 'rotate' not in text_transform and not ctx.use_transform_matrix:@@ -2645,53 +2432,18 @@ def convert_text(elem: ET.Element, ctx: ConvertContext) -> ShapeResult | None:if paragraph_runs is not None:# SVG dy(px) -> hundredths-of-a-point: dy_pt = dy_px * 0.75, then x100.- metric_scale = ctx.scale_y if explicit_paragraph else 1.0- if line_height_px is not None and line_height_px > 0:- if explicit_paragraph:- line_spc_val = _contract_round(- line_height_px,- FONT_PX_TO_HUNDREDTHS_PT,- metric_scale,- )- else:- line_spc_val = round(- line_height_px * FONT_PX_TO_HUNDREDTHS_PT- )- ln_spc_xml = f'<a:lnSpc><a:spcPts val="{line_spc_val}"/></a:lnSpc>'- else:- ln_spc_xml = ''+ line_spc_val = round(line_height_px * FONT_PX_TO_HUNDREDTHS_PT)+ ln_spc_xml = f'<a:lnSpc><a:spcPts val="{line_spc_val}"/></a:lnSpc>'paragraph_xml_chunks = []- for line, extra_px, bullet in zip(- paragraph_runs,- paragraph_space_before,- paragraph_bullets,- ):+ for line, extra_px, bullet in zip(paragraph_runs, paragraph_space_before, paragraph_bullets):spc_bef_xml = ''if extra_px > 0:- if explicit_paragraph:- spc_bef_val = _contract_round(- extra_px,- FONT_PX_TO_HUNDREDTHS_PT,- metric_scale,- )- else:- spc_bef_val = round(- extra_px * FONT_PX_TO_HUNDREDTHS_PT- )+ spc_bef_val = round(extra_px * FONT_PX_TO_HUNDREDTHS_PT)spc_bef_xml = f'<a:spcBef><a:spcPts val="{spc_bef_val}"/></a:spcBef>'- runs_inner = '\n'.join(- '<a:br/>'- if item.get('_break')- else _build_run_xml(item, fonts, ctx, text_effect_xml)- for item in line- )- first_run = next((item for item in line if not item.get('_break')), None)+ runs_inner = '\n'.join(_build_run_xml(r, fonts, ctx, text_effect_xml) for r in line)p_pr_xml = _paragraph_pr_xml(algn=algn,- font_size=(- float(first_run.get('font_size', font_size))- if first_run is not None else font_size- ),+ font_size=float(line[0].get('font_size', font_size)) if line else font_size,body_xml=f'{ln_spc_xml}{spc_bef_xml}',bullet=bullet,ctx=ctx,@@ -2710,28 +2462,20 @@ def convert_text(elem: ET.Element, ctx: ConvertContext) -> ShapeResult | None:)paragraphs_xml = f'<a:p>\n{p_pr_xml}\n{runs_xml}\n</a:p>'- if explicit_bounds is not None:- off_x = _contract_round(box_x, 9525)- off_y = _contract_round(box_y, 9525)- ext_cx = _contract_round(box_w, 9525)- ext_cy = _contract_round(box_h, 9525)- else:- off_x = px_to_emu(box_x)- off_y = px_to_emu(box_y)- ext_cx = px_to_emu(box_w)- ext_cy = px_to_emu(box_h)-- # Hard breaks preserve the authored initial layout; wrap="square" keeps the- # frame editable so deleting a break or changing text restores normal- # PowerPoint wrapping. Paragraph frames avoid spAutoFit, which could resize- # them past their SVG bounds. Single-line text keeps its tight-frame path.+ off_x = px_to_emu(box_x)+ off_y = px_to_emu(box_y)+ ext_cx = px_to_emu(box_w)+ ext_cy = px_to_emu(box_h)++ # Paragraph mode: wrap="square" so text reflows when the user resizes,+ # but NO spAutoFit — otherwise PowerPoint expands the frame to fit a+ # long joined-up <a:p> on one line, blowing past the canvas. The cx we+ # write below is the longest source SVG line without single-line renderer+ # headroom; PowerPoint wraps long paragraphs inside this design width.+ # Single-line text keeps wrap="none" + spAutoFit for tight fidelity.if paragraph_runs is not None:- top_inset = 0- if explicit_paragraph:- top_inset_px = _f(elem.get(_TEXT_CONTRACT_TOP_INSET_ATTR))- top_inset = _contract_round(top_inset_px, 9525, ctx.scale_y)body_pr_xml = (- f'<a:bodyPr wrap="square" lIns="0" tIns="{top_inset}" rIns="0" bIns="0" '+ '<a:bodyPr wrap="square" lIns="0" tIns="0" rIns="0" bIns="0" ''anchor="t" anchorCtr="0"/>')else:@@ -2740,7 +2484,7 @@ def convert_text(elem: ET.Element, ctx: ConvertContext) -> ShapeResult | None:'anchor="t" anchorCtr="0">\n<a:spAutoFit/>\n</a:bodyPr>')- shape_xml = f'''<p:sp>+ return ShapeResult(xml=f'''<p:sp><p:nvSpPr><p:cNvPr id="{shape_id}" name="TextBox {shape_id}"/><p:cNvSpPr txBox="1"/><p:nvPr/>@@ -2758,57 +2502,7 @@ def convert_text(elem: ET.Element, ctx: ConvertContext) -> ShapeResult | None:<a:lstStyle/>{paragraphs_xml}</p:txBody>-</p:sp>'''- bounds_emu = (off_x, off_y, off_x + ext_cx, off_y + ext_cy)- paragraph_count = len(paragraph_runs) if paragraph_runs is not None else 1- break_count = (- sum(- 1- for paragraph in paragraph_runs- for item in paragraph- if item.get('_break')- )- if paragraph_runs is not None else 0- )- trace_metadata: dict[str, Any] = {}- if explicit_contract:- effective_paragraphs = (- paragraph_runs- if paragraph_runs is not None- else [runs]- )- effective_bullets = (- paragraph_bullets- if paragraph_runs is not None- else [single_bullet]- )- trace_metadata = {- 'text_contract': True,- 'source_text_id': elem.get(_TEXT_CONTRACT_SOURCE_ATTR),- 'requested_mode': requested_mode,- 'effective_mode': effective_mode,- 'resolution_reason': elem.get(_TEXT_CONTRACT_REASON_ATTR),- 'visual_line_count': int(elem.get(_TEXT_CONTRACT_LINE_COUNT_ATTR) or '1'),- 'line_index': int(elem.get(_TEXT_CONTRACT_LINE_INDEX_ATTR) or '1'),- 'paragraph_count': paragraph_count,- 'break_count': break_count,- 'content_sha256': _paragraph_content_sha256(- effective_paragraphs,- effective_bullets,- ),- 'readback': {'status': 'passed'},- }- _validate_compiled_text_shape(- shape_xml,- expected_paragraphs=paragraph_count,- expected_breaks=break_count,- expected_bounds=bounds_emu,- )- return ShapeResult(- xml=shape_xml,- bounds_emu=bounds_emu,- trace_metadata=trace_metadata,- )+</p:sp>''', bounds_emu=(off_x, off_y, off_x + ext_cx, off_y + ext_cy))# ---------------------------------------------------------------------------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.pyindex 8033a9e8..b9e0d85d 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@@ -108,7 +108,6 @@ from .template_structure import (template_placeholder_bindings,)from .template_validation import validate_pptx_template_package-from .text_validation import validate_pptx_text_contractsSLIDE_LAYOUT_REL_TYPE = ("http://schemas.openxmlformats.org/officeDocument/2006/relationships/slideLayout"@@ -4132,12 +4131,9 @@ def create_pptx_with_native_svg(separators=(',', ':'),)animation_rng = random.Random(animation_seed)- runtime_trace: list[dict[str, Any]] | None = [] if use_native_shapes else None- conversion_trace: list[dict[str, Any]] | None = (- runtime_trace if conversion_trace_path else None- )+ conversion_trace: list[dict[str, Any]] | None = [] if conversion_trace_path else Nonestructure_trace: list[dict[str, Any]] | None = (- runtime_trace+ []if use_native_shapes and pptx_structure in {"baseline", "structured", "preserve"}else None)@@ -4217,7 +4213,9 @@ def create_pptx_with_native_svg(animation_group_overrides=explicit_animation_groups,theme_font_spec=active_theme_font_spec,theme_color_spec=active_theme_color_spec,- trace_out=runtime_trace,+ trace_out=conversion_trace+ if conversion_trace is not None+ else structure_trace,))# Order matters: OOXML schema requires <p:transition>@@ -4802,12 +4800,6 @@ def create_pptx_with_native_svg(raise RuntimeError(f'PPTX animation package validation failed: {exc}') from exc- try:- validate_pptx_text_contracts(temp_output_path, runtime_trace)- except ValueError as exc:- raise RuntimeError(- f'PPTX text contract validation failed: {exc}'- ) from excshutil.move(str(temp_output_path), str(output_path))permission_warnings = _relax_output_permissions(output_path)diff --git a/skills/ppt-master/scripts/svg_to_pptx/pptx_package/text_validation.py b/skills/ppt-master/scripts/svg_to_pptx/pptx_package/text_validation.pydeleted file mode 100644index 03a43f4b..00000000--- a/skills/ppt-master/scripts/svg_to_pptx/pptx_package/text_validation.py+++ /dev/null@@ -1,176 +0,0 @@-"""Post-package validation for explicit SVG text-layout contracts."""--from __future__ import annotations--import hashlib-import json-import zipfile-from pathlib import Path-from typing import Any-from xml.etree import ElementTree as ET---P_NS = 'http://schemas.openxmlformats.org/presentationml/2006/main'-A_NS = 'http://schemas.openxmlformats.org/drawingml/2006/main'-NS = {'p': P_NS, 'a': A_NS}---def _shape_id(shape: ET.Element) -> int | None:- c_nv_pr = shape.find('./p:nvSpPr/p:cNvPr', NS)- if c_nv_pr is None:- c_nv_pr = shape.find('./p:nvGrpSpPr/p:cNvPr', NS)- if c_nv_pr is None:- return None- try:- return int(c_nv_pr.get('id', ''))- except ValueError:- return None---def _shape_map(root: ET.Element) -> dict[int, ET.Element]:- shapes: dict[int, ET.Element] = {}- for element in root.iter():- if element.tag not in {- f'{{{P_NS}}}sp',- f'{{{P_NS}}}grpSp',- }:- continue- shape_id = _shape_id(element)- if shape_id is not None:- shapes[shape_id] = element- return shapes---def _shape_bounds(shape: ET.Element) -> list[int]:- off = shape.find('./p:spPr/a:xfrm/a:off', NS)- ext = shape.find('./p:spPr/a:xfrm/a:ext', NS)- if off is None or ext is None:- raise ValueError('[TEXT_PACKAGE_BOUNDS] Text shape has no complete a:xfrm')- x = int(off.get('x', '0'))- y = int(off.get('y', '0'))- return [x, y, x + int(ext.get('cx', '0')), y + int(ext.get('cy', '0'))]---def _content_sha256(shape: ET.Element) -> str:- txbody = shape.find('./p:txBody', NS)- if txbody is None:- raise ValueError('[TEXT_PACKAGE_TXBODY] Contract output has no direct p:txBody')- paragraphs: list[list[dict[str, str]]] = []- for paragraph in txbody.findall('./a:p', NS):- tokens: list[dict[str, str]] = []- bullet = paragraph.find('./a:pPr/a:buChar', NS)- if bullet is not None:- tokens.append({- 'kind': 'bullet',- 'text': bullet.get('char', '•'),- })- for child in paragraph:- if child.tag == f'{{{A_NS}}}br':- tokens.append({'kind': 'break'})- continue- if child.tag not in {f'{{{A_NS}}}r', f'{{{A_NS}}}fld'}:- continue- text = child.find('./a:t', NS)- text_value = text.text if text is not None and text.text is not None else ''- tokens.append({'kind': 'run', 'text': text_value})- paragraphs.append(tokens)- payload = json.dumps(paragraphs, ensure_ascii=False, separators=(',', ':'))- return hashlib.sha256(payload.encode('utf-8')).hexdigest()---def _combined_content_sha256(shapes: list[ET.Element]) -> str:- hashes = [_content_sha256(shape) for shape in shapes]- if len(hashes) == 1:- return hashes[0]- payload = json.dumps(hashes, separators=(',', ':'))- return hashlib.sha256(payload.encode('utf-8')).hexdigest()---def _union_bounds(shapes: list[ET.Element]) -> list[int]:- bounds = [_shape_bounds(shape) for shape in shapes]- return [- min(value[0] for value in bounds),- min(value[1] for value in bounds),- max(value[2] for value in bounds),- max(value[3] for value in bounds),- ]---def _validate_block(- block: dict[str, Any],- shapes_by_id: dict[int, ET.Element],-) -> None:- source = block.get('source') or {}- source_id = source.get('element_id') or '<unknown>'- shape_ids = [int(value) for value in block.get('output_shape_ids') or []]- missing = [shape_id for shape_id in shape_ids if shape_id not in shapes_by_id]- if missing:- raise ValueError(- f'[TEXT_PACKAGE_SHAPE_MAP] {source_id}: missing output shape id(s) {missing}'- )- shapes = [shapes_by_id[shape_id] for shape_id in shape_ids]- effective_mode = block.get('effective_mode')- expected_shape_count = (- 1- if effective_mode == 'paragraph'- else int(block.get('visual_line_count') or 0)- )- if len(shapes) != expected_shape_count:- raise ValueError(- f'[TEXT_PACKAGE_SHAPE_COUNT] {source_id}: '- f'{len(shapes)} != {expected_shape_count}'- )- paragraph_count = sum(- len(shape.findall('./p:txBody/a:p', NS))- for shape in shapes- )- break_count = sum(- len(shape.findall('./p:txBody/a:p/a:br', NS))- for shape in shapes- )- if paragraph_count != int(block.get('paragraph_count') or 0):- raise ValueError(- f'[TEXT_PACKAGE_PARAGRAPHS] {source_id}: '- f'{paragraph_count} != {block.get("paragraph_count")}'- )- if break_count != int(block.get('break_count') or 0):- raise ValueError(- f'[TEXT_PACKAGE_BREAKS] {source_id}: '- f'{break_count} != {block.get("break_count")}'- )- expected_bounds = block.get('slide_bounds_emu')- if expected_bounds is not None and _union_bounds(shapes) != expected_bounds:- raise ValueError(- f'[TEXT_PACKAGE_BOUNDS] {source_id}: final package bounds changed'- )- actual_hash = _combined_content_sha256(shapes)- if actual_hash != block.get('content_sha256'):- raise ValueError(- f'[TEXT_PACKAGE_CONTENT] {source_id}: final text content hash changed'- )- block['readback'] = {'status': 'passed'}---def validate_pptx_text_contracts(- pptx_path: Path,- conversion_traces: list[dict[str, Any]] | None,-) -> None:- """Validate explicit text blocks in a temporary PPTX before publication."""- if not conversion_traces:- return- with zipfile.ZipFile(pptx_path) as package:- for slide_trace in conversion_traces:- blocks = slide_trace.get('text_blocks') or []- if not blocks:- continue- slide_num = int(slide_trace['slide_num'])- slide_path = f'ppt/slides/slide{slide_num}.xml'- try:- root = ET.fromstring(package.read(slide_path))- except KeyError as exc:- raise ValueError(- f'[TEXT_PACKAGE_SLIDE] Missing {slide_path}'- ) from exc- shapes_by_id = _shape_map(root)- for block in blocks:- _validate_block(block, shapes_by_id)diff --git a/skills/ppt-master/scripts/svg_to_pptx/text_contract.py b/skills/ppt-master/scripts/svg_to_pptx/text_contract.pydeleted file mode 100644index 6ef5330f..00000000--- a/skills/ppt-master/scripts/svg_to_pptx/text_contract.py+++ /dev/null@@ -1,1322 +0,0 @@-#!/usr/bin/env python3-"""-PPT Master - SVG Text Layout Contract--Parse and validate explicit SVG paragraph/line semantics before any tspan-normalization mutates the author source tree.--Usage:- Import ``validate_text_contracts`` from the SVG quality/export pipeline.--Examples:- contracts = validate_text_contracts(svg_root, merge_paragraphs=True)--Dependencies:- None (only uses the Python standard library)-"""--from __future__ import annotations--import math-import re-from dataclasses import dataclass-from xml.etree import ElementTree as ET---SVG_NS = "http://www.w3.org/2000/svg"-XML_SPACE_ATTR = "{http://www.w3.org/XML/1998/namespace}space"--TEXT_MODE_ATTR = "data-pptx-text-mode"-TEXT_BOUNDS_ATTR = "data-pptx-text-bounds"-TEXT_BREAK_ATTR = "data-pptx-break"-TEXT_JOIN_ATTR = "data-pptx-join"--RUNTIME_CONTRACT_ATTR = "data-pptx-runtime-text-contract"-RUNTIME_SOURCE_ID_ATTR = "data-pptx-runtime-source-text-id"-RUNTIME_REQUESTED_MODE_ATTR = "data-pptx-runtime-requested-mode"-RUNTIME_EFFECTIVE_MODE_ATTR = "data-pptx-runtime-effective-mode"-RUNTIME_LINE_INDEX_ATTR = "data-pptx-runtime-line-index"-RUNTIME_LINE_COUNT_ATTR = "data-pptx-runtime-line-count"-RUNTIME_MODE_REASON_ATTR = "data-pptx-runtime-mode-reason"-RUNTIME_ATTRS = frozenset(- {- RUNTIME_CONTRACT_ATTR,- RUNTIME_SOURCE_ID_ATTR,- RUNTIME_REQUESTED_MODE_ATTR,- RUNTIME_EFFECTIVE_MODE_ATTR,- RUNTIME_LINE_INDEX_ATTR,- RUNTIME_LINE_COUNT_ATTR,- RUNTIME_MODE_REASON_ATTR,- }-)--TEXT_MODES = frozenset({"auto", "paragraph", "lines"})-BREAK_KINDS = frozenset({"soft", "line", "paragraph"})-JOIN_KINDS = frozenset({"space", "none"})--DY_TOLERANCE_PX = 0.5-MAX_DY_MULTIPLIER = 3.0--_NUMBER = r"[+-]?(?:\d+(?:\.\d*)?|\.\d+)(?:[eE][+-]?\d+)?"-_COORDINATE_RE = re.compile(rf"^({_NUMBER})(?:px)?$")-_FOUR_NUMBERS_RE = re.compile(- rf"^\s*({_NUMBER})(?:\s*,\s*|\s+)"- rf"({_NUMBER})(?:\s*,\s*|\s+)"- rf"({_NUMBER})(?:\s*,\s*|\s+)"- rf"({_NUMBER})\s*$"-)-_TRANSLATE_RE = re.compile(- rf"translate\(\s*({_NUMBER})(?:(?:\s*,\s*|\s+)({_NUMBER}))?\s*\)"-)---@dataclass-class TextContractDiagnostic:- """Describe one stable, actionable text-contract violation."""-- code: str- message: str- source_name: str | None = None- element_id: str | None = None- element_tag: str = "text"- attribute: str | None = None- hint: str | None = None-- def format(self) -> str:- """Format the diagnostic for checker/exporter user-facing output."""- location = f"<{self.element_tag}>"- if self.element_id:- location = f"<{self.element_tag} id={self.element_id!r}>"- if self.source_name:- location = f"{self.source_name} {location}"- attribute = f" {self.attribute}:" if self.attribute else ":"- rendered = f"[{self.code}] {location}{attribute} {self.message}"- if self.hint:- rendered += f"; fix: {self.hint}"- return rendered---class TextContractError(ValueError):- """Aggregate explicit text-contract diagnostics without partial mutation."""-- def __init__(- self,- diagnostics: list[TextContractDiagnostic] | tuple[TextContractDiagnostic, ...],- ):- self.diagnostics = tuple(diagnostics)- super().__init__(self.format())-- def format(self) -> str:- """Format every diagnostic in deterministic discovery order."""- return "\n".join(diagnostic.format() for diagnostic in self.diagnostics)---@dataclass-class TextVisualLine:- """Normalized author-source view of one explicit paragraph visual line."""-- starter: ET.Element- inline_tspans: tuple[ET.Element, ...]- break_kind: str | None- join_kind: str | None- dy: float- space_before: float = 0.0-- @property- def elements(self) -> tuple[ET.Element, ...]:- """Return the line starter followed by direct-child inline runs."""- return (self.starter, *self.inline_tspans)---@dataclass-class TextBlockContract:- """Validated explicit contract attached to one SVG ``<text>`` block."""-- element: ET.Element- element_id: str- requested_mode: str- effective_mode: str- mode_reason: str- bounds: tuple[float, float, float, float] | None = None- anchor_x: float | None = None- anchor_y: float | None = None- block_font_size: float | None = None- base_line_height: float | None = None- top_inset: float | None = None- preserve_space: bool = False- visual_lines: tuple[TextVisualLine, ...] = ()---def _local_name(elem: ET.Element) -> str:- return elem.tag.rsplit("}", 1)[-1]---def _style_map(elem: ET.Element) -> dict[str, str]:- parsed: dict[str, str] = {}- for declaration in (elem.get("style") or "").split(";"):- if ":" not in declaration:- continue- name, value = declaration.split(":", 1)- parsed[name.strip()] = value.strip()- return parsed---def _element_chain(- elem: ET.Element,- parent_map: dict[ET.Element, ET.Element],-) -> list[ET.Element]:- chain = [elem]- while chain[-1] in parent_map:- chain.append(parent_map[chain[-1]])- chain.reverse()- return chain---def _effective_property(- elem: ET.Element,- name: str,- parent_map: dict[ET.Element, ET.Element],-) -> str | None:- value: str | None = None- for node in _element_chain(elem, parent_map):- if node.get(name) is not None:- value = node.get(name)- style_value = _style_map(node).get(name)- if style_value is not None:- value = style_value- return value---def _preserves_space(- elem: ET.Element,- parent_map: dict[ET.Element, ET.Element],-) -> bool:- preserve = False- for node in _element_chain(elem, parent_map):- value = node.get(XML_SPACE_ATTR) or node.get("xml:space")- if value == "preserve":- preserve = True- elif value == "default":- preserve = False- return preserve---def _parse_coordinate(value: str | None) -> float | None:- if value is None:- return None- match = _COORDINATE_RE.fullmatch(value.strip())- if not match:- return None- parsed = float(match.group(1))- return parsed if math.isfinite(parsed) else None---def _parse_bounds(value: str | None) -> tuple[float, float, float, float] | None:- if value is None:- return None- match = _FOUR_NUMBERS_RE.fullmatch(value)- if not match:- return None- parsed = tuple(float(part) for part in match.groups())- if not all(math.isfinite(part) for part in parsed):- return None- return parsed # type: ignore[return-value]---def _parse_length_px(value: str | None, inherited: float = 16.0) -> float | None:- if value is None:- return inherited- raw = value.strip().lower()- match = re.fullmatch(rf"({_NUMBER})(px|pt|em|rem)?", raw)- if not match:- return None- amount = float(match.group(1))- if not math.isfinite(amount):- return None- unit = match.group(2) or "px"- if unit == "pt":- return amount * 4.0 / 3.0- if unit in {"em", "rem"}:- return amount * inherited- return amount---def _effective_font_size(- elem: ET.Element,- parent_map: dict[ET.Element, ET.Element],-) -> float | None:- size = 16.0- for node in _element_chain(elem, parent_map):- raw = node.get("font-size")- style_raw = _style_map(node).get("font-size")- if style_raw is not None:- raw = style_raw- if raw is None:- continue- parsed = _parse_length_px(raw, size)- if parsed is None or parsed <= 0:- return None- size = parsed- return size---def _translate_only(value: str) -> bool:- position = 0- matched = False- while position < len(value):- separator = re.match(r"[\s,]*", value[position:])- position += separator.end() if separator else 0- if position >= len(value):- break- match = _TRANSLATE_RE.match(value, position)- if not match:- return False- matched = True- position = match.end()- return matched---def _has_visible_direct_text(value: str | None, preserve_space: bool) -> bool:- if not value:- return False- return bool(value) if preserve_space else bool(value.strip())---def _nearest_text(- elem: ET.Element,- parent_map: dict[ET.Element, ET.Element],-) -> ET.Element | None:- current = parent_map.get(elem)- while current is not None:- if _local_name(current) == "text":- return current- current = parent_map.get(current)- return None---def _direct_metadata(elem: ET.Element, part: str) -> ET.Element | None:- for child in elem:- if (- _local_name(child) == "metadata"- and child.get("data-pptx-part") == part- ):- return child- return None---def _native_text_carrier(- text: ET.Element,- parent_map: dict[ET.Element, ET.Element],-) -> ET.Element | None:- current = parent_map.get(text)- while current is not None:- if _direct_metadata(current, "txbody") is not None:- return current- current = parent_map.get(current)- return None---def _validate_native_carrier_contract(- text: ET.Element,- mode: str,- merge_paragraphs: bool,- parent_map: dict[ET.Element, ET.Element],- diagnostics: list[TextContractDiagnostic],-) -> None:- """Fail closed when an imported one-shape text carrier cannot be rebuilt."""- carrier = _native_text_carrier(text, parent_map)- if carrier is None:- return-- if mode in {"auto", "lines"} or not merge_paragraphs:- _add_diagnostic(- diagnostics,- "TEXT_NATIVE_MODE_UNSUPPORTED",- text,- "native text carriers only support effective paragraph mode",- attribute=TEXT_MODE_ATTR,- hint="use paragraph mode without --no-merge, or remove the native carrier metadata",- )- return-- visible_texts = [- descendant- for descendant in carrier.iter()- if _local_name(descendant) == "text"- and "".join(descendant.itertext()).strip()- ]- if len(visible_texts) != 1:- _add_diagnostic(- diagnostics,- "TEXT_NATIVE_TEXT_COUNT",- text,- "native text carriers require exactly one non-empty text block",- hint="keep one non-empty explicit paragraph inside the imported carrier",- )-- carrier_metadata = _direct_metadata(carrier, "text-carrier")- if (- carrier_metadata is None- or carrier_metadata.get("data-pptx-text-carrier-version") != "1"- ):- _add_diagnostic(- diagnostics,- "TEXT_NATIVE_CARRIER_METADATA_REQUIRED",- text,- "native paragraph recompilation requires version-1 text-carrier metadata",- hint="re-import the source PPTX with the current pptx_to_svg converter",- )- return- if carrier_metadata.get("data-pptx-text-carrier-eligible") != "1":- reason = carrier_metadata.get("data-pptx-text-carrier-reason") or "unknown"- _add_diagnostic(- diagnostics,- "TEXT_NATIVE_CARRIER_INELIGIBLE",- text,- f"the imported native text carrier is ineligible: {reason}",- hint="use ordinary SVG text or remove the unsupported carrier feature",- )---def _line_has_positional_descendant(starter: ET.Element) -> bool:- for descendant in starter.iter():- if descendant is starter or _local_name(descendant) != "tspan":- continue- if any(descendant.get(name) is not None for name in ("x", "y", "dy")):- return True- return False---def _add_diagnostic(- diagnostics: list[TextContractDiagnostic],- code: str,- elem: ET.Element,- message: str,- *,- attribute: str | None = None,- hint: str | None = None,- owner: ET.Element | None = None,-) -> None:- identity = owner or elem- diagnostics.append(- TextContractDiagnostic(- code=code,- message=message,- element_id=identity.get("id"),- element_tag=_local_name(elem),- attribute=attribute,- hint=hint,- )- )---def _validate_transforms_and_baseline(- text: ET.Element,- parent_map: dict[ET.Element, ET.Element],- diagnostics: list[TextContractDiagnostic],-) -> None:- for node in _element_chain(text, parent_map):- style_transform = _style_map(node).get("transform")- if style_transform is not None:- _add_diagnostic(- diagnostics,- "TEXT_TRANSFORM_UNSUPPORTED",- node,- "CSS style transforms are unsupported for explicit paragraph bounds",- attribute="style/transform",- hint="move a pure translate into the SVG transform attribute",- owner=text,- )- continue- transform = node.get("transform")- if transform and not _translate_only(transform):- _add_diagnostic(- diagnostics,- "TEXT_TRANSFORM_UNSUPPORTED",- node,- "explicit paragraph bounds only support pure translate transforms",- attribute="transform",- hint="materialize the transform into coordinates or use mode='lines'",- owner=text,- )-- effective_dominant = _effective_property(text, "dominant-baseline", parent_map)- effective_alignment = _effective_property(text, "alignment-baseline", parent_map)- effective_shift = _effective_property(text, "baseline-shift", parent_map)- if effective_dominant not in {None, "auto", "alphabetic"}:- _add_diagnostic(- diagnostics,- "TEXT_BASELINE_UNSUPPORTED",- text,- f"unsupported inherited dominant-baseline {effective_dominant!r}",- attribute="dominant-baseline",- hint="use the default alphabetic baseline",- )- if effective_alignment not in {None, "auto", "baseline", "alphabetic"}:- _add_diagnostic(- diagnostics,- "TEXT_BASELINE_UNSUPPORTED",- text,- f"unsupported inherited alignment-baseline {effective_alignment!r}",- attribute="alignment-baseline",- hint="use the default alphabetic baseline",- )- if effective_shift not in {- None,- "baseline",- "0",- "0px",- "+0",- "+0px",- "-0",- "-0px",- }:- _add_diagnostic(- diagnostics,- "TEXT_BASELINE_UNSUPPORTED",- text,- f"unsupported inherited baseline-shift {effective_shift!r}",- attribute="baseline-shift",- hint="remove baseline-shift from explicit paragraph text",- )-- for node in text.iter():- if node is text:- continue- if node is not text:- transform = _style_map(node).get("transform") or node.get("transform")- if transform:- _add_diagnostic(- diagnostics,- "TEXT_TRANSFORM_UNSUPPORTED",- node,- "tspan transforms are unsupported in explicit paragraph mode",- attribute="transform",- hint="remove the tspan transform and express line positions with x/dy",- owner=text,- )-- styles = _style_map(node)- dominant = styles.get("dominant-baseline") or node.get("dominant-baseline")- alignment = styles.get("alignment-baseline") or node.get("alignment-baseline")- shift = styles.get("baseline-shift") or node.get("baseline-shift")- if dominant not in {None, "auto", "alphabetic"}:- _add_diagnostic(- diagnostics,- "TEXT_BASELINE_UNSUPPORTED",- node,- f"unsupported dominant-baseline {dominant!r}",- attribute="dominant-baseline",- hint="use the default alphabetic baseline",- owner=text,- )- if alignment not in {None, "auto", "baseline", "alphabetic"}:- _add_diagnostic(- diagnostics,- "TEXT_BASELINE_UNSUPPORTED",- node,- f"unsupported alignment-baseline {alignment!r}",- attribute="alignment-baseline",- hint="use the default alphabetic baseline",- owner=text,- )- if shift not in {None, "baseline", "0", "0px", "+0", "+0px", "-0", "-0px"}:- _add_diagnostic(- diagnostics,- "TEXT_BASELINE_UNSUPPORTED",- node,- f"unsupported baseline-shift {shift!r}",- attribute="baseline-shift",- hint="remove baseline-shift from explicit paragraph text",- owner=text,- )---def _validate_unsupported_text_features(- text: ET.Element,- parent_map: dict[ET.Element, ET.Element],- diagnostics: list[TextContractDiagnostic],-) -> None:- """Reject SVG text features that have no deterministic DrawingML mapping."""- for attribute in ("dx", "dy", "rotate"):- if text.get(attribute) is not None:- _add_diagnostic(- diagnostics,- "TEXT_BLOCK_OFFSET_UNSUPPORTED",- text,- f"block-level {attribute} is unsupported in explicit paragraph mode",- attribute=attribute,- hint="fold the offset into x/y or use a pure translate",- )-- for node in text.iter():- for attribute in ("textLength", "lengthAdjust"):- if node.get(attribute) is not None:- _add_diagnostic(- diagnostics,- "TEXT_LENGTH_ADJUST_UNSUPPORTED",- node,- "SVG text-length adjustment has no native paragraph mapping",- attribute=attribute,- hint="remove textLength/lengthAdjust and use normal font metrics",- owner=text,- )- if node is not text and node.get("rotate") is not None:- _add_diagnostic(- diagnostics,- "TEXT_BLOCK_OFFSET_UNSUPPORTED",- node,- "per-character tspan rotation is unsupported in paragraph mode",- attribute="rotate",- hint="remove the rotate attribute or use mode='lines'",- owner=text,- )- if node is text:- continue- styles = _style_map(node)- if node.get("dx") is not None:- _add_diagnostic(- diagnostics,- "TEXT_INLINE_OFFSET_UNSUPPORTED",- node,- "tspan dx offsets have no deterministic paragraph-run mapping",- attribute="dx",- hint="fold the offset into line x or remove it",- owner=text,- )- local_writing = styles.get("writing-mode") or node.get("writing-mode")- local_direction = styles.get("direction") or node.get("direction")- local_bidi = styles.get("unicode-bidi") or node.get("unicode-bidi")- if local_writing not in {None, "horizontal-tb"}:- _add_diagnostic(- diagnostics,- "TEXT_WRITING_MODE_UNSUPPORTED",- node,- f"unsupported tspan writing-mode {local_writing!r}",- attribute="writing-mode",- hint="remove the run-level writing mode",- owner=text,- )- if local_direction not in {None, "ltr"} or local_bidi not in {None, "normal"}:- _add_diagnostic(- diagnostics,- "TEXT_DIRECTION_UNSUPPORTED",- node,- "run-level bidirectional controls are unsupported in v1",- attribute="direction/unicode-bidi",- hint="remove the run-level bidi controls",- owner=text,- )-- writing_mode = _effective_property(text, "writing-mode", parent_map)- if writing_mode not in {None, "horizontal-tb"}:- _add_diagnostic(- diagnostics,- "TEXT_WRITING_MODE_UNSUPPORTED",- text,- f"unsupported writing-mode {writing_mode!r}",- attribute="writing-mode",- hint="use horizontal-tb text or mode='lines'",- )- direction = _effective_property(text, "direction", parent_map)- unicode_bidi = _effective_property(text, "unicode-bidi", parent_map)- if direction not in {None, "ltr"} or unicode_bidi not in {None, "normal"}:- _add_diagnostic(- diagnostics,- "TEXT_DIRECTION_UNSUPPORTED",- text,- "bidirectional paragraph controls are unsupported in v1",- attribute="direction/unicode-bidi",- hint="use default left-to-right text or mode='lines'",- )---def _parse_paragraph_contract(- text: ET.Element,- element_id: str,- parent_map: dict[ET.Element, ET.Element],- merge_paragraphs: bool,- diagnostics: list[TextContractDiagnostic],-) -> TextBlockContract | None:- bounds_raw = text.get(TEXT_BOUNDS_ATTR)- bounds = _parse_bounds(bounds_raw)- if bounds_raw is None:- _add_diagnostic(- diagnostics,- "TEXT_BOUNDS_REQUIRED",- text,- "paragraph mode requires four explicit text-frame bounds",- attribute=TEXT_BOUNDS_ATTR,- hint="add data-pptx-text-bounds='x y width height'",- )- elif bounds is None:- _add_diagnostic(- diagnostics,- "TEXT_BOUNDS_INVALID",- text,- "bounds must contain exactly four finite unitless numbers",- attribute=TEXT_BOUNDS_ATTR,- hint="use data-pptx-text-bounds='x y width height'",- )- elif bounds[2] <= 0 or bounds[3] <= 0:- _add_diagnostic(- diagnostics,- "TEXT_BOUNDS_NONPOSITIVE",- text,- "text-frame width and height must both be positive",- attribute=TEXT_BOUNDS_ATTR,- hint="provide positive width and height values",- )-- _validate_transforms_and_baseline(text, parent_map, diagnostics)- _validate_unsupported_text_features(text, parent_map, diagnostics)-- effective_font_size_raw = _effective_property(text, "font-size", parent_map)- if (- effective_font_size_raw is not None- and re.fullmatch(rf"{_NUMBER}(?:em|rem)", effective_font_size_raw.strip().lower())- ):- _add_diagnostic(- diagnostics,- "TEXT_FONT_SIZE_RELATIVE",- text,- "relative block font sizes are unsupported in explicit paragraph mode",- attribute="font-size",- hint="resolve the block font size to unitless px or pt",- )-- font_size = _effective_font_size(text, parent_map)- if font_size is None:- _add_diagnostic(- diagnostics,- "TEXT_FONT_SIZE_INVALID",- text,- "the inherited block font-size must resolve to a positive px value",- attribute="font-size",- hint="use a positive unitless, px, or pt font-size",- )-- preserve_space = _preserves_space(text, parent_map)- if _has_visible_direct_text(text.text, preserve_space):- _add_diagnostic(- diagnostics,- "TEXT_PARAGRAPH_DIRECT_TEXT",- text,- "paragraph content must start inside a direct child tspan",- hint="move the direct text into the first line tspan",- )-- direct_children = list(text)- for child in direct_children:- if _local_name(child) != "tspan":- _add_diagnostic(- diagnostics,- "TEXT_PARAGRAPH_CHILD_INVALID",- child,- "paragraph mode only accepts direct child tspan elements",- hint="remove the non-tspan child or move it outside the text block",- owner=text,- )- if _has_visible_direct_text(child.tail, preserve_space):- _add_diagnostic(- diagnostics,- "TEXT_PARAGRAPH_DIRECT_TEXT",- child,- "direct tspan tail text is ambiguous paragraph content",- hint="move the tail into the preceding or following line tspan",- owner=text,- )-- tspan_children = [child for child in direct_children if _local_name(child) == "tspan"]- line_groups: list[list[ET.Element]] = []- for tspan in tspan_children:- positional = any(tspan.get(name) is not None for name in ("x", "y", "dy"))- if positional:- line_groups.append([tspan])- elif not line_groups:- _add_diagnostic(- diagnostics,- "TEXT_LINE_START_REQUIRED",- tspan,- "an inline tspan appears before the first positioned visual line",- hint="give the first line tspan the required x/y or x/dy anchor",- owner=text,- )- else:- line_groups[-1].append(tspan)-- if not line_groups:- _add_diagnostic(- diagnostics,- "TEXT_LINE_START_REQUIRED",- text,- "paragraph mode requires at least one positioned direct child tspan",- hint="wrap the first visual line in a positioned tspan",- )- return None-- for group in line_groups:- for line_part in group:- if _line_has_positional_descendant(line_part):- _add_diagnostic(- diagnostics,- "TEXT_POSITIONAL_NESTED",- line_part,- "nested positional tspans are unsupported in paragraph mode",- hint="keep x/y/dy only on direct visual-line starters",- owner=text,- )- content = "".join("".join(part.itertext()) for part in group)- if not content.strip():- _add_diagnostic(- diagnostics,- "TEXT_LINE_EMPTY",- group[0],- "empty or whitespace-only visual lines are unsupported in v1",- hint="remove the empty line and express spacing on the next paragraph dy",- owner=text,- )-- parent_x_raw = text.get("x")- parent_y_raw = text.get("y")- first = line_groups[0][0]- anchor_x: float | None = None- anchor_y: float | None = None- parent_anchor = parent_x_raw is not None or parent_y_raw is not None- if parent_anchor and (parent_x_raw is None or parent_y_raw is None):- _add_diagnostic(- diagnostics,- "TEXT_BLOCK_ANCHOR_INVALID",- text,- "parent text must provide both x and y, or neither",- attribute="x/y",- hint="use <text x y> with first <tspan x dy='0'>, or put x/y on the first tspan",- )- elif parent_anchor:- anchor_x = _parse_coordinate(parent_x_raw)- anchor_y = _parse_coordinate(parent_y_raw)- first_x = _parse_coordinate(first.get("x"))- first_dy = _parse_coordinate(first.get("dy"))- if anchor_x is None or anchor_y is None or first_x is None:- _add_diagnostic(- diagnostics,- "TEXT_BLOCK_ANCHOR_INVALID",- first,- "parent and first-line anchors must be finite unitless or px coordinates",- attribute="x/y/dy",- hint="use numeric parent x/y and repeat parent x on the first tspan",- owner=text,- )- if first.get("y") is not None or first.get("dy") is None or first_dy != 0:- _add_diagnostic(- diagnostics,- "TEXT_FIRST_LINE_GEOMETRY",- first,- "with parent x/y, the first line must use x and dy='0' without y",- attribute="x/y/dy",- hint="repeat the parent x and set dy='0' on the first tspan",- owner=text,- )- if anchor_x is not None and first_x is not None and abs(anchor_x - first_x) > 1e-6:- _add_diagnostic(- diagnostics,- "TEXT_LINE_X_MISMATCH",- first,- "first-line x does not repeat the block anchor",- attribute="x",- hint="set the first-line x equal to the parent text x",- owner=text,- )- else:- anchor_x = _parse_coordinate(first.get("x"))- anchor_y = _parse_coordinate(first.get("y"))- if anchor_x is None or anchor_y is None or first.get("dy") is not None:- _add_diagnostic(- diagnostics,- "TEXT_FIRST_LINE_GEOMETRY",- first,- "without parent x/y, the first line must provide finite x/y and no dy",- attribute="x/y/dy",- hint="put x and y on the first tspan, or use the parent-anchor form",- owner=text,- )-- if first.get(TEXT_BREAK_ATTR) is not None:- _add_diagnostic(- diagnostics,- "TEXT_BREAK_FIRST",- first,- "the first visual line cannot declare a preceding break",- attribute=TEXT_BREAK_ATTR,- hint="remove data-pptx-break from the first line",- owner=text,- )- if first.get(TEXT_JOIN_ATTR) is not None:- _add_diagnostic(- diagnostics,- "TEXT_JOIN_FORBIDDEN",- first,- "the first visual line cannot declare a join",- attribute=TEXT_JOIN_ATTR,- hint="remove data-pptx-join from the first line",- owner=text,- )-- raw_lines: list[tuple[ET.Element, tuple[ET.Element, ...], str | None, str | None, float]] = [- (first, tuple(line_groups[0][1:]), None, None, 0.0)- ]- for group in line_groups[1:]:- starter = group[0]- line_x = _parse_coordinate(starter.get("x"))- line_dy = _parse_coordinate(starter.get("dy"))- if starter.get("y") is not None or starter.get("x") is None or line_x is None:- _add_diagnostic(- diagnostics,- "TEXT_LINE_GEOMETRY_INVALID",- starter,- "later visual lines require a finite x and must not set y",- attribute="x/y",- hint="repeat the block anchor x and use a positive relative dy",- owner=text,- )- if anchor_x is not None and line_x is not None and abs(anchor_x - line_x) > 1e-6:- _add_diagnostic(- diagnostics,- "TEXT_LINE_X_MISMATCH",- starter,- "visual-line x does not repeat the block anchor",- attribute="x",- hint="set every line starter x equal to the block anchor",- owner=text,- )- if line_dy is None or line_dy <= 0:- _add_diagnostic(- diagnostics,- "TEXT_LINE_DY_INVALID",- starter,- "later visual lines require a positive relative dy",- attribute="dy",- hint="set dy to the positive distance from the preceding baseline",- owner=text,- )- line_dy = 0.0- break_kind = starter.get(TEXT_BREAK_ATTR)- join_kind = starter.get(TEXT_JOIN_ATTR)- if break_kind is None:- _add_diagnostic(- diagnostics,- "TEXT_BREAK_REQUIRED",- starter,- "every visual line after the first must declare its break semantics",- attribute=TEXT_BREAK_ATTR,- hint="set data-pptx-break to soft, line, or paragraph",- owner=text,- )- elif break_kind not in BREAK_KINDS:- _add_diagnostic(- diagnostics,- "TEXT_BREAK_INVALID",- starter,- f"unknown break kind {break_kind!r}",- attribute=TEXT_BREAK_ATTR,- hint="use exactly soft, line, or paragraph",- owner=text,- )- if break_kind == "soft":- if join_kind is None:- _add_diagnostic(- diagnostics,- "TEXT_JOIN_REQUIRED",- starter,- "soft visual breaks require an explicit join policy",- attribute=TEXT_JOIN_ATTR,- hint="set data-pptx-join to space or none",- owner=text,- )- elif join_kind not in JOIN_KINDS:- _add_diagnostic(- diagnostics,- "TEXT_JOIN_INVALID",- starter,- f"unknown join kind {join_kind!r}",- attribute=TEXT_JOIN_ATTR,- hint="use exactly space or none",- owner=text,- )- elif join_kind is not None:- _add_diagnostic(- diagnostics,- "TEXT_JOIN_FORBIDDEN",- starter,- "join is only valid with data-pptx-break='soft'",- attribute=TEXT_JOIN_ATTR,- hint="remove data-pptx-join or change the break to soft",- owner=text,- )- raw_lines.append((starter, tuple(group[1:]), break_kind, join_kind, line_dy))-- for group in line_groups:- for inline in group[1:]:- for attribute in (TEXT_BREAK_ATTR, TEXT_JOIN_ATTR):- if inline.get(attribute) is not None:- _add_diagnostic(- diagnostics,- (- "TEXT_BREAK_PLACEMENT"- if attribute == TEXT_BREAK_ATTR- else "TEXT_JOIN_PLACEMENT"- ),- inline,- "break/join metadata is only valid on a direct visual-line starter",- attribute=attribute,- hint="move the attribute to the positioned line-start tspan",- owner=text,- )-- base_line_height: float | None = None- if len(raw_lines) > 1:- positive_dys = [line[4] for line in raw_lines[1:] if line[4] > 0]- if positive_dys:- base_line_height = min(positive_dys)- if font_size is not None and (- base_line_height > font_size * MAX_DY_MULTIPLIER + DY_TOLERANCE_PX- ):- _add_diagnostic(- diagnostics,- "TEXT_LINE_HEIGHT_INVALID",- text,- "base line height exceeds three times the block font size",- attribute="dy",- hint="reduce line dy or split the content into separate text blocks",- )- for starter, _, break_kind, _, line_dy in raw_lines[1:]:- if line_dy > base_line_height * MAX_DY_MULTIPLIER + DY_TOLERANCE_PX:- _add_diagnostic(- diagnostics,- "TEXT_LINE_HEIGHT_INVALID",- starter,- "line dy exceeds three times the normalized base line height",- attribute="dy",- hint="reduce the gap or use a separate text block",- owner=text,- )- if break_kind in {"soft", "line"} and (- abs(line_dy - base_line_height) > DY_TOLERANCE_PX- ):- _add_diagnostic(- diagnostics,- "TEXT_BREAK_SPACING_INVALID",- starter,- "soft and line breaks cannot carry paragraph spacing",- attribute="dy",- hint="use the base line height, or declare a paragraph break",- owner=text,- )-- visual_lines: list[TextVisualLine] = []- for starter, inline, break_kind, join_kind, line_dy in raw_lines:- space_before = 0.0- if break_kind == "paragraph" and base_line_height is not None:- space_before = max(0.0, line_dy - base_line_height)- visual_lines.append(- TextVisualLine(- starter=starter,- inline_tspans=inline,- break_kind=break_kind,- join_kind=join_kind,- dy=line_dy,- space_before=space_before,- )- )-- top_inset: float | None = None- if bounds is not None and bounds[2] > 0 and bounds[3] > 0:- text_anchor = _effective_property(text, "text-anchor", parent_map) or "start"- if text_anchor not in {"start", "middle", "end"}:- _add_diagnostic(- diagnostics,- "TEXT_ANCHOR_INVALID",- text,- f"unsupported text-anchor {text_anchor!r}",- attribute="text-anchor",- hint="use start, middle, or end",- )- if anchor_x is not None:- expected_x = {- "start": bounds[0],- "middle": bounds[0] + bounds[2] / 2.0,- "end": bounds[0] + bounds[2],- }.get(text_anchor)- if expected_x is not None and abs(anchor_x - expected_x) > DY_TOLERANCE_PX:- _add_diagnostic(- diagnostics,- "TEXT_BOUNDS_ANCHOR_MISMATCH",- text,- "block anchor x does not match the declared text frame and text-anchor",- attribute=TEXT_BOUNDS_ATTR,- hint="align line x with the frame left, center, or right anchor",- )- if anchor_y is not None:- final_baseline = anchor_y + sum(line[4] for line in raw_lines[1:])- bounds_bottom = bounds[1] + bounds[3]- if (- anchor_y < bounds[1] - DY_TOLERANCE_PX- or anchor_y > bounds_bottom + DY_TOLERANCE_PX- or final_baseline < bounds[1] - DY_TOLERANCE_PX- or final_baseline > bounds_bottom + DY_TOLERANCE_PX- ):- _add_diagnostic(- diagnostics,- "TEXT_BOUNDS_BASELINE_OUTSIDE",- text,- "first or final visual baseline falls outside the declared text frame",- attribute=TEXT_BOUNDS_ATTR,- hint="increase frame height or correct the line baselines",- )- if font_size is not None:- top_inset = anchor_y - bounds[1] - 0.85 * font_size- if top_inset < -1e-9:- _add_diagnostic(- diagnostics,- "TEXT_TOP_INSET_NEGATIVE",- text,- "the baseline-to-frame offset produces a negative top inset",- attribute=TEXT_BOUNDS_ATTR,- hint="move bounds.y upward or move the first baseline downward",- )- else:- top_inset = max(0.0, top_inset)-- effective_mode = "paragraph" if merge_paragraphs else "lines"- mode_reason = "explicit-paragraph" if merge_paragraphs else "cli-no-merge"- return TextBlockContract(- element=text,- element_id=element_id,- requested_mode="paragraph",- effective_mode=effective_mode,- mode_reason=mode_reason,- bounds=bounds,- anchor_x=anchor_x,- anchor_y=anchor_y,- block_font_size=font_size,- base_line_height=base_line_height,- top_inset=top_inset,- preserve_space=preserve_space,- visual_lines=tuple(visual_lines),- )---def validate_text_contracts(- root: ET.Element,- merge_paragraphs: bool = True,- *,- enforce_native_carrier: bool = True,- source_name: str | None = None,-) -> dict[ET.Element, TextBlockContract]:- """Validate every explicit text contract and return contracts by element.-- Validation is all-or-nothing. Every discoverable violation is collected and- raised as one ``TextContractError`` before callers mutate the SVG tree.- Unmarked legacy text is intentionally omitted from the result. Disable- ``enforce_native_carrier`` only for strict visual-preview generation;- formal PPTX export must keep it enabled.- """- diagnostics: list[TextContractDiagnostic] = []- parent_map = {child: parent for parent in root.iter() for child in parent}- id_counts: dict[str, int] = {}- for elem in root.iter():- elem_id = elem.get("id")- if elem_id:- id_counts[elem_id] = id_counts.get(elem_id, 0) + 1-- for elem in root.iter():- tag = _local_name(elem)- for attribute in elem.attrib:- if attribute.startswith("data-paragraph-"):- _add_diagnostic(- diagnostics,- "TEXT_INTERNAL_ATTR_FORBIDDEN",- elem,- "data-paragraph-* attributes are compiler-internal and cannot be authored",- attribute=attribute,- hint="remove the internal attribute and use the public text contract",- owner=_nearest_text(elem, parent_map),- )- elif attribute in RUNTIME_ATTRS:- _add_diagnostic(- diagnostics,- "TEXT_RUNTIME_ATTR_FORBIDDEN",- elem,- "data-pptx-runtime-* attributes are compiler-internal and cannot be authored",- attribute=attribute,- hint="remove the runtime attribute from svg_output",- owner=_nearest_text(elem, parent_map),- )- if elem.get(TEXT_MODE_ATTR) is not None and tag != "text":- _add_diagnostic(- diagnostics,- "TEXT_MODE_PLACEMENT",- elem,- "text mode is only valid on an SVG text element",- attribute=TEXT_MODE_ATTR,- hint="move data-pptx-text-mode to the owning text element",- owner=_nearest_text(elem, parent_map),- )- if elem.get(TEXT_BOUNDS_ATTR) is not None and tag != "text":- _add_diagnostic(- diagnostics,- "TEXT_BOUNDS_PLACEMENT",- elem,- "text bounds are only valid on an SVG text element",- attribute=TEXT_BOUNDS_ATTR,- hint="move data-pptx-text-bounds to the owning text element",- owner=_nearest_text(elem, parent_map),- )- for attribute, code in (- (TEXT_BREAK_ATTR, "TEXT_BREAK_PLACEMENT"),- (TEXT_JOIN_ATTR, "TEXT_JOIN_PLACEMENT"),- ):- if elem.get(attribute) is None:- continue- owner = _nearest_text(elem, parent_map)- if tag != "tspan" or owner is None or parent_map.get(elem) is not owner:- _add_diagnostic(- diagnostics,- code,- elem,- "break/join metadata is only valid on a direct visual-line tspan",- attribute=attribute,- hint="move the attribute to a direct line-start tspan in paragraph mode",- owner=owner,- )- elif owner.get(TEXT_MODE_ATTR) != "paragraph":- _add_diagnostic(- diagnostics,- code,- elem,- "break/join metadata requires data-pptx-text-mode='paragraph'",- attribute=attribute,- hint="remove the attribute or opt the owning text into paragraph mode",- owner=owner,- )-- contracts: dict[ET.Element, TextBlockContract] = {}- for text in root.iter(f"{{{SVG_NS}}}text"):- mode = text.get(TEXT_MODE_ATTR)- if mode is None:- if text.get(TEXT_BOUNDS_ATTR) is not None:- _add_diagnostic(- diagnostics,- "TEXT_BOUNDS_MODE",- text,- "text bounds do not implicitly enable paragraph mode",- attribute=TEXT_BOUNDS_ATTR,- hint="add data-pptx-text-mode='paragraph' or remove bounds",- )- continue- if mode not in TEXT_MODES:- _add_diagnostic(- diagnostics,- "TEXT_MODE_INVALID",- text,- f"unknown text mode {mode!r}",- attribute=TEXT_MODE_ATTR,- hint="use exactly auto, paragraph, or lines without surrounding whitespace",- )- continue-- element_id = text.get("id") or ""- if not element_id:- _add_diagnostic(- diagnostics,- "TEXT_ID_REQUIRED",- text,- "an explicit text contract requires a non-empty stable id",- attribute="id",- hint="add a unique id to the text element",- )- elif id_counts.get(element_id, 0) != 1:- _add_diagnostic(- diagnostics,- "TEXT_ID_DUPLICATE",- text,- f"explicit text id {element_id!r} is not unique in this SVG",- attribute="id",- hint="rename the text element so its id is unique",- )-- if mode != "paragraph" and text.get(TEXT_BOUNDS_ATTR) is not None:- _add_diagnostic(- diagnostics,- "TEXT_BOUNDS_FORBIDDEN",- text,- f"text bounds are not valid in {mode!r} mode",- attribute=TEXT_BOUNDS_ATTR,- hint="remove bounds or use paragraph mode",- )-- if enforce_native_carrier:- _validate_native_carrier_contract(- text,- mode,- merge_paragraphs,- parent_map,- diagnostics,- )-- if mode == "paragraph":- contract = _parse_paragraph_contract(- text,- element_id,- parent_map,- merge_paragraphs,- diagnostics,- )- if contract is not None:- contracts[text] = contract- else:- effective_mode = "lines" if not merge_paragraphs or mode == "lines" else "auto"- if not merge_paragraphs:- reason = "cli-no-merge"- elif mode == "lines":- reason = "explicit-lines"- else:- reason = "legacy-auto"- contracts[text] = TextBlockContract(- element=text,- element_id=element_id,- requested_mode=mode,- effective_mode=effective_mode,- mode_reason=reason,- preserve_space=_preserves_space(text, parent_map),- )-- if diagnostics:- if source_name:- for diagnostic in diagnostics:- diagnostic.source_name = source_name- raise TextContractError(diagnostics)- return contracts---__all__ = [- "BREAK_KINDS",- "JOIN_KINDS",- "RUNTIME_CONTRACT_ATTR",- "RUNTIME_ATTRS",- "RUNTIME_EFFECTIVE_MODE_ATTR",- "RUNTIME_LINE_COUNT_ATTR",- "RUNTIME_LINE_INDEX_ATTR",- "RUNTIME_MODE_REASON_ATTR",- "RUNTIME_REQUESTED_MODE_ATTR",- "RUNTIME_SOURCE_ID_ATTR",- "TEXT_BOUNDS_ATTR",- "TEXT_BREAK_ATTR",- "TEXT_JOIN_ATTR",- "TEXT_MODE_ATTR",- "TextBlockContract",- "TextContractDiagnostic",- "TextContractError",- "TextVisualLine",- "validate_text_contracts",-]diff --git a/skills/ppt-master/scripts/svg_to_pptx/tspan_flattener.py b/skills/ppt-master/scripts/svg_to_pptx/tspan_flattener.pyindex 641cdff0..45b6d1f5 100644--- a/skills/ppt-master/scripts/svg_to_pptx/tspan_flattener.py+++ b/skills/ppt-master/scripts/svg_to_pptx/tspan_flattener.py@@ -31,15 +31,13 @@ from xml.etree import ElementTree as ETdef flatten_positional_tspans(tree: ET.ElementTree,merge_paragraphs: bool = False,- *,- source_name: str | None = None,) -> bool:"""Flatten positional ``<tspan>`` elements into independent ``<text>``.Delegates to ``svg_finalize.flatten_tspan.flatten_text_with_tspans`` sothe in-memory transform exactly matches the on-disk one. When- ``merge_paragraphs`` is True, eligible blocks remain one <text> for- downstream hard-line and paragraph conversion.+ ``merge_paragraphs`` is True, mergeable paragraph blocks are preserved+ as a single <text> for downstream multi-<a:p> conversion.Returns True if any tspan was rewritten."""@@ -47,8 +45,4 @@ def flatten_positional_tspans(if str(scripts_dir) not in sys.path:sys.path.insert(0, str(scripts_dir))from svg_finalize.flatten_tspan import flatten_text_with_tspans # type: ignore- return flatten_text_with_tspans(- tree,- merge_paragraphs=merge_paragraphs,- source_name=source_name,- )+ return flatten_text_with_tspans(tree, merge_paragraphs=merge_paragraphs)