| 1 | """Import PPT Master-owned Office Math into canonical SVG formula markers. |
| 2 | |
| 3 | The reverse contract is deliberately narrow: accept only the closed OMML |
| 4 | vocabulary already validated by the native formula compiler, serialize that |
| 5 | structure to compiler-accepted LaTeX, and build a plain-text SVG preview. |
| 6 | Unknown third-party OMML remains outside this module's reconstruction claim. |
| 7 | """ |
| 8 | |
| 9 | from __future__ import annotations |
| 10 | |
| 11 | from collections import Counter |
| 12 | from dataclasses import dataclass |
| 13 | from xml.etree import ElementTree as ET |
| 14 | |
| 15 | from svg_to_pptx.native_objects.formula_compiler import ( |
| 16 | FormulaCompileError, |
| 17 | compile_latex_to_inline_omml, |
| 18 | compile_latex_to_omml, |
| 19 | validate_omml_fragment, |
| 20 | ) |
| 21 | from svg_to_pptx.native_objects.formula_profile import ( |
| 22 | ACCENT_COMMANDS, |
| 23 | NARY_COMMANDS, |
| 24 | ) |
| 25 | |
| 26 | |
| 27 | MATH_NS = "http://schemas.openxmlformats.org/officeDocument/2006/math" |
| 28 | DRAWING_NS = "http://schemas.openxmlformats.org/drawingml/2006/main" |
| 29 | A14_NS = "http://schemas.microsoft.com/office/drawing/2010/main" |
| 30 | |
| 31 | _M = f"{{{MATH_NS}}}" |
| 32 | _A = f"{{{DRAWING_NS}}}" |
| 33 | _A14_M = f"{{{A14_NS}}}m" |
| 34 | _PROPERTY_SUFFIX = "Pr" |
| 35 | |
| 36 | _NARY_COMMAND_BY_SYMBOL = { |
| 37 | symbol: command |
| 38 | for command, (symbol, _category) in NARY_COMMANDS.items() |
| 39 | } |
| 40 | _ACCENT_COMMAND_BY_CHARACTER = { |
| 41 | character: command |
| 42 | for command, character in ACCENT_COMMANDS.items() |
| 43 | } |
| 44 | _GROUP_COMMANDS = { |
| 45 | ("⏞", "top", "bot"): "overbrace", |
| 46 | ("⏟", "bot", "top"): "underbrace", |
| 47 | ("→", "bot", "top"): "underrightarrow", |
| 48 | ("←", "bot", "top"): "underleftarrow", |
| 49 | ("↔", "bot", "top"): "underleftrightarrow", |
| 50 | ("→", "top", "bot"): "overrightarrow", |
| 51 | ("←", "top", "bot"): "overleftarrow", |
| 52 | ("↔", "top", "bot"): "overleftrightarrow", |
| 53 | } |
| 54 | _DELIMITER_COMMANDS = { |
| 55 | "{": r"\{", |
| 56 | "}": r"\}", |
| 57 | "⟨": r"\langle", |
| 58 | "⟩": r"\rangle", |
| 59 | "⌊": r"\lfloor", |
| 60 | "⌋": r"\rfloor", |
| 61 | "⌈": r"\lceil", |
| 62 | "⌉": r"\rceil", |
| 63 | "‖": r"\Vert", |
| 64 | "⟦": "⟦", |
| 65 | "⟧": "⟧", |
| 66 | } |
| 67 | _SPACING_LATEX = { |
| 68 | "\u200b": r"\!", |
| 69 | "\u2009": r"\,", |
| 70 | "\u205f": r"\:", |
| 71 | "\u2004": r"\;", |
| 72 | "\u2002": r"\enspace{}", |
| 73 | "\u2003": r"\quad{}", |
| 74 | "\u00a0": "~", |
| 75 | } |
| 76 | |
| 77 | |
| 78 | class FormulaImportError(ValueError): |
| 79 | """Raised when source OMML is outside the reversible project contract.""" |
| 80 | |
| 81 | |
| 82 | @dataclass(frozen=True) |
| 83 | class FormulaImport: |
| 84 | """Canonical formula source plus a visible SVG-preview representation.""" |
| 85 | |
| 86 | latex: str |
| 87 | preview: str |
| 88 | font_size_px: float |
| 89 | color: str |
| 90 | align: str |
| 91 | language: str |
| 92 | |
| 93 | |
| 94 | def import_formula(math_zone: ET.Element, *, display: bool) -> FormulaImport: |
| 95 | """Validate and reconstruct one ``a14:m`` formula zone.""" |
| 96 | if math_zone.tag != _A14_M: |
| 97 | raise FormulaImportError("formula carrier must be one a14:m element") |
| 98 | children = [child for child in math_zone if isinstance(child.tag, str)] |
| 99 | if len(children) != 1: |
| 100 | raise FormulaImportError("a14:m must contain exactly one Office Math root") |
| 101 | root = children[0] |
| 102 | expected = f"{_M}{'oMathPara' if display else 'oMath'}" |
| 103 | if root.tag != expected: |
| 104 | kind = "block" if display else "inline" |
| 105 | raise FormulaImportError( |
| 106 | f"{kind} formula requires m:{expected.rsplit('}', 1)[-1]}" |
| 107 | ) |
| 108 | |
| 109 | try: |
| 110 | canonical = validate_omml_fragment( |
| 111 | ET.tostring(root, encoding="unicode", short_empty_elements=True) |
| 112 | ) |
| 113 | validated_root = ET.fromstring(canonical) |
| 114 | latex = _serialize_root(validated_root) |
| 115 | if display: |
| 116 | compile_latex_to_omml(latex) |
| 117 | else: |
| 118 | compile_latex_to_inline_omml(latex) |
| 119 | except (FormulaCompileError, ET.ParseError, RecursionError) as exc: |
| 120 | raise FormulaImportError(str(exc)) from exc |
| 121 | |
| 122 | preview = _preview_root(validated_root).strip() |
| 123 | if not latex.strip() or not preview: |
| 124 | raise FormulaImportError("formula reconstruction produced empty content") |
| 125 | font_size_px, color, language = _dominant_run_style(validated_root) |
| 126 | return FormulaImport( |
| 127 | latex=latex, |
| 128 | preview=preview, |
| 129 | font_size_px=font_size_px, |
| 130 | color=color, |
| 131 | align=_formula_alignment(validated_root), |
| 132 | language=language, |
| 133 | ) |
| 134 | |
| 135 | |
| 136 | def opaque_formula_preview(math_zone: ET.Element) -> str: |
| 137 | """Return readable source text without claiming native reconstruction.""" |
| 138 | text = "".join( |
| 139 | item.text or "" |
| 140 | for item in math_zone.iter(f"{_M}t") |
| 141 | ).strip() |
| 142 | return text or "[unsupported formula]" |
| 143 | |
| 144 | |
| 145 | def _serialize_root(root: ET.Element) -> str: |
| 146 | if root.tag == f"{_M}oMathPara": |
| 147 | math = root.find(f"{_M}oMath") |
| 148 | if math is None: |
| 149 | raise FormulaImportError("m:oMathPara is missing m:oMath") |
| 150 | return _serialize_children(math) |
| 151 | if root.tag == f"{_M}oMath": |
| 152 | return _serialize_children(root) |
| 153 | raise FormulaImportError("unsupported Office Math root") |
| 154 | |
| 155 | |
| 156 | def _serialize_children(parent: ET.Element, *, alignment: bool = False) -> str: |
| 157 | return "".join( |
| 158 | _serialize_node(child, alignment=alignment) |
| 159 | for child in parent |
| 160 | if _local_name(child) not in {_PROPERTY_SUFFIX, "ctrlPr"} |
| 161 | and not _local_name(child).endswith(_PROPERTY_SUFFIX) |
| 162 | ) |
| 163 | |
| 164 | |
| 165 | def _serialize_argument(owner: ET.Element, name: str, *, alignment: bool = False) -> str: |
| 166 | element = owner.find(f"{_M}{name}") |
| 167 | if element is None: |
| 168 | raise FormulaImportError( |
| 169 | f"m:{_local_name(owner)} is missing required m:{name}" |
| 170 | ) |
| 171 | return _serialize_children(element, alignment=alignment) |
| 172 | |
| 173 | |
| 174 | def _serialize_node(element: ET.Element, *, alignment: bool = False) -> str: |
| 175 | name = _local_name(element) |
| 176 | if name == "r": |
| 177 | return _serialize_run(element, alignment=alignment) |
| 178 | if name == "f": |
| 179 | kind = _property_value(element, "fPr", "type", "bar") |
| 180 | command = {"bar": "frac", "skw": "ifrac"}.get(kind) |
| 181 | numerator = _serialize_argument(element, "num") |
| 182 | denominator = _serialize_argument(element, "den") |
| 183 | if command is not None: |
| 184 | return f"\\{command}{{{numerator}}}{{{denominator}}}" |
| 185 | if kind == "noBar": |
| 186 | return ( |
| 187 | r"\genfrac{}{}{0pt}{}" |
| 188 | f"{{{numerator}}}{{{denominator}}}" |
| 189 | ) |
| 190 | raise FormulaImportError(f"unsupported reversible fraction type: {kind!r}") |
| 191 | if name == "rad": |
| 192 | body = _serialize_argument(element, "e") |
| 193 | degree = _serialize_argument(element, "deg") |
| 194 | return f"\\sqrt[{degree}]{{{body}}}" if degree else f"\\sqrt{{{body}}}" |
| 195 | if name in {"sSub", "sSup", "sSubSup"}: |
| 196 | base = _serialize_argument(element, "e") |
| 197 | subscript = _serialize_optional_argument(element, "sub") |
| 198 | superscript = _serialize_optional_argument(element, "sup") |
| 199 | return _scripted(f"{{{base}}}", subscript, superscript) |
| 200 | if name == "sPre": |
| 201 | base = _serialize_argument(element, "e") |
| 202 | subscript = _serialize_optional_argument(element, "sub") |
| 203 | superscript = _serialize_optional_argument(element, "sup") |
| 204 | return _scripted("", subscript, superscript) + f"{{{base}}}" |
| 205 | if name == "nary": |
| 206 | symbol = _property_value(element, "naryPr", "chr", "") |
| 207 | command = _NARY_COMMAND_BY_SYMBOL.get(symbol) |
| 208 | if command is None: |
| 209 | raise FormulaImportError(f"unsupported reversible n-ary symbol: {symbol!r}") |
| 210 | source = f"\\{command}" |
| 211 | # Serialize the source location explicitly when it uses under/over |
| 212 | # limits. ``subSup`` is the neutral default in both formula contexts. |
| 213 | default_location = "subSup" |
| 214 | location = _property_value(element, "naryPr", "limLoc", default_location) |
| 215 | if location != default_location: |
| 216 | source += r"\limits" if location == "undOvr" else r"\nolimits" |
| 217 | source = _scripted( |
| 218 | source, |
| 219 | _serialize_optional_argument(element, "sub"), |
| 220 | _serialize_optional_argument(element, "sup"), |
| 221 | ) |
| 222 | body = _serialize_optional_argument(element, "e") |
| 223 | return source + (f"{{{body}}}" if body else "") |
| 224 | if name == "d": |
| 225 | left = _property_value(element, "dPr", "begChr", "(") |
| 226 | right = _property_value(element, "dPr", "endChr", ")") |
| 227 | separator = _property_value(element, "dPr", "sepChr", "") |
| 228 | segments = [ |
| 229 | _serialize_children(child) |
| 230 | for child in element.findall(f"{_M}e") |
| 231 | ] |
| 232 | middle = f"\\middle{_delimiter(separator, side='middle')}" if separator else "" |
| 233 | return ( |
| 234 | f"\\left{_delimiter(left, side='left')}" |
| 235 | + middle.join(segments) |
| 236 | + f"\\right{_delimiter(right, side='right')}" |
| 237 | ) |
| 238 | if name == "m": |
| 239 | rows = [] |
| 240 | for row in element.findall(f"{_M}mr"): |
| 241 | rows.append("&".join( |
| 242 | _serialize_children(cell) |
| 243 | for cell in row.findall(f"{_M}e") |
| 244 | )) |
| 245 | return r"\begin{matrix}" + r"\\".join(rows) + r"\end{matrix}" |
| 246 | if name == "eqArr": |
| 247 | rows = [ |
| 248 | _serialize_children(row, alignment=True) |
| 249 | for row in element.findall(f"{_M}e") |
| 250 | ] |
| 251 | return r"\begin{aligned}" + r"\\".join(rows) + r"\end{aligned}" |
| 252 | if name == "acc": |
| 253 | character = _property_value(element, "accPr", "chr", "") |
| 254 | command = _ACCENT_COMMAND_BY_CHARACTER.get(character) |
| 255 | if character == "̸": |
| 256 | command = "not" |
| 257 | if command is None: |
| 258 | raise FormulaImportError(f"unsupported reversible accent: {character!r}") |
| 259 | return f"\\{command}{{{_serialize_argument(element, 'e')}}}" |
| 260 | if name == "bar": |
| 261 | position = _property_value(element, "barPr", "pos", "top") |
| 262 | command = "overline" if position == "top" else "underline" |
| 263 | return f"\\{command}{{{_serialize_argument(element, 'e')}}}" |
| 264 | if name == "groupChr": |
| 265 | key = ( |
| 266 | _property_value(element, "groupChrPr", "chr", ""), |
| 267 | _property_value(element, "groupChrPr", "pos", "top"), |
| 268 | _property_value(element, "groupChrPr", "vertJc", "bot"), |
| 269 | ) |
| 270 | command = _GROUP_COMMANDS.get(key) |
| 271 | if command is None: |
| 272 | raise FormulaImportError(f"unsupported reversible group character: {key!r}") |
| 273 | return f"\\{command}{{{_serialize_argument(element, 'e')}}}" |
| 274 | if name == "limLow": |
| 275 | base = _serialize_argument(element, "e") |
| 276 | lower = _serialize_argument(element, "lim") |
| 277 | return f"\\underset{{{lower}}}{{{base}}}" |
| 278 | if name == "limUpp": |
| 279 | base = _serialize_argument(element, "e") |
| 280 | upper = _serialize_argument(element, "lim") |
| 281 | return f"\\overset{{{upper}}}{{{base}}}" |
| 282 | if name == "func": |
| 283 | function_name = _serialize_argument(element, "fName") |
| 284 | body = _serialize_optional_argument(element, "e") |
| 285 | source = f"\\mathop{{{function_name}}}" |
| 286 | return source + (f"{{{body}}}" if body else "") |
| 287 | if name == "box": |
| 288 | enabled = _property_value(element, "boxPr", "opEmu", "off") |
| 289 | if enabled not in {"on", "true", "1"}: |
| 290 | raise FormulaImportError("only operator-emulator m:box is reversible") |
| 291 | return f"\\mathrel{{{_serialize_argument(element, 'e')}}}" |
| 292 | if name == "phant": |
| 293 | properties = element.find(f"{_M}phantPr") |
| 294 | zero_width = _on_property(properties, "zeroWid") |
| 295 | zero_ascent = _on_property(properties, "zeroAsc") |
| 296 | zero_descent = _on_property(properties, "zeroDesc") |
| 297 | if zero_width: |
| 298 | command = "vphantom" |
| 299 | elif zero_ascent and zero_descent: |
| 300 | command = "hphantom" |
| 301 | else: |
| 302 | command = "phantom" |
| 303 | return f"\\{command}{{{_serialize_argument(element, 'e')}}}" |
| 304 | if name == "borderBox": |
| 305 | properties = element.find(f"{_M}borderBoxPr") |
| 306 | rising = _on_property(properties, "strikeBLTR") |
| 307 | falling = _on_property(properties, "strikeTLBR") |
| 308 | hidden = all( |
| 309 | _on_property(properties, side) |
| 310 | for side in ("hideTop", "hideBot", "hideLeft", "hideRight") |
| 311 | ) |
| 312 | if hidden and rising and falling: |
| 313 | command = "xcancel" |
| 314 | elif hidden and rising: |
| 315 | command = "cancel" |
| 316 | elif hidden and falling: |
| 317 | command = "bcancel" |
| 318 | elif not hidden and not rising and not falling: |
| 319 | command = "boxed" |
| 320 | else: |
| 321 | raise FormulaImportError("unsupported reversible border-box properties") |
| 322 | return f"\\{command}{{{_serialize_argument(element, 'e')}}}" |
| 323 | if name in {"e", "deg", "den", "fName", "lim", "num", "sub", "sup"}: |
| 324 | return _serialize_children(element, alignment=alignment) |
| 325 | raise FormulaImportError(f"unsupported reversible Office Math element: m:{name}") |
| 326 | |
| 327 | |
| 328 | def _serialize_optional_argument(owner: ET.Element, name: str) -> str: |
| 329 | element = owner.find(f"{_M}{name}") |
| 330 | return _serialize_children(element) if element is not None else "" |
| 331 | |
| 332 | |
| 333 | def _scripted(base: str, subscript: str, superscript: str) -> str: |
| 334 | if subscript: |
| 335 | base += f"_{{{subscript}}}" |
| 336 | if superscript: |
| 337 | base += f"^{{{superscript}}}" |
| 338 | return base |
| 339 | |
| 340 | |
| 341 | def _serialize_run(run: ET.Element, *, alignment: bool) -> str: |
| 342 | text_element = run.find(f"{_M}t") |
| 343 | value = text_element.text or "" if text_element is not None else "" |
| 344 | properties = run.find(f"{_M}rPr") |
| 345 | literal = _on_property(properties, "lit") |
| 346 | normal = _on_property(properties, "nor") |
| 347 | source = _escape_text(value, alignment=alignment and not literal, text_mode=normal) |
| 348 | if not source: |
| 349 | return "" |
| 350 | drawing = run.find(f"{_A}rPr") |
| 351 | if normal: |
| 352 | typeface = "" |
| 353 | if drawing is not None: |
| 354 | latin = drawing.find(f"{_A}latin") |
| 355 | typeface = latin.get("typeface", "") if latin is not None else "" |
| 356 | text_command = { |
| 357 | "Arial": "textsf", |
| 358 | "Courier New": "texttt", |
| 359 | }.get(typeface, "text") |
| 360 | source = f"\\{text_command}{{{source}}}" |
| 361 | if drawing is not None and _drawing_on(drawing, "b"): |
| 362 | source = f"\\textbf{{{source}}}" |
| 363 | if drawing is not None and _drawing_on(drawing, "i"): |
| 364 | source = f"\\textit{{{source}}}" |
| 365 | else: |
| 366 | script = _property_child_value(properties, "scr") |
| 367 | style = _property_child_value(properties, "sty") |
| 368 | command = { |
| 369 | "sans-serif": "mathsf", |
| 370 | "monospace": "mathtt", |
| 371 | "double-struck": "mathbb", |
| 372 | "script": "mathcal", |
| 373 | "fraktur": "mathfrak", |
| 374 | }.get(script or "") |
| 375 | if command is None: |
| 376 | command = { |
| 377 | "p": "mathrm", |
| 378 | "b": "mathbf", |
| 379 | "i": "mathit", |
| 380 | "bi": "boldsymbol", |
| 381 | }.get(style or "") |
| 382 | if command is not None: |
| 383 | source = f"\\{command}{{{source}}}" |
| 384 | color = _drawing_color(drawing) |
| 385 | if color is not None: |
| 386 | source = f"\\textcolor{{#{color}}}{{{source}}}" |
| 387 | return source |
| 388 | |
| 389 | |
| 390 | def _escape_text(value: str, *, alignment: bool, text_mode: bool) -> str: |
| 391 | parts: list[str] = [] |
| 392 | for character in value: |
| 393 | if character in _SPACING_LATEX and not text_mode: |
| 394 | parts.append(_SPACING_LATEX[character]) |
| 395 | elif character == " " and not text_mode: |
| 396 | parts.append(r"\ ") |
| 397 | elif character == "&": |
| 398 | parts.append("&" if alignment else r"\&") |
| 399 | elif character in "{}_%#$": |
| 400 | parts.append("\\" + character) |
| 401 | elif character == "\\": |
| 402 | parts.append(r"\backslash{}") |
| 403 | elif character == "~" and not text_mode: |
| 404 | parts.append(r"\text{~}") |
| 405 | elif character == "^" and not text_mode: |
| 406 | parts.append(r"\text{^}") |
| 407 | else: |
| 408 | parts.append(character) |
| 409 | return "".join(parts) |
| 410 | |
| 411 | |
| 412 | def _delimiter(value: str, *, side: str) -> str: |
| 413 | if not value: |
| 414 | return "." |
| 415 | if value == "|": |
| 416 | return "|" |
| 417 | command = _DELIMITER_COMMANDS.get(value) |
| 418 | if command is not None: |
| 419 | if value == "‖" and side == "left": |
| 420 | command = r"\lVert" |
| 421 | if value == "‖" and side == "right": |
| 422 | command = r"\rVert" |
| 423 | if command.startswith("\\") and command[1:].isalpha(): |
| 424 | return command + " " |
| 425 | return command |
| 426 | if len(value) == 1: |
| 427 | return value |
| 428 | raise FormulaImportError(f"unsupported reversible delimiter: {value!r}") |
| 429 | |
| 430 | |
| 431 | def _preview_root(root: ET.Element) -> str: |
| 432 | math = root.find(f"{_M}oMath") if root.tag == f"{_M}oMathPara" else root |
| 433 | if math is None: |
| 434 | return "" |
| 435 | return _preview_children(math) |
| 436 | |
| 437 | |
| 438 | def _preview_children(parent: ET.Element) -> str: |
| 439 | return "".join( |
| 440 | _preview_node(child) |
| 441 | for child in parent |
| 442 | if not _local_name(child).endswith(_PROPERTY_SUFFIX) |
| 443 | and _local_name(child) != "ctrlPr" |
| 444 | ) |
| 445 | |
| 446 | |
| 447 | def _preview_argument(owner: ET.Element, name: str) -> str: |
| 448 | child = owner.find(f"{_M}{name}") |
| 449 | return _preview_children(child) if child is not None else "" |
| 450 | |
| 451 | |
| 452 | def _preview_node(element: ET.Element) -> str: |
| 453 | name = _local_name(element) |
| 454 | if name == "r": |
| 455 | text = element.find(f"{_M}t") |
| 456 | return text.text or "" if text is not None else "" |
| 457 | if name == "f": |
| 458 | return f"({_preview_argument(element, 'num')})/({_preview_argument(element, 'den')})" |
| 459 | if name == "rad": |
| 460 | degree = _preview_argument(element, "deg") |
| 461 | prefix = f"√[{degree}]" if degree else "√" |
| 462 | return f"{prefix}({_preview_argument(element, 'e')})" |
| 463 | if name in {"sSub", "sSup", "sSubSup"}: |
| 464 | base = _preview_argument(element, "e") |
| 465 | subscript = _preview_argument(element, "sub") |
| 466 | superscript = _preview_argument(element, "sup") |
| 467 | return _preview_script(base, subscript, superscript) |
| 468 | if name == "sPre": |
| 469 | prefix = _preview_script( |
| 470 | "", |
| 471 | _preview_argument(element, "sub"), |
| 472 | _preview_argument(element, "sup"), |
| 473 | ) |
| 474 | return prefix + _preview_argument(element, "e") |
| 475 | if name == "nary": |
| 476 | symbol = _property_value(element, "naryPr", "chr", "") |
| 477 | return _preview_script( |
| 478 | symbol, |
| 479 | _preview_argument(element, "sub"), |
| 480 | _preview_argument(element, "sup"), |
| 481 | ) + _preview_argument(element, "e") |
| 482 | if name == "d": |
| 483 | left = _property_value(element, "dPr", "begChr", "(") |
| 484 | right = _property_value(element, "dPr", "endChr", ")") |
| 485 | separator = _property_value(element, "dPr", "sepChr", "") |
| 486 | segments = [_preview_children(child) for child in element.findall(f"{_M}e")] |
| 487 | return left + separator.join(segments) + right |
| 488 | if name == "m": |
| 489 | rows = [ |
| 490 | ", ".join(_preview_children(cell) for cell in row.findall(f"{_M}e")) |
| 491 | for row in element.findall(f"{_M}mr") |
| 492 | ] |
| 493 | return "[" + "; ".join(rows) + "]" |
| 494 | if name == "eqArr": |
| 495 | return " ".join( |
| 496 | _preview_children(row) |
| 497 | for row in element.findall(f"{_M}e") |
| 498 | ) |
| 499 | if name == "acc": |
| 500 | return _preview_argument(element, "e") + _property_value(element, "accPr", "chr", "") |
| 501 | if name == "bar": |
| 502 | marker = "¯" if _property_value(element, "barPr", "pos", "top") == "top" else "_" |
| 503 | return marker + f"({_preview_argument(element, 'e')})" |
| 504 | if name == "groupChr": |
| 505 | marker = _property_value(element, "groupChrPr", "chr", "") |
| 506 | return marker + f"({_preview_argument(element, 'e')})" |
| 507 | if name == "limLow": |
| 508 | return _preview_script( |
| 509 | _preview_argument(element, "e"), |
| 510 | _preview_argument(element, "lim"), |
| 511 | "", |
| 512 | ) |
| 513 | if name == "limUpp": |
| 514 | return _preview_script( |
| 515 | _preview_argument(element, "e"), |
| 516 | "", |
| 517 | _preview_argument(element, "lim"), |
| 518 | ) |
| 519 | if name == "func": |
| 520 | return _preview_argument(element, "fName") + _preview_argument(element, "e") |
| 521 | if name in {"box", "phant"}: |
| 522 | return _preview_argument(element, "e") |
| 523 | if name == "borderBox": |
| 524 | return "□(" + _preview_argument(element, "e") + ")" |
| 525 | if name in {"e", "deg", "den", "fName", "lim", "num", "sub", "sup"}: |
| 526 | return _preview_children(element) |
| 527 | return "" |
| 528 | |
| 529 | |
| 530 | def _preview_script(base: str, subscript: str, superscript: str) -> str: |
| 531 | if subscript: |
| 532 | base += f"_({subscript})" |
| 533 | if superscript: |
| 534 | base += f"^({superscript})" |
| 535 | return base |
| 536 | |
| 537 | |
| 538 | def _dominant_run_style(root: ET.Element) -> tuple[float, str, str]: |
| 539 | sizes: list[str] = [] |
| 540 | colors: list[str] = [] |
| 541 | languages: list[str] = [] |
| 542 | for properties in root.iter(f"{_A}rPr"): |
| 543 | if properties.get("sz"): |
| 544 | sizes.append(properties.get("sz") or "") |
| 545 | if properties.get("lang"): |
| 546 | languages.append(properties.get("lang") or "") |
| 547 | color = properties.find(f"{_A}solidFill/{_A}srgbClr") |
| 548 | if color is not None and color.get("val"): |
| 549 | colors.append(color.get("val") or "") |
| 550 | raw_size = _dominant(sizes, "2100") |
| 551 | try: |
| 552 | font_size = max(1.0, min(400.0, int(raw_size) / 75.0)) |
| 553 | except ValueError: |
| 554 | font_size = 28.0 |
| 555 | return ( |
| 556 | font_size, |
| 557 | f"#{_dominant(colors, '000000').upper()}", |
| 558 | _dominant(languages, "en-US"), |
| 559 | ) |
| 560 | |
| 561 | |
| 562 | def _formula_alignment(root: ET.Element) -> str: |
| 563 | if root.tag != f"{_M}oMathPara": |
| 564 | return "left" |
| 565 | value = _property_value(root, "oMathParaPr", "jc", "center") |
| 566 | return {"left": "left", "right": "right"}.get(value, "center") |
| 567 | |
| 568 | |
| 569 | def _dominant(values: list[str], default: str) -> str: |
| 570 | return Counter(values).most_common(1)[0][0] if values else default |
| 571 | |
| 572 | |
| 573 | def _property_value(owner: ET.Element, properties: str, name: str, default: str) -> str: |
| 574 | container = owner.find(f"{_M}{properties}") |
| 575 | value = _property_child_value(container, name) |
| 576 | return value if value is not None else default |
| 577 | |
| 578 | |
| 579 | def _property_child_value(owner: ET.Element | None, name: str) -> str | None: |
| 580 | if owner is None: |
| 581 | return None |
| 582 | child = owner.find(f"{_M}{name}") |
| 583 | return child.get(f"{_M}val") if child is not None else None |
| 584 | |
| 585 | |
| 586 | def _on_property(owner: ET.Element | None, name: str) -> bool: |
| 587 | return _property_child_value(owner, name) in {"on", "true", "1"} |
| 588 | |
| 589 | |
| 590 | def _drawing_on(owner: ET.Element, name: str) -> bool: |
| 591 | return owner.get(name) in {"on", "true", "1"} |
| 592 | |
| 593 | |
| 594 | def _drawing_color(owner: ET.Element | None) -> str | None: |
| 595 | if owner is None: |
| 596 | return None |
| 597 | color = owner.find(f"{_A}solidFill/{_A}srgbClr") |
| 598 | return color.get("val") if color is not None else None |
| 599 | |
| 600 | |
| 601 | def _local_name(element: ET.Element) -> str: |
| 602 | return element.tag.rsplit("}", 1)[-1] |
| 603 | |
| 604 | |
| 605 | __all__ = [ |
| 606 | "A14_NS", |
| 607 | "FormulaImport", |
| 608 | "FormulaImportError", |
| 609 | "import_formula", |
| 610 | "opaque_formula_preview", |
| 611 | ] |
| 612 |