tex2typst 0.2.7 → 0.2.8

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
package/dist/index.js CHANGED
@@ -1,614 +1,3 @@
1
- // src/parser.ts
2
- function assert(condition, message = "") {
3
- if (!condition) {
4
- throw new LatexParserError(message);
5
- }
6
- }
7
- function get_command_param_num(command) {
8
- if (UNARY_COMMANDS.includes(command)) {
9
- return 1;
10
- } else if (BINARY_COMMANDS.includes(command)) {
11
- return 2;
12
- } else {
13
- return 0;
14
- }
15
- }
16
- function find_closing_curly_bracket(tokens, start) {
17
- assert(token_eq(tokens[start], LEFT_CURLY_BRACKET));
18
- let count = 1;
19
- let pos = start + 1;
20
- while (count > 0) {
21
- if (pos >= tokens.length) {
22
- throw new LatexParserError("Unmatched curly brackets");
23
- }
24
- if (token_eq(tokens[pos], LEFT_CURLY_BRACKET)) {
25
- count += 1;
26
- } else if (token_eq(tokens[pos], RIGHT_CURLY_BRACKET)) {
27
- count -= 1;
28
- }
29
- pos += 1;
30
- }
31
- return pos - 1;
32
- }
33
- function find_closing_square_bracket(tokens, start) {
34
- assert(token_eq(tokens[start], LEFT_SQUARE_BRACKET));
35
- let count = 1;
36
- let pos = start + 1;
37
- while (count > 0) {
38
- if (pos >= tokens.length) {
39
- throw new LatexParserError("Unmatched square brackets");
40
- }
41
- if (token_eq(tokens[pos], LEFT_SQUARE_BRACKET)) {
42
- count += 1;
43
- } else if (token_eq(tokens[pos], RIGHT_SQUARE_BRACKET)) {
44
- count -= 1;
45
- }
46
- pos += 1;
47
- }
48
- return pos - 1;
49
- }
50
- function isalpha(char) {
51
- return "abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ".includes(char);
52
- }
53
- function isdigit(char) {
54
- return "0123456789".includes(char);
55
- }
56
- function eat_whitespaces(tokens, start) {
57
- let pos = start;
58
- while (pos < tokens.length && ["whitespace", "newline"].includes(tokens[pos].type)) {
59
- pos++;
60
- }
61
- return tokens.slice(start, pos);
62
- }
63
- function eat_parenthesis(tokens, start) {
64
- const firstToken = tokens[start];
65
- if (firstToken.type === "element" && ["(", ")", "[", "]", "|", "\\{", "\\}"].includes(firstToken.value)) {
66
- return firstToken;
67
- } else if (firstToken.type === "command" && ["lfloor", "rfloor", "lceil", "rceil", "langle", "rangle"].includes(firstToken.value.slice(1))) {
68
- return firstToken;
69
- } else {
70
- return null;
71
- }
72
- }
73
- function eat_primes(tokens, start) {
74
- let pos = start;
75
- while (pos < tokens.length && token_eq(tokens[pos], { type: "element", value: "'" })) {
76
- pos += 1;
77
- }
78
- return pos - start;
79
- }
80
- function eat_command_name(latex, start) {
81
- let pos = start;
82
- while (pos < latex.length && isalpha(latex[pos])) {
83
- pos += 1;
84
- }
85
- return latex.substring(start, pos);
86
- }
87
- function find_closing_right_command(tokens, start) {
88
- let count = 1;
89
- let pos = start;
90
- while (count > 0) {
91
- if (pos >= tokens.length) {
92
- return -1;
93
- }
94
- if (token_eq(tokens[pos], LEFT_COMMAND)) {
95
- count += 1;
96
- } else if (token_eq(tokens[pos], RIGHT_COMMAND)) {
97
- count -= 1;
98
- }
99
- pos += 1;
100
- }
101
- return pos - 1;
102
- }
103
- function find_closing_end_command(tokens, start) {
104
- let count = 1;
105
- let pos = start;
106
- while (count > 0) {
107
- if (pos >= tokens.length) {
108
- return -1;
109
- }
110
- if (token_eq(tokens[pos], BEGIN_COMMAND)) {
111
- count += 1;
112
- } else if (token_eq(tokens[pos], END_COMMAND)) {
113
- count -= 1;
114
- }
115
- pos += 1;
116
- }
117
- return pos - 1;
118
- }
119
- function find_closing_curly_bracket_char(latex, start) {
120
- assert(latex[start] === "{");
121
- let count = 1;
122
- let pos = start + 1;
123
- while (count > 0) {
124
- if (pos >= latex.length) {
125
- throw new LatexParserError("Unmatched curly brackets");
126
- }
127
- if (pos + 1 < latex.length && ["\\{", "\\}"].includes(latex.substring(pos, pos + 2))) {
128
- pos += 2;
129
- continue;
130
- }
131
- if (latex[pos] === "{") {
132
- count += 1;
133
- } else if (latex[pos] === "}") {
134
- count -= 1;
135
- }
136
- pos += 1;
137
- }
138
- return pos - 1;
139
- }
140
- function tokenize(latex) {
141
- const tokens = [];
142
- let pos = 0;
143
- while (pos < latex.length) {
144
- const firstChar = latex[pos];
145
- let token;
146
- switch (firstChar) {
147
- case "%": {
148
- let newPos = pos + 1;
149
- while (newPos < latex.length && latex[newPos] !== "\n") {
150
- newPos += 1;
151
- }
152
- token = { type: "comment", value: latex.slice(pos + 1, newPos) };
153
- pos = newPos;
154
- break;
155
- }
156
- case "{":
157
- case "}":
158
- case "_":
159
- case "^":
160
- case "&":
161
- token = { type: "control", value: firstChar };
162
- pos++;
163
- break;
164
- case "\n":
165
- token = { type: "newline", value: firstChar };
166
- pos++;
167
- break;
168
- case "\r": {
169
- if (pos + 1 < latex.length && latex[pos + 1] === "\n") {
170
- token = { type: "newline", value: "\n" };
171
- pos += 2;
172
- } else {
173
- token = { type: "newline", value: "\n" };
174
- pos++;
175
- }
176
- break;
177
- }
178
- case " ": {
179
- let newPos = pos;
180
- while (newPos < latex.length && latex[newPos] === " ") {
181
- newPos += 1;
182
- }
183
- token = { type: "whitespace", value: latex.slice(pos, newPos) };
184
- pos = newPos;
185
- break;
186
- }
187
- case "\\": {
188
- if (pos + 1 >= latex.length) {
189
- throw new LatexParserError("Expecting command name after \\");
190
- }
191
- const firstTwoChars = latex.slice(pos, pos + 2);
192
- if (["\\\\", "\\,"].includes(firstTwoChars)) {
193
- token = { type: "control", value: firstTwoChars };
194
- } else if (["\\{", "\\}", "\\%", "\\$", "\\&", "\\#", "\\_"].includes(firstTwoChars)) {
195
- token = { type: "element", value: firstTwoChars };
196
- } else {
197
- const command = eat_command_name(latex, pos + 1);
198
- token = { type: "command", value: "\\" + command };
199
- }
200
- pos += token.value.length;
201
- break;
202
- }
203
- default: {
204
- if (isdigit(firstChar)) {
205
- let newPos = pos;
206
- while (newPos < latex.length && isdigit(latex[newPos])) {
207
- newPos += 1;
208
- }
209
- token = { type: "element", value: latex.slice(pos, newPos) };
210
- } else if (isalpha(firstChar)) {
211
- token = { type: "element", value: firstChar };
212
- } else if ("+-*/=\'<>!.,;?()[]|".includes(firstChar)) {
213
- token = { type: "element", value: firstChar };
214
- } else {
215
- token = { type: "unknown", value: firstChar };
216
- }
217
- pos += token.value.length;
218
- }
219
- }
220
- tokens.push(token);
221
- if (token.type === "command" && ["\\text", "\\begin", "\\end"].includes(token.value)) {
222
- if (pos >= latex.length || latex[pos] !== "{") {
223
- throw new LatexParserError(`No content for ${token.value} command`);
224
- }
225
- tokens.push({ type: "control", value: "{" });
226
- const posClosingBracket = find_closing_curly_bracket_char(latex, pos);
227
- pos++;
228
- let textInside = latex.slice(pos, posClosingBracket);
229
- const chars = ["{", "}", "\\", "$", "&", "#", "_", "%"];
230
- for (const char of chars) {
231
- textInside = textInside.replaceAll("\\" + char, char);
232
- }
233
- tokens.push({ type: "text", value: textInside });
234
- tokens.push({ type: "control", value: "}" });
235
- pos = posClosingBracket + 1;
236
- }
237
- }
238
- return tokens;
239
- }
240
- function token_eq(token1, token2) {
241
- return token1.type == token2.type && token1.value == token2.value;
242
- }
243
- function passIgnoreWhitespaceBeforeScriptMark(tokens) {
244
- const is_script_mark = (token) => token_eq(token, SUB_SYMBOL) || token_eq(token, SUP_SYMBOL);
245
- let out_tokens = [];
246
- for (let i = 0;i < tokens.length; i++) {
247
- if (tokens[i].type === "whitespace" && i + 1 < tokens.length && is_script_mark(tokens[i + 1])) {
248
- continue;
249
- }
250
- if (tokens[i].type === "whitespace" && i - 1 >= 0 && is_script_mark(tokens[i - 1])) {
251
- continue;
252
- }
253
- out_tokens.push(tokens[i]);
254
- }
255
- return out_tokens;
256
- }
257
- function passExpandCustomTexMacros(tokens, customTexMacros) {
258
- let out_tokens = [];
259
- for (const token of tokens) {
260
- if (token.type === "command" && customTexMacros[token.value]) {
261
- const expanded_tokens = tokenize(customTexMacros[token.value]);
262
- out_tokens = out_tokens.concat(expanded_tokens);
263
- } else {
264
- out_tokens.push(token);
265
- }
266
- }
267
- return out_tokens;
268
- }
269
- function parseTex(tex, customTexMacros) {
270
- const parser = new LatexParser;
271
- let tokens = tokenize(tex);
272
- tokens = passIgnoreWhitespaceBeforeScriptMark(tokens);
273
- tokens = passExpandCustomTexMacros(tokens, customTexMacros);
274
- return parser.parse(tokens);
275
- }
276
- var UNARY_COMMANDS = [
277
- "sqrt",
278
- "text",
279
- "bar",
280
- "bold",
281
- "boldsymbol",
282
- "ddot",
283
- "dot",
284
- "hat",
285
- "mathbb",
286
- "mathbf",
287
- "mathcal",
288
- "mathfrak",
289
- "mathit",
290
- "mathrm",
291
- "mathscr",
292
- "mathsf",
293
- "mathtt",
294
- "operatorname",
295
- "overbrace",
296
- "overline",
297
- "pmb",
298
- "rm",
299
- "tilde",
300
- "underbrace",
301
- "underline",
302
- "vec",
303
- "widehat",
304
- "widetilde"
305
- ];
306
- var BINARY_COMMANDS = [
307
- "frac",
308
- "tfrac",
309
- "binom",
310
- "dbinom",
311
- "dfrac",
312
- "tbinom"
313
- ];
314
- var EMPTY_NODE = { type: "empty", content: "" };
315
- var LEFT_CURLY_BRACKET = { type: "control", value: "{" };
316
- var RIGHT_CURLY_BRACKET = { type: "control", value: "}" };
317
- var LEFT_SQUARE_BRACKET = { type: "element", value: "[" };
318
- var RIGHT_SQUARE_BRACKET = { type: "element", value: "]" };
319
- var LEFT_COMMAND = { type: "command", value: "\\left" };
320
- var RIGHT_COMMAND = { type: "command", value: "\\right" };
321
- var BEGIN_COMMAND = { type: "command", value: "\\begin" };
322
- var END_COMMAND = { type: "command", value: "\\end" };
323
-
324
- class LatexParserError extends Error {
325
- constructor(message) {
326
- super(message);
327
- this.name = "LatexParserError";
328
- }
329
- }
330
- var SUB_SYMBOL = { type: "control", value: "_" };
331
- var SUP_SYMBOL = { type: "control", value: "^" };
332
-
333
- class LatexParser {
334
- space_sensitive;
335
- newline_sensitive;
336
- constructor(space_sensitive = false, newline_sensitive = true) {
337
- this.space_sensitive = space_sensitive;
338
- this.newline_sensitive = newline_sensitive;
339
- }
340
- parse(tokens) {
341
- const results = [];
342
- let pos = 0;
343
- while (pos < tokens.length) {
344
- const results2 = [];
345
- let pos2 = 0;
346
- while (pos2 < tokens.length) {
347
- const [res, newPos] = this.parseNextExpr(tokens, pos2);
348
- pos2 = newPos;
349
- if (!this.space_sensitive && res.type === "whitespace") {
350
- continue;
351
- }
352
- if (!this.newline_sensitive && res.type === "newline") {
353
- continue;
354
- }
355
- if (res.type === "control" && res.content === "&") {
356
- throw new LatexParserError("Unexpected & outside of an alignment");
357
- }
358
- results2.push(res);
359
- }
360
- if (results2.length === 0) {
361
- return EMPTY_NODE;
362
- } else if (results2.length === 1) {
363
- return results2[0];
364
- } else {
365
- return { type: "ordgroup", content: "", args: results2 };
366
- }
367
- }
368
- if (results.length === 0) {
369
- return EMPTY_NODE;
370
- } else if (results.length === 1) {
371
- return results[0];
372
- } else {
373
- return { type: "ordgroup", content: "", args: results };
374
- }
375
- }
376
- parseNextExpr(tokens, start) {
377
- let [base, pos] = this.parseNextExprWithoutSupSub(tokens, start);
378
- let sub = null;
379
- let sup = null;
380
- let num_prime = 0;
381
- num_prime += eat_primes(tokens, pos);
382
- pos += num_prime;
383
- if (pos < tokens.length && token_eq(tokens[pos], SUB_SYMBOL)) {
384
- [sub, pos] = this.parseNextExprWithoutSupSub(tokens, pos + 1);
385
- num_prime += eat_primes(tokens, pos);
386
- pos += num_prime;
387
- if (pos < tokens.length && token_eq(tokens[pos], SUP_SYMBOL)) {
388
- [sup, pos] = this.parseNextExprWithoutSupSub(tokens, pos + 1);
389
- if (eat_primes(tokens, pos) > 0) {
390
- throw new LatexParserError("Double superscript");
391
- }
392
- }
393
- } else if (pos < tokens.length && token_eq(tokens[pos], SUP_SYMBOL)) {
394
- [sup, pos] = this.parseNextExprWithoutSupSub(tokens, pos + 1);
395
- if (eat_primes(tokens, pos) > 0) {
396
- throw new LatexParserError("Double superscript");
397
- }
398
- if (pos < tokens.length && token_eq(tokens[pos], SUB_SYMBOL)) {
399
- [sub, pos] = this.parseNextExprWithoutSupSub(tokens, pos + 1);
400
- if (eat_primes(tokens, pos) > 0) {
401
- throw new LatexParserError("Double superscript");
402
- }
403
- }
404
- }
405
- if (sub !== null || sup !== null || num_prime > 0) {
406
- const res = { base };
407
- if (sub) {
408
- res.sub = sub;
409
- }
410
- if (num_prime > 0) {
411
- res.sup = { type: "ordgroup", content: "", args: [] };
412
- for (let i = 0;i < num_prime; i++) {
413
- res.sup.args.push({ type: "symbol", content: "\\prime" });
414
- }
415
- if (sup) {
416
- res.sup.args.push(sup);
417
- }
418
- if (res.sup.args.length === 1) {
419
- res.sup = res.sup.args[0];
420
- }
421
- } else if (sup) {
422
- res.sup = sup;
423
- }
424
- return [{ type: "supsub", content: "", data: res }, pos];
425
- } else {
426
- return [base, pos];
427
- }
428
- }
429
- parseNextExprWithoutSupSub(tokens, start) {
430
- const firstToken = tokens[start];
431
- const tokenType = firstToken.type;
432
- switch (tokenType) {
433
- case "element":
434
- case "text":
435
- case "comment":
436
- case "whitespace":
437
- case "newline":
438
- return [{ type: tokenType, content: firstToken.value }, start + 1];
439
- case "command":
440
- if (token_eq(firstToken, BEGIN_COMMAND)) {
441
- return this.parseBeginEndExpr(tokens, start);
442
- } else if (token_eq(firstToken, LEFT_COMMAND)) {
443
- return this.parseLeftRightExpr(tokens, start);
444
- } else {
445
- return this.parseCommandExpr(tokens, start);
446
- }
447
- case "control":
448
- const controlChar = firstToken.value;
449
- switch (controlChar) {
450
- case "{":
451
- const posClosingBracket = find_closing_curly_bracket(tokens, start);
452
- const exprInside = tokens.slice(start + 1, posClosingBracket);
453
- return [this.parse(exprInside), posClosingBracket + 1];
454
- case "}":
455
- throw new LatexParserError("Unmatched '}'");
456
- case "\\\\":
457
- return [{ type: "control", content: "\\\\" }, start + 1];
458
- case "\\,":
459
- return [{ type: "control", content: "\\," }, start + 1];
460
- case "_": {
461
- return [EMPTY_NODE, start];
462
- }
463
- case "^": {
464
- return [EMPTY_NODE, start];
465
- }
466
- case "&":
467
- return [{ type: "control", content: "&" }, start + 1];
468
- default:
469
- throw new LatexParserError("Unknown control sequence");
470
- }
471
- default:
472
- throw new LatexParserError("Unknown token type");
473
- }
474
- }
475
- parseCommandExpr(tokens, start) {
476
- assert(tokens[start].type === "command");
477
- const command = tokens[start].value;
478
- let pos = start + 1;
479
- if (["left", "right", "begin", "end"].includes(command.slice(1))) {
480
- throw new LatexParserError("Unexpected command: " + command);
481
- }
482
- const paramNum = get_command_param_num(command.slice(1));
483
- if (paramNum === 0) {
484
- return [{ type: "symbol", content: command }, pos];
485
- } else if (paramNum === 1) {
486
- if (command === "\\sqrt" && pos < tokens.length && token_eq(tokens[pos], LEFT_SQUARE_BRACKET)) {
487
- const posLeftSquareBracket = pos;
488
- const posRightSquareBracket = find_closing_square_bracket(tokens, pos);
489
- const exprInside = tokens.slice(posLeftSquareBracket + 1, posRightSquareBracket);
490
- const exponent = this.parse(exprInside);
491
- const [arg12, newPos2] = this.parseNextExprWithoutSupSub(tokens, posRightSquareBracket + 1);
492
- return [{ type: "unaryFunc", content: command, args: [arg12], data: exponent }, newPos2];
493
- } else if (command === "\\text") {
494
- if (pos + 2 >= tokens.length) {
495
- throw new LatexParserError("Expecting content for \\text command");
496
- }
497
- assert(token_eq(tokens[pos], LEFT_CURLY_BRACKET));
498
- assert(tokens[pos + 1].type === "text");
499
- assert(token_eq(tokens[pos + 2], RIGHT_CURLY_BRACKET));
500
- const text = tokens[pos + 1].value;
501
- return [{ type: "text", content: text }, pos + 3];
502
- }
503
- let [arg1, newPos] = this.parseNextExprWithoutSupSub(tokens, pos);
504
- return [{ type: "unaryFunc", content: command, args: [arg1] }, newPos];
505
- } else if (paramNum === 2) {
506
- const [arg1, pos1] = this.parseNextExprWithoutSupSub(tokens, pos);
507
- const [arg2, pos2] = this.parseNextExprWithoutSupSub(tokens, pos1);
508
- return [{ type: "binaryFunc", content: command, args: [arg1, arg2] }, pos2];
509
- } else {
510
- throw new Error("Invalid number of parameters");
511
- }
512
- }
513
- parseLeftRightExpr(tokens, start) {
514
- assert(token_eq(tokens[start], LEFT_COMMAND));
515
- let pos = start + 1;
516
- pos += eat_whitespaces(tokens, pos).length;
517
- if (pos >= tokens.length) {
518
- throw new LatexParserError("Expecting delimiter after \\left");
519
- }
520
- const leftDelimiter = eat_parenthesis(tokens, pos);
521
- if (leftDelimiter === null) {
522
- throw new LatexParserError("Invalid delimiter after \\left");
523
- }
524
- pos++;
525
- const exprInsideStart = pos;
526
- const idx = find_closing_right_command(tokens, pos);
527
- if (idx === -1) {
528
- throw new LatexParserError("No matching \\right");
529
- }
530
- const exprInsideEnd = idx;
531
- pos = idx + 1;
532
- pos += eat_whitespaces(tokens, pos).length;
533
- if (pos >= tokens.length) {
534
- throw new LatexParserError("Expecting \\right after \\left");
535
- }
536
- const rightDelimiter = eat_parenthesis(tokens, pos);
537
- if (rightDelimiter === null) {
538
- throw new LatexParserError("Invalid delimiter after \\right");
539
- }
540
- pos++;
541
- const exprInside = tokens.slice(exprInsideStart, exprInsideEnd);
542
- const body = this.parse(exprInside);
543
- const args = [
544
- { type: "element", content: leftDelimiter.value },
545
- body,
546
- { type: "element", content: rightDelimiter.value }
547
- ];
548
- const res = { type: "leftright", content: "", args };
549
- return [res, pos];
550
- }
551
- parseBeginEndExpr(tokens, start) {
552
- assert(token_eq(tokens[start], BEGIN_COMMAND));
553
- let pos = start + 1;
554
- assert(token_eq(tokens[pos], LEFT_CURLY_BRACKET));
555
- assert(tokens[pos + 1].type === "text");
556
- assert(token_eq(tokens[pos + 2], RIGHT_CURLY_BRACKET));
557
- const envName = tokens[pos + 1].value;
558
- pos += 3;
559
- pos += eat_whitespaces(tokens, pos).length;
560
- const exprInsideStart = pos;
561
- const endIdx = find_closing_end_command(tokens, pos);
562
- if (endIdx === -1) {
563
- throw new LatexParserError("No matching \\end");
564
- }
565
- const exprInsideEnd = endIdx;
566
- pos = endIdx + 1;
567
- assert(token_eq(tokens[pos], LEFT_CURLY_BRACKET));
568
- assert(tokens[pos + 1].type === "text");
569
- assert(token_eq(tokens[pos + 2], RIGHT_CURLY_BRACKET));
570
- if (tokens[pos + 1].value !== envName) {
571
- throw new LatexParserError("Mismatched \\begin and \\end environments");
572
- }
573
- pos += 3;
574
- const exprInside = tokens.slice(exprInsideStart, exprInsideEnd);
575
- while (exprInside.length > 0 && ["whitespace", "newline"].includes(exprInside[exprInside.length - 1].type)) {
576
- exprInside.pop();
577
- }
578
- const body = this.parseAligned(exprInside);
579
- const res = { type: "beginend", content: envName, data: body };
580
- return [res, pos];
581
- }
582
- parseAligned(tokens) {
583
- let pos = 0;
584
- const allRows = [];
585
- let row = [];
586
- allRows.push(row);
587
- let group = { type: "ordgroup", content: "", args: [] };
588
- row.push(group);
589
- while (pos < tokens.length) {
590
- const [res, newPos] = this.parseNextExpr(tokens, pos);
591
- pos = newPos;
592
- if (res.type === "whitespace") {
593
- continue;
594
- } else if (res.type === "newline" && !this.newline_sensitive) {
595
- continue;
596
- } else if (res.type === "control" && res.content === "\\\\") {
597
- row = [];
598
- group = { type: "ordgroup", content: "", args: [] };
599
- row.push(group);
600
- allRows.push(row);
601
- } else if (res.type === "control" && res.content === "&") {
602
- group = { type: "ordgroup", content: "", args: [] };
603
- row.push(group);
604
- } else {
605
- group.args.push(res);
606
- }
607
- }
608
- return allRows;
609
- }
610
- }
611
-
612
1
  // src/map.ts
613
2
  var symbolMap = new Map([
614
3
  ["nonumber", ""],
@@ -909,7 +298,788 @@ var symbolMap = new Map([
909
298
  ["TeX", "#TeX"]
910
299
  ]);
911
300
 
301
+ // src/parser.ts
302
+ function assert(condition, message = "") {
303
+ if (!condition) {
304
+ throw new LatexParserError(message);
305
+ }
306
+ }
307
+ function get_command_param_num(command) {
308
+ if (UNARY_COMMANDS.includes(command)) {
309
+ return 1;
310
+ } else if (BINARY_COMMANDS.includes(command)) {
311
+ return 2;
312
+ } else {
313
+ return 0;
314
+ }
315
+ }
316
+ function find_closing_curly_bracket(tokens, start) {
317
+ assert(token_eq(tokens[start], LEFT_CURLY_BRACKET));
318
+ let count = 1;
319
+ let pos = start + 1;
320
+ while (count > 0) {
321
+ if (pos >= tokens.length) {
322
+ throw new LatexParserError("Unmatched curly brackets");
323
+ }
324
+ if (token_eq(tokens[pos], LEFT_CURLY_BRACKET)) {
325
+ count += 1;
326
+ } else if (token_eq(tokens[pos], RIGHT_CURLY_BRACKET)) {
327
+ count -= 1;
328
+ }
329
+ pos += 1;
330
+ }
331
+ return pos - 1;
332
+ }
333
+ function find_closing_square_bracket(tokens, start) {
334
+ assert(token_eq(tokens[start], LEFT_SQUARE_BRACKET));
335
+ let count = 1;
336
+ let pos = start + 1;
337
+ while (count > 0) {
338
+ if (pos >= tokens.length) {
339
+ throw new LatexParserError("Unmatched square brackets");
340
+ }
341
+ if (token_eq(tokens[pos], LEFT_SQUARE_BRACKET)) {
342
+ count += 1;
343
+ } else if (token_eq(tokens[pos], RIGHT_SQUARE_BRACKET)) {
344
+ count -= 1;
345
+ }
346
+ pos += 1;
347
+ }
348
+ return pos - 1;
349
+ }
350
+ function isalpha(char) {
351
+ return "abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ".includes(char);
352
+ }
353
+ function isdigit(char) {
354
+ return "0123456789".includes(char);
355
+ }
356
+ function eat_whitespaces(tokens, start) {
357
+ let pos = start;
358
+ while (pos < tokens.length && [4 /* WHITESPACE */, 5 /* NEWLINE */].includes(tokens[pos].type)) {
359
+ pos++;
360
+ }
361
+ return tokens.slice(start, pos);
362
+ }
363
+ function eat_parenthesis(tokens, start) {
364
+ const firstToken = tokens[start];
365
+ if (firstToken.type === 0 /* ELEMENT */ && ["(", ")", "[", "]", "|", "\\{", "\\}"].includes(firstToken.value)) {
366
+ return firstToken;
367
+ } else if (firstToken.type === 1 /* COMMAND */ && ["lfloor", "rfloor", "lceil", "rceil", "langle", "rangle"].includes(firstToken.value.slice(1))) {
368
+ return firstToken;
369
+ } else {
370
+ return null;
371
+ }
372
+ }
373
+ function eat_primes(tokens, start) {
374
+ let pos = start;
375
+ while (pos < tokens.length && token_eq(tokens[pos], { type: 0 /* ELEMENT */, value: "'" })) {
376
+ pos += 1;
377
+ }
378
+ return pos - start;
379
+ }
380
+ function eat_command_name(latex, start) {
381
+ let pos = start;
382
+ while (pos < latex.length && isalpha(latex[pos])) {
383
+ pos += 1;
384
+ }
385
+ return latex.substring(start, pos);
386
+ }
387
+ function find_closing_right_command(tokens, start) {
388
+ let count = 1;
389
+ let pos = start;
390
+ while (count > 0) {
391
+ if (pos >= tokens.length) {
392
+ return -1;
393
+ }
394
+ if (token_eq(tokens[pos], LEFT_COMMAND)) {
395
+ count += 1;
396
+ } else if (token_eq(tokens[pos], RIGHT_COMMAND)) {
397
+ count -= 1;
398
+ }
399
+ pos += 1;
400
+ }
401
+ return pos - 1;
402
+ }
403
+ function find_closing_end_command(tokens, start) {
404
+ let count = 1;
405
+ let pos = start;
406
+ while (count > 0) {
407
+ if (pos >= tokens.length) {
408
+ return -1;
409
+ }
410
+ if (token_eq(tokens[pos], BEGIN_COMMAND)) {
411
+ count += 1;
412
+ } else if (token_eq(tokens[pos], END_COMMAND)) {
413
+ count -= 1;
414
+ }
415
+ pos += 1;
416
+ }
417
+ return pos - 1;
418
+ }
419
+ function find_closing_curly_bracket_char(latex, start) {
420
+ assert(latex[start] === "{");
421
+ let count = 1;
422
+ let pos = start + 1;
423
+ while (count > 0) {
424
+ if (pos >= latex.length) {
425
+ throw new LatexParserError("Unmatched curly brackets");
426
+ }
427
+ if (pos + 1 < latex.length && ["\\{", "\\}"].includes(latex.substring(pos, pos + 2))) {
428
+ pos += 2;
429
+ continue;
430
+ }
431
+ if (latex[pos] === "{") {
432
+ count += 1;
433
+ } else if (latex[pos] === "}") {
434
+ count -= 1;
435
+ }
436
+ pos += 1;
437
+ }
438
+ return pos - 1;
439
+ }
440
+ function tokenize(latex) {
441
+ const tokens = [];
442
+ let pos = 0;
443
+ while (pos < latex.length) {
444
+ const firstChar = latex[pos];
445
+ let token;
446
+ switch (firstChar) {
447
+ case "%": {
448
+ let newPos = pos + 1;
449
+ while (newPos < latex.length && latex[newPos] !== "\n") {
450
+ newPos += 1;
451
+ }
452
+ token = { type: 3 /* COMMENT */, value: latex.slice(pos + 1, newPos) };
453
+ pos = newPos;
454
+ break;
455
+ }
456
+ case "{":
457
+ case "}":
458
+ case "_":
459
+ case "^":
460
+ case "&":
461
+ token = { type: 6 /* CONTROL */, value: firstChar };
462
+ pos++;
463
+ break;
464
+ case "\n":
465
+ token = { type: 5 /* NEWLINE */, value: firstChar };
466
+ pos++;
467
+ break;
468
+ case "\r": {
469
+ if (pos + 1 < latex.length && latex[pos + 1] === "\n") {
470
+ token = { type: 5 /* NEWLINE */, value: "\n" };
471
+ pos += 2;
472
+ } else {
473
+ token = { type: 5 /* NEWLINE */, value: "\n" };
474
+ pos++;
475
+ }
476
+ break;
477
+ }
478
+ case " ": {
479
+ let newPos = pos;
480
+ while (newPos < latex.length && latex[newPos] === " ") {
481
+ newPos += 1;
482
+ }
483
+ token = { type: 4 /* WHITESPACE */, value: latex.slice(pos, newPos) };
484
+ pos = newPos;
485
+ break;
486
+ }
487
+ case "\\": {
488
+ if (pos + 1 >= latex.length) {
489
+ throw new LatexParserError("Expecting command name after \\");
490
+ }
491
+ const firstTwoChars = latex.slice(pos, pos + 2);
492
+ if (["\\\\", "\\,"].includes(firstTwoChars)) {
493
+ token = { type: 6 /* CONTROL */, value: firstTwoChars };
494
+ } else if (["\\{", "\\}", "\\%", "\\$", "\\&", "\\#", "\\_"].includes(firstTwoChars)) {
495
+ token = { type: 0 /* ELEMENT */, value: firstTwoChars };
496
+ } else {
497
+ const command = eat_command_name(latex, pos + 1);
498
+ token = { type: 1 /* COMMAND */, value: "\\" + command };
499
+ }
500
+ pos += token.value.length;
501
+ break;
502
+ }
503
+ default: {
504
+ if (isdigit(firstChar)) {
505
+ let newPos = pos;
506
+ while (newPos < latex.length && isdigit(latex[newPos])) {
507
+ newPos += 1;
508
+ }
509
+ token = { type: 0 /* ELEMENT */, value: latex.slice(pos, newPos) };
510
+ } else if (isalpha(firstChar)) {
511
+ token = { type: 0 /* ELEMENT */, value: firstChar };
512
+ } else if ("+-*/=\'<>!.,;?()[]|".includes(firstChar)) {
513
+ token = { type: 0 /* ELEMENT */, value: firstChar };
514
+ } else {
515
+ token = { type: 7 /* UNKNOWN */, value: firstChar };
516
+ }
517
+ pos += token.value.length;
518
+ }
519
+ }
520
+ tokens.push(token);
521
+ if (token.type === 1 /* COMMAND */ && ["\\text", "\\operatorname", "\\begin", "\\end"].includes(token.value)) {
522
+ if (pos >= latex.length || latex[pos] !== "{") {
523
+ throw new LatexParserError(`No content for ${token.value} command`);
524
+ }
525
+ tokens.push({ type: 6 /* CONTROL */, value: "{" });
526
+ const posClosingBracket = find_closing_curly_bracket_char(latex, pos);
527
+ pos++;
528
+ let textInside = latex.slice(pos, posClosingBracket);
529
+ const chars = ["{", "}", "\\", "$", "&", "#", "_", "%"];
530
+ for (const char of chars) {
531
+ textInside = textInside.replaceAll("\\" + char, char);
532
+ }
533
+ tokens.push({ type: 2 /* TEXT */, value: textInside });
534
+ tokens.push({ type: 6 /* CONTROL */, value: "}" });
535
+ pos = posClosingBracket + 1;
536
+ }
537
+ }
538
+ return tokens;
539
+ }
540
+ function token_eq(token1, token2) {
541
+ return token1.type == token2.type && token1.value == token2.value;
542
+ }
543
+ function passIgnoreWhitespaceBeforeScriptMark(tokens) {
544
+ const is_script_mark = (token) => token_eq(token, SUB_SYMBOL) || token_eq(token, SUP_SYMBOL);
545
+ let out_tokens = [];
546
+ for (let i = 0;i < tokens.length; i++) {
547
+ if (tokens[i].type === 4 /* WHITESPACE */ && i + 1 < tokens.length && is_script_mark(tokens[i + 1])) {
548
+ continue;
549
+ }
550
+ if (tokens[i].type === 4 /* WHITESPACE */ && i - 1 >= 0 && is_script_mark(tokens[i - 1])) {
551
+ continue;
552
+ }
553
+ out_tokens.push(tokens[i]);
554
+ }
555
+ return out_tokens;
556
+ }
557
+ function passExpandCustomTexMacros(tokens, customTexMacros) {
558
+ let out_tokens = [];
559
+ for (const token of tokens) {
560
+ if (token.type === 1 /* COMMAND */ && customTexMacros[token.value]) {
561
+ const expanded_tokens = tokenize(customTexMacros[token.value]);
562
+ out_tokens = out_tokens.concat(expanded_tokens);
563
+ } else {
564
+ out_tokens.push(token);
565
+ }
566
+ }
567
+ return out_tokens;
568
+ }
569
+ function parseTex(tex, customTexMacros) {
570
+ const parser = new LatexParser;
571
+ let tokens = tokenize(tex);
572
+ tokens = passIgnoreWhitespaceBeforeScriptMark(tokens);
573
+ tokens = passExpandCustomTexMacros(tokens, customTexMacros);
574
+ return parser.parse(tokens);
575
+ }
576
+ var UNARY_COMMANDS = [
577
+ "sqrt",
578
+ "text",
579
+ "bar",
580
+ "bold",
581
+ "boldsymbol",
582
+ "ddot",
583
+ "dot",
584
+ "hat",
585
+ "mathbb",
586
+ "mathbf",
587
+ "mathcal",
588
+ "mathfrak",
589
+ "mathit",
590
+ "mathrm",
591
+ "mathscr",
592
+ "mathsf",
593
+ "mathtt",
594
+ "operatorname",
595
+ "overbrace",
596
+ "overline",
597
+ "pmb",
598
+ "rm",
599
+ "tilde",
600
+ "underbrace",
601
+ "underline",
602
+ "vec",
603
+ "widehat",
604
+ "widetilde"
605
+ ];
606
+ var BINARY_COMMANDS = [
607
+ "frac",
608
+ "tfrac",
609
+ "binom",
610
+ "dbinom",
611
+ "dfrac",
612
+ "tbinom"
613
+ ];
614
+ var EMPTY_NODE = { type: "empty", content: "" };
615
+ var LEFT_CURLY_BRACKET = { type: 6 /* CONTROL */, value: "{" };
616
+ var RIGHT_CURLY_BRACKET = { type: 6 /* CONTROL */, value: "}" };
617
+ var LEFT_SQUARE_BRACKET = { type: 0 /* ELEMENT */, value: "[" };
618
+ var RIGHT_SQUARE_BRACKET = { type: 0 /* ELEMENT */, value: "]" };
619
+ var LEFT_COMMAND = { type: 1 /* COMMAND */, value: "\\left" };
620
+ var RIGHT_COMMAND = { type: 1 /* COMMAND */, value: "\\right" };
621
+ var BEGIN_COMMAND = { type: 1 /* COMMAND */, value: "\\begin" };
622
+ var END_COMMAND = { type: 1 /* COMMAND */, value: "\\end" };
623
+
624
+ class LatexParserError extends Error {
625
+ constructor(message) {
626
+ super(message);
627
+ this.name = "LatexParserError";
628
+ }
629
+ }
630
+ var SUB_SYMBOL = { type: 6 /* CONTROL */, value: "_" };
631
+ var SUP_SYMBOL = { type: 6 /* CONTROL */, value: "^" };
632
+
633
+ class LatexParser {
634
+ space_sensitive;
635
+ newline_sensitive;
636
+ constructor(space_sensitive = false, newline_sensitive = true) {
637
+ this.space_sensitive = space_sensitive;
638
+ this.newline_sensitive = newline_sensitive;
639
+ }
640
+ parse(tokens) {
641
+ const results = [];
642
+ let pos = 0;
643
+ while (pos < tokens.length) {
644
+ const results2 = [];
645
+ let pos2 = 0;
646
+ while (pos2 < tokens.length) {
647
+ const [res, newPos] = this.parseNextExpr(tokens, pos2);
648
+ pos2 = newPos;
649
+ if (!this.space_sensitive && res.type === "whitespace") {
650
+ continue;
651
+ }
652
+ if (!this.newline_sensitive && res.type === "newline") {
653
+ continue;
654
+ }
655
+ if (res.type === "control" && res.content === "&") {
656
+ throw new LatexParserError("Unexpected & outside of an alignment");
657
+ }
658
+ results2.push(res);
659
+ }
660
+ if (results2.length === 0) {
661
+ return EMPTY_NODE;
662
+ } else if (results2.length === 1) {
663
+ return results2[0];
664
+ } else {
665
+ return { type: "ordgroup", content: "", args: results2 };
666
+ }
667
+ }
668
+ if (results.length === 0) {
669
+ return EMPTY_NODE;
670
+ } else if (results.length === 1) {
671
+ return results[0];
672
+ } else {
673
+ return { type: "ordgroup", content: "", args: results };
674
+ }
675
+ }
676
+ parseNextExpr(tokens, start) {
677
+ let [base, pos] = this.parseNextExprWithoutSupSub(tokens, start);
678
+ let sub = null;
679
+ let sup = null;
680
+ let num_prime = 0;
681
+ num_prime += eat_primes(tokens, pos);
682
+ pos += num_prime;
683
+ if (pos < tokens.length && token_eq(tokens[pos], SUB_SYMBOL)) {
684
+ [sub, pos] = this.parseNextExprWithoutSupSub(tokens, pos + 1);
685
+ num_prime += eat_primes(tokens, pos);
686
+ pos += num_prime;
687
+ if (pos < tokens.length && token_eq(tokens[pos], SUP_SYMBOL)) {
688
+ [sup, pos] = this.parseNextExprWithoutSupSub(tokens, pos + 1);
689
+ if (eat_primes(tokens, pos) > 0) {
690
+ throw new LatexParserError("Double superscript");
691
+ }
692
+ }
693
+ } else if (pos < tokens.length && token_eq(tokens[pos], SUP_SYMBOL)) {
694
+ [sup, pos] = this.parseNextExprWithoutSupSub(tokens, pos + 1);
695
+ if (eat_primes(tokens, pos) > 0) {
696
+ throw new LatexParserError("Double superscript");
697
+ }
698
+ if (pos < tokens.length && token_eq(tokens[pos], SUB_SYMBOL)) {
699
+ [sub, pos] = this.parseNextExprWithoutSupSub(tokens, pos + 1);
700
+ if (eat_primes(tokens, pos) > 0) {
701
+ throw new LatexParserError("Double superscript");
702
+ }
703
+ }
704
+ }
705
+ if (sub !== null || sup !== null || num_prime > 0) {
706
+ const res = { base };
707
+ if (sub) {
708
+ res.sub = sub;
709
+ }
710
+ if (num_prime > 0) {
711
+ res.sup = { type: "ordgroup", content: "", args: [] };
712
+ for (let i = 0;i < num_prime; i++) {
713
+ res.sup.args.push({ type: "element", content: "'" });
714
+ }
715
+ if (sup) {
716
+ res.sup.args.push(sup);
717
+ }
718
+ if (res.sup.args.length === 1) {
719
+ res.sup = res.sup.args[0];
720
+ }
721
+ } else if (sup) {
722
+ res.sup = sup;
723
+ }
724
+ return [{ type: "supsub", content: "", data: res }, pos];
725
+ } else {
726
+ return [base, pos];
727
+ }
728
+ }
729
+ parseNextExprWithoutSupSub(tokens, start) {
730
+ const firstToken = tokens[start];
731
+ const tokenType = firstToken.type;
732
+ switch (tokenType) {
733
+ case 0 /* ELEMENT */:
734
+ return [{ type: "element", content: firstToken.value }, start + 1];
735
+ case 2 /* TEXT */:
736
+ return [{ type: "text", content: firstToken.value }, start + 1];
737
+ case 3 /* COMMENT */:
738
+ return [{ type: "comment", content: firstToken.value }, start + 1];
739
+ case 4 /* WHITESPACE */:
740
+ return [{ type: "whitespace", content: firstToken.value }, start + 1];
741
+ case 5 /* NEWLINE */:
742
+ return [{ type: "newline", content: firstToken.value }, start + 1];
743
+ case 1 /* COMMAND */:
744
+ if (token_eq(firstToken, BEGIN_COMMAND)) {
745
+ return this.parseBeginEndExpr(tokens, start);
746
+ } else if (token_eq(firstToken, LEFT_COMMAND)) {
747
+ return this.parseLeftRightExpr(tokens, start);
748
+ } else {
749
+ return this.parseCommandExpr(tokens, start);
750
+ }
751
+ case 6 /* CONTROL */:
752
+ const controlChar = firstToken.value;
753
+ switch (controlChar) {
754
+ case "{":
755
+ const posClosingBracket = find_closing_curly_bracket(tokens, start);
756
+ const exprInside = tokens.slice(start + 1, posClosingBracket);
757
+ return [this.parse(exprInside), posClosingBracket + 1];
758
+ case "}":
759
+ throw new LatexParserError("Unmatched '}'");
760
+ case "\\\\":
761
+ return [{ type: "control", content: "\\\\" }, start + 1];
762
+ case "\\,":
763
+ return [{ type: "control", content: "\\," }, start + 1];
764
+ case "_": {
765
+ return [EMPTY_NODE, start];
766
+ }
767
+ case "^": {
768
+ return [EMPTY_NODE, start];
769
+ }
770
+ case "&":
771
+ return [{ type: "control", content: "&" }, start + 1];
772
+ default:
773
+ throw new LatexParserError("Unknown control sequence");
774
+ }
775
+ default:
776
+ throw new LatexParserError("Unknown token type");
777
+ }
778
+ }
779
+ parseCommandExpr(tokens, start) {
780
+ assert(tokens[start].type === 1 /* COMMAND */);
781
+ const command = tokens[start].value;
782
+ let pos = start + 1;
783
+ if (["left", "right", "begin", "end"].includes(command.slice(1))) {
784
+ throw new LatexParserError("Unexpected command: " + command);
785
+ }
786
+ const paramNum = get_command_param_num(command.slice(1));
787
+ switch (paramNum) {
788
+ case 0:
789
+ if (!symbolMap.has(command.slice(1))) {
790
+ return [{ type: "unknownMacro", content: command }, pos];
791
+ }
792
+ return [{ type: "symbol", content: command }, pos];
793
+ case 1: {
794
+ if (command === "\\sqrt" && pos < tokens.length && token_eq(tokens[pos], LEFT_SQUARE_BRACKET)) {
795
+ const posLeftSquareBracket = pos;
796
+ const posRightSquareBracket = find_closing_square_bracket(tokens, pos);
797
+ const exprInside = tokens.slice(posLeftSquareBracket + 1, posRightSquareBracket);
798
+ const exponent = this.parse(exprInside);
799
+ const [arg12, newPos2] = this.parseNextExprWithoutSupSub(tokens, posRightSquareBracket + 1);
800
+ return [{ type: "unaryFunc", content: command, args: [arg12], data: exponent }, newPos2];
801
+ } else if (command === "\\text") {
802
+ if (pos + 2 >= tokens.length) {
803
+ throw new LatexParserError("Expecting content for \\text command");
804
+ }
805
+ assert(token_eq(tokens[pos], LEFT_CURLY_BRACKET));
806
+ assert(tokens[pos + 1].type === 2 /* TEXT */);
807
+ assert(token_eq(tokens[pos + 2], RIGHT_CURLY_BRACKET));
808
+ const text = tokens[pos + 1].value;
809
+ return [{ type: "text", content: text }, pos + 3];
810
+ }
811
+ let [arg1, newPos] = this.parseNextExprWithoutSupSub(tokens, pos);
812
+ return [{ type: "unaryFunc", content: command, args: [arg1] }, newPos];
813
+ }
814
+ case 2: {
815
+ const [arg1, pos1] = this.parseNextExprWithoutSupSub(tokens, pos);
816
+ const [arg2, pos2] = this.parseNextExprWithoutSupSub(tokens, pos1);
817
+ return [{ type: "binaryFunc", content: command, args: [arg1, arg2] }, pos2];
818
+ }
819
+ default:
820
+ throw new Error("Invalid number of parameters");
821
+ }
822
+ }
823
+ parseLeftRightExpr(tokens, start) {
824
+ assert(token_eq(tokens[start], LEFT_COMMAND));
825
+ let pos = start + 1;
826
+ pos += eat_whitespaces(tokens, pos).length;
827
+ if (pos >= tokens.length) {
828
+ throw new LatexParserError("Expecting delimiter after \\left");
829
+ }
830
+ const leftDelimiter = eat_parenthesis(tokens, pos);
831
+ if (leftDelimiter === null) {
832
+ throw new LatexParserError("Invalid delimiter after \\left");
833
+ }
834
+ pos++;
835
+ const exprInsideStart = pos;
836
+ const idx = find_closing_right_command(tokens, pos);
837
+ if (idx === -1) {
838
+ throw new LatexParserError("No matching \\right");
839
+ }
840
+ const exprInsideEnd = idx;
841
+ pos = idx + 1;
842
+ pos += eat_whitespaces(tokens, pos).length;
843
+ if (pos >= tokens.length) {
844
+ throw new LatexParserError("Expecting \\right after \\left");
845
+ }
846
+ const rightDelimiter = eat_parenthesis(tokens, pos);
847
+ if (rightDelimiter === null) {
848
+ throw new LatexParserError("Invalid delimiter after \\right");
849
+ }
850
+ pos++;
851
+ const exprInside = tokens.slice(exprInsideStart, exprInsideEnd);
852
+ const body = this.parse(exprInside);
853
+ const args = [
854
+ { type: "element", content: leftDelimiter.value },
855
+ body,
856
+ { type: "element", content: rightDelimiter.value }
857
+ ];
858
+ const res = { type: "leftright", content: "", args };
859
+ return [res, pos];
860
+ }
861
+ parseBeginEndExpr(tokens, start) {
862
+ assert(token_eq(tokens[start], BEGIN_COMMAND));
863
+ let pos = start + 1;
864
+ assert(token_eq(tokens[pos], LEFT_CURLY_BRACKET));
865
+ assert(tokens[pos + 1].type === 2 /* TEXT */);
866
+ assert(token_eq(tokens[pos + 2], RIGHT_CURLY_BRACKET));
867
+ const envName = tokens[pos + 1].value;
868
+ pos += 3;
869
+ pos += eat_whitespaces(tokens, pos).length;
870
+ const exprInsideStart = pos;
871
+ const endIdx = find_closing_end_command(tokens, pos);
872
+ if (endIdx === -1) {
873
+ throw new LatexParserError("No matching \\end");
874
+ }
875
+ const exprInsideEnd = endIdx;
876
+ pos = endIdx + 1;
877
+ assert(token_eq(tokens[pos], LEFT_CURLY_BRACKET));
878
+ assert(tokens[pos + 1].type === 2 /* TEXT */);
879
+ assert(token_eq(tokens[pos + 2], RIGHT_CURLY_BRACKET));
880
+ if (tokens[pos + 1].value !== envName) {
881
+ throw new LatexParserError("Mismatched \\begin and \\end environments");
882
+ }
883
+ pos += 3;
884
+ const exprInside = tokens.slice(exprInsideStart, exprInsideEnd);
885
+ while (exprInside.length > 0 && [4 /* WHITESPACE */, 5 /* NEWLINE */].includes(exprInside[exprInside.length - 1].type)) {
886
+ exprInside.pop();
887
+ }
888
+ const body = this.parseAligned(exprInside);
889
+ const res = { type: "beginend", content: envName, data: body };
890
+ return [res, pos];
891
+ }
892
+ parseAligned(tokens) {
893
+ let pos = 0;
894
+ const allRows = [];
895
+ let row = [];
896
+ allRows.push(row);
897
+ let group = { type: "ordgroup", content: "", args: [] };
898
+ row.push(group);
899
+ while (pos < tokens.length) {
900
+ const [res, newPos] = this.parseNextExpr(tokens, pos);
901
+ pos = newPos;
902
+ if (res.type === "whitespace") {
903
+ continue;
904
+ } else if (res.type === "newline" && !this.newline_sensitive) {
905
+ continue;
906
+ } else if (res.type === "control" && res.content === "\\\\") {
907
+ row = [];
908
+ group = { type: "ordgroup", content: "", args: [] };
909
+ row.push(group);
910
+ allRows.push(row);
911
+ } else if (res.type === "control" && res.content === "&") {
912
+ group = { type: "ordgroup", content: "", args: [] };
913
+ row.push(group);
914
+ } else {
915
+ group.args.push(res);
916
+ }
917
+ }
918
+ return allRows;
919
+ }
920
+ }
921
+
912
922
  // src/writer.ts
923
+ function convertTree(node) {
924
+ switch (node.type) {
925
+ case "empty":
926
+ case "whitespace":
927
+ return { type: "empty", content: "" };
928
+ case "ordgroup":
929
+ return {
930
+ type: "group",
931
+ content: "",
932
+ args: node.args.map(convertTree)
933
+ };
934
+ case "element":
935
+ case "symbol":
936
+ return { type: "symbol", content: convertToken(node.content) };
937
+ case "text":
938
+ return { type: "text", content: node.content };
939
+ case "comment":
940
+ return { type: "comment", content: node.content };
941
+ case "supsub": {
942
+ let { base, sup, sub } = node.data;
943
+ if (base && base.type === "unaryFunc" && base.content === "\\overbrace" && sup) {
944
+ return {
945
+ type: "binaryFunc",
946
+ content: "overbrace",
947
+ args: [convertTree(base.args[0]), convertTree(sup)]
948
+ };
949
+ } else if (base && base.type === "unaryFunc" && base.content === "\\underbrace" && sub) {
950
+ return {
951
+ type: "binaryFunc",
952
+ content: "underbrace",
953
+ args: [convertTree(base.args[0]), convertTree(sub)]
954
+ };
955
+ }
956
+ const data = {
957
+ base: convertTree(base)
958
+ };
959
+ if (data.base.type === "empty") {
960
+ data.base = { type: "text", content: "" };
961
+ }
962
+ if (sup) {
963
+ data.sup = convertTree(sup);
964
+ }
965
+ if (sub) {
966
+ data.sub = convertTree(sub);
967
+ }
968
+ return {
969
+ type: "supsub",
970
+ content: "",
971
+ data
972
+ };
973
+ }
974
+ case "leftright": {
975
+ const [left, body, right] = node.args;
976
+ const group = {
977
+ type: "group",
978
+ content: "",
979
+ args: node.args.map(convertTree)
980
+ };
981
+ if (["[]", "()", "\\{\\}", "\\lfloor\\rfloor", "\\lceil\\rceil"].includes(left.content + right.content)) {
982
+ return group;
983
+ }
984
+ return {
985
+ type: "unaryFunc",
986
+ content: "lr",
987
+ args: [group]
988
+ };
989
+ }
990
+ case "binaryFunc": {
991
+ return {
992
+ type: "binaryFunc",
993
+ content: convertToken(node.content),
994
+ args: node.args.map(convertTree)
995
+ };
996
+ }
997
+ case "unaryFunc": {
998
+ const arg0 = convertTree(node.args[0]);
999
+ if (node.content === "\\sqrt" && node.data) {
1000
+ const data = convertTree(node.data);
1001
+ return {
1002
+ type: "binaryFunc",
1003
+ content: "root",
1004
+ args: [data, arg0]
1005
+ };
1006
+ }
1007
+ if (node.content === "\\mathbf") {
1008
+ const inner = {
1009
+ type: "unaryFunc",
1010
+ content: "bold",
1011
+ args: [arg0]
1012
+ };
1013
+ return {
1014
+ type: "unaryFunc",
1015
+ content: "upright",
1016
+ args: [inner]
1017
+ };
1018
+ }
1019
+ if (node.content === "\\mathbb" && arg0.type === "symbol" && /^[A-Z]$/.test(arg0.content)) {
1020
+ return {
1021
+ type: "symbol",
1022
+ content: arg0.content + arg0.content
1023
+ };
1024
+ }
1025
+ if (node.content === "\\operatorname") {
1026
+ const body = node.args;
1027
+ if (body.length !== 1 || body[0].type !== "text") {
1028
+ throw new TypstWriterError(`Expecting body of \\operatorname to be text but got`, node);
1029
+ }
1030
+ const text = body[0].content;
1031
+ if (TYPST_INTRINSIC_SYMBOLS.includes(text)) {
1032
+ return {
1033
+ type: "symbol",
1034
+ content: text
1035
+ };
1036
+ } else {
1037
+ return {
1038
+ type: "unaryFunc",
1039
+ content: "op",
1040
+ args: [{ type: "text", content: text }]
1041
+ };
1042
+ }
1043
+ }
1044
+ return {
1045
+ type: "unaryFunc",
1046
+ content: convertToken(node.content),
1047
+ args: node.args.map(convertTree)
1048
+ };
1049
+ }
1050
+ case "newline":
1051
+ return { type: "newline", content: "\n" };
1052
+ case "beginend": {
1053
+ const matrix = node.data;
1054
+ const data = matrix.map((row) => row.map(convertTree));
1055
+ if (node.content.startsWith("align")) {
1056
+ return {
1057
+ type: "align",
1058
+ content: "",
1059
+ data
1060
+ };
1061
+ } else {
1062
+ return {
1063
+ type: "matrix",
1064
+ content: "mat",
1065
+ data
1066
+ };
1067
+ }
1068
+ }
1069
+ case "unknownMacro":
1070
+ return { type: "unknown", content: convertToken(node.content) };
1071
+ case "control":
1072
+ if (node.content === "\\\\") {
1073
+ return { type: "symbol", content: "\\" };
1074
+ } else if (node.content === "\\,") {
1075
+ return { type: "symbol", content: "thin" };
1076
+ } else {
1077
+ throw new TypstWriterError(`Unknown control sequence: ${node.content}`, node);
1078
+ }
1079
+ default:
1080
+ throw new TypstWriterError(`Unimplemented node type: ${node.type}`, node);
1081
+ }
1082
+ }
913
1083
  function convertToken(token) {
914
1084
  if (/^[a-zA-Z0-9]$/.test(token)) {
915
1085
  return token;
@@ -982,142 +1152,74 @@ class TypstWriter {
982
1152
  this.buffer += str;
983
1153
  }
984
1154
  append(node) {
985
- if (node.type === "empty" || node.type === "whitespace") {
986
- return;
987
- } else if (node.type === "ordgroup") {
988
- node.args.forEach((arg) => this.append(arg));
989
- } else if (node.type === "element") {
990
- let content = node.content;
991
- if (node.content === "," && this.insideFunctionDepth > 0) {
992
- content = "comma";
993
- }
994
- this.queue.push({ type: "symbol", content });
995
- } else if (node.type === "symbol") {
996
- this.queue.push({ type: "symbol", content: node.content });
997
- } else if (node.type === "text") {
998
- this.queue.push(node);
999
- } else if (node.type === "supsub") {
1000
- let { base, sup, sub } = node.data;
1001
- if (base && base.type === "unaryFunc" && base.content === "\\overbrace" && sup) {
1002
- this.append({ type: "binaryFunc", content: "\\overbrace", args: [base.args[0], sup] });
1003
- return;
1004
- } else if (base && base.type === "unaryFunc" && base.content === "\\underbrace" && sub) {
1005
- this.append({ type: "binaryFunc", content: "\\underbrace", args: [base.args[0], sub] });
1006
- return;
1155
+ switch (node.type) {
1156
+ case "empty":
1157
+ break;
1158
+ case "symbol": {
1159
+ let content = node.content;
1160
+ if (node.content === "," && this.insideFunctionDepth > 0) {
1161
+ content = "comma";
1162
+ }
1163
+ this.queue.push({ type: "symbol", content });
1164
+ break;
1007
1165
  }
1008
- if (base.type === "empty") {
1009
- this.queue.push({ type: "text", content: "" });
1010
- } else {
1166
+ case "text":
1167
+ case "comment":
1168
+ case "newline":
1169
+ this.queue.push(node);
1170
+ break;
1171
+ case "group":
1172
+ for (const item of node.args) {
1173
+ this.append(item);
1174
+ }
1175
+ break;
1176
+ case "supsub": {
1177
+ let { base, sup, sub } = node.data;
1011
1178
  this.appendWithBracketsIfNeeded(base);
1179
+ let trailing_space_needed = false;
1180
+ const has_prime = sup && sup.type === "symbol" && sup.content === "\'";
1181
+ if (has_prime) {
1182
+ this.queue.push({ type: "atom", content: "\'" });
1183
+ trailing_space_needed = false;
1184
+ }
1185
+ if (sub) {
1186
+ this.queue.push({ type: "atom", content: "_" });
1187
+ trailing_space_needed = this.appendWithBracketsIfNeeded(sub);
1188
+ }
1189
+ if (sup && !has_prime) {
1190
+ this.queue.push({ type: "atom", content: "^" });
1191
+ trailing_space_needed = this.appendWithBracketsIfNeeded(sup);
1192
+ }
1193
+ if (trailing_space_needed) {
1194
+ this.queue.push({ type: "softSpace", content: "" });
1195
+ }
1196
+ break;
1012
1197
  }
1013
- let trailing_space_needed = false;
1014
- const has_prime = sup && sup.type === "symbol" && sup.content === "\\prime";
1015
- if (has_prime) {
1016
- this.queue.push({ type: "atom", content: "\'" });
1017
- trailing_space_needed = false;
1018
- }
1019
- if (sub) {
1020
- this.queue.push({ type: "atom", content: "_" });
1021
- trailing_space_needed = this.appendWithBracketsIfNeeded(sub);
1022
- }
1023
- if (sup && !has_prime) {
1024
- this.queue.push({ type: "atom", content: "^" });
1025
- trailing_space_needed = this.appendWithBracketsIfNeeded(sup);
1026
- }
1027
- if (trailing_space_needed) {
1028
- this.queue.push({ type: "softSpace", content: "" });
1029
- }
1030
- } else if (node.type === "leftright") {
1031
- const [left, body, right] = node.args;
1032
- if (["[]", "()", "\\{\\}", "\\lfloor\\rfloor", "\\lceil\\rceil"].includes(left.content + right.content)) {
1033
- this.append(left);
1034
- this.append(body);
1035
- this.append(right);
1036
- return;
1037
- }
1038
- const func_symbol = { type: "symbol", content: "lr" };
1039
- this.queue.push(func_symbol);
1040
- this.insideFunctionDepth++;
1041
- this.queue.push({ type: "atom", content: "(" });
1042
- this.append(left);
1043
- this.append(body);
1044
- this.append(right);
1045
- this.queue.push({ type: "atom", content: ")" });
1046
- this.insideFunctionDepth--;
1047
- } else if (node.type === "binaryFunc") {
1048
- const func_symbol = { type: "symbol", content: node.content };
1049
- const [arg0, arg1] = node.args;
1050
- this.queue.push(func_symbol);
1051
- this.insideFunctionDepth++;
1052
- this.queue.push({ type: "atom", content: "(" });
1053
- this.append(arg0);
1054
- this.queue.push({ type: "atom", content: "," });
1055
- this.append(arg1);
1056
- this.queue.push({ type: "atom", content: ")" });
1057
- this.insideFunctionDepth--;
1058
- } else if (node.type === "unaryFunc") {
1059
- const func_symbol = { type: "symbol", content: node.content };
1060
- const arg0 = node.args[0];
1061
- if (node.content === "\\sqrt" && node.data) {
1062
- func_symbol.content = "root";
1198
+ case "binaryFunc": {
1199
+ const func_symbol = { type: "symbol", content: node.content };
1200
+ const [arg0, arg1] = node.args;
1063
1201
  this.queue.push(func_symbol);
1064
1202
  this.insideFunctionDepth++;
1065
1203
  this.queue.push({ type: "atom", content: "(" });
1066
- this.append(node.data);
1067
- this.queue.push({ type: "atom", content: "," });
1068
1204
  this.append(arg0);
1205
+ this.queue.push({ type: "atom", content: "," });
1206
+ this.append(arg1);
1069
1207
  this.queue.push({ type: "atom", content: ")" });
1070
1208
  this.insideFunctionDepth--;
1071
- return;
1072
- } else if (node.content === "\\mathbf") {
1073
- this.append({ type: "symbol", content: "upright" });
1074
- this.insideFunctionDepth++;
1075
- this.queue.push({ type: "atom", content: "(" });
1209
+ break;
1210
+ }
1211
+ case "unaryFunc": {
1212
+ const func_symbol = { type: "symbol", content: node.content };
1213
+ const arg0 = node.args[0];
1076
1214
  this.queue.push(func_symbol);
1077
1215
  this.insideFunctionDepth++;
1078
1216
  this.queue.push({ type: "atom", content: "(" });
1079
1217
  this.append(arg0);
1080
1218
  this.queue.push({ type: "atom", content: ")" });
1081
1219
  this.insideFunctionDepth--;
1082
- this.queue.push({ type: "atom", content: ")" });
1083
- this.insideFunctionDepth--;
1084
- return;
1085
- } else if (node.content === "\\mathbb") {
1086
- const body = node.args[0];
1087
- if (body.type === "element" && /^[A-Z]$/.test(body.content)) {
1088
- this.queue.push({ type: "symbol", content: body.content + body.content });
1089
- return;
1090
- }
1091
- } else if (node.content === "\\operatorname") {
1092
- let body = node.args;
1093
- if (body.length === 1 && body[0].type == "ordgroup") {
1094
- body = body[0].args;
1095
- }
1096
- const text = body.reduce((buff, n) => {
1097
- buff += convertToken(n.content);
1098
- return buff;
1099
- }, "");
1100
- if (this.preferTypstIntrinsic && TYPST_INTRINSIC_SYMBOLS.includes(text)) {
1101
- this.queue.push({ type: "symbol", content: text });
1102
- } else {
1103
- this.queue.push({ type: "symbol", content: "op" });
1104
- this.queue.push({ type: "atom", content: "(" });
1105
- this.queue.push({ type: "text", content: text });
1106
- this.queue.push({ type: "atom", content: ")" });
1107
- }
1108
- return;
1220
+ break;
1109
1221
  }
1110
- this.queue.push(func_symbol);
1111
- this.insideFunctionDepth++;
1112
- this.queue.push({ type: "atom", content: "(" });
1113
- this.append(arg0);
1114
- this.queue.push({ type: "atom", content: ")" });
1115
- this.insideFunctionDepth--;
1116
- } else if (node.type === "newline") {
1117
- this.queue.push({ type: "newline", content: "\n" });
1118
- return;
1119
- } else if (node.type === "beginend") {
1120
- if (node.content.startsWith("align")) {
1222
+ case "align": {
1121
1223
  const matrix = node.data;
1122
1224
  matrix.forEach((row, i) => {
1123
1225
  row.forEach((cell, j) => {
@@ -1127,10 +1229,12 @@ class TypstWriter {
1127
1229
  this.append(cell);
1128
1230
  });
1129
1231
  if (i < matrix.length - 1) {
1130
- this.queue.push({ type: "symbol", content: "\\\\" });
1232
+ this.queue.push({ type: "symbol", content: "\\" });
1131
1233
  }
1132
1234
  });
1133
- } else {
1235
+ break;
1236
+ }
1237
+ case "matrix": {
1134
1238
  const matrix = node.data;
1135
1239
  this.queue.push({ type: "symbol", content: "mat" });
1136
1240
  this.insideFunctionDepth++;
@@ -1138,10 +1242,6 @@ class TypstWriter {
1138
1242
  this.queue.push({ type: "symbol", content: "delim: #none, " });
1139
1243
  matrix.forEach((row, i) => {
1140
1244
  row.forEach((cell, j) => {
1141
- if (cell.type === "ordgroup" && cell.args.length === 0) {
1142
- this.queue.push({ type: "atom", content: "," });
1143
- return;
1144
- }
1145
1245
  this.append(cell);
1146
1246
  if (j < row.length - 1) {
1147
1247
  this.queue.push({ type: "atom", content: "," });
@@ -1154,37 +1254,44 @@ class TypstWriter {
1154
1254
  });
1155
1255
  this.queue.push({ type: "atom", content: ")" });
1156
1256
  this.insideFunctionDepth--;
1257
+ break;
1157
1258
  }
1158
- } else if (node.type === "matrix") {
1159
- } else if (node.type === "unknownMacro") {
1160
- if (this.nonStrict) {
1161
- this.queue.push({ type: "symbol", content: node.content });
1162
- } else {
1163
- throw new TypstWriterError(`Unknown macro: ${node.content}`, node);
1164
- }
1165
- } else if (node.type === "control") {
1166
- if (node.content === "\\\\") {
1167
- this.queue.push({ type: "symbol", content: node.content });
1168
- } else if (node.content === "\\,") {
1169
- this.queue.push({ type: "symbol", content: "thin" });
1170
- } else {
1171
- throw new TypstWriterError(`Unknown control sequence: ${node.content}`, node);
1259
+ case "unknown": {
1260
+ if (this.nonStrict) {
1261
+ this.queue.push({ type: "symbol", content: node.content });
1262
+ } else {
1263
+ throw new TypstWriterError(`Unknown macro: ${node.content}`, node);
1264
+ }
1265
+ break;
1172
1266
  }
1173
- } else if (node.type === "comment") {
1174
- this.queue.push({ type: "comment", content: node.content });
1267
+ default:
1268
+ throw new TypstWriterError(`Unimplemented node type to append: ${node.type}`, node);
1269
+ }
1270
+ }
1271
+ appendWithBracketsIfNeeded(node) {
1272
+ const is_single = !["group", "supsub", "empty"].includes(node.type);
1273
+ if (is_single) {
1274
+ this.append(node);
1175
1275
  } else {
1176
- throw new TypstWriterError(`Unimplemented node type to append: ${node.type}`, node);
1276
+ this.queue.push({
1277
+ type: "atom",
1278
+ content: "("
1279
+ });
1280
+ this.append(node);
1281
+ this.queue.push({
1282
+ type: "atom",
1283
+ content: ")"
1284
+ });
1177
1285
  }
1286
+ return is_single;
1178
1287
  }
1179
1288
  flushQueue() {
1180
1289
  this.queue.forEach((node) => {
1181
1290
  let str = "";
1182
1291
  switch (node.type) {
1183
1292
  case "atom":
1184
- str = node.content;
1185
- break;
1186
1293
  case "symbol":
1187
- str = convertToken(node.content);
1294
+ str = node.content;
1188
1295
  break;
1189
1296
  case "text":
1190
1297
  str = `"${node.content}"`;
@@ -1208,23 +1315,6 @@ class TypstWriter {
1208
1315
  });
1209
1316
  this.queue = [];
1210
1317
  }
1211
- appendWithBracketsIfNeeded(node) {
1212
- const is_single = ["symbol", "element", "unaryFunc", "binaryFunc", "leftright"].includes(node.type);
1213
- if (is_single) {
1214
- this.append(node);
1215
- } else {
1216
- this.queue.push({
1217
- type: "atom",
1218
- content: "("
1219
- });
1220
- this.append(node);
1221
- this.queue.push({
1222
- type: "atom",
1223
- content: ")"
1224
- });
1225
- }
1226
- return is_single;
1227
- }
1228
1318
  finalize() {
1229
1319
  this.flushQueue();
1230
1320
  const smartFloorPass = function(input) {
@@ -1246,7 +1336,7 @@ class TypstWriter {
1246
1336
  // src/index.ts
1247
1337
  function tex2typst(tex, options) {
1248
1338
  const opt = {
1249
- nonStrict: false,
1339
+ nonStrict: true,
1250
1340
  preferTypstIntrinsic: true,
1251
1341
  customTexMacros: {}
1252
1342
  };
@@ -1261,9 +1351,10 @@ function tex2typst(tex, options) {
1261
1351
  opt.customTexMacros = options.customTexMacros;
1262
1352
  }
1263
1353
  }
1264
- const t = parseTex(tex, opt.customTexMacros);
1354
+ const texTree = parseTex(tex, opt.customTexMacros);
1355
+ const typstTree = convertTree(texTree);
1265
1356
  const writer2 = new TypstWriter(opt.nonStrict, opt.preferTypstIntrinsic);
1266
- writer2.append(t);
1357
+ writer2.append(typstTree);
1267
1358
  return writer2.finalize();
1268
1359
  }
1269
1360
  export {