tex2typst 0.2.6 → 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
@@ -1,607 +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 parseTex(tex, customTexMacros) {
244
- const parser = new LatexParser;
245
- const original_tokens = tokenize(tex);
246
- let processed_tokens = [];
247
- for (const token of original_tokens) {
248
- if (token.type === "command" && customTexMacros[token.value]) {
249
- const expanded_tokens = tokenize(customTexMacros[token.value]);
250
- processed_tokens = processed_tokens.concat(expanded_tokens);
251
- } else {
252
- processed_tokens.push(token);
253
- }
254
- }
255
- return parser.parse(processed_tokens);
256
- }
257
- var UNARY_COMMANDS = [
258
- "sqrt",
259
- "text",
260
- "bar",
261
- "bold",
262
- "boldsymbol",
263
- "ddot",
264
- "dot",
265
- "hat",
266
- "mathbb",
267
- "mathbf",
268
- "mathcal",
269
- "mathfrak",
270
- "mathit",
271
- "mathrm",
272
- "mathscr",
273
- "mathsf",
274
- "mathtt",
275
- "operatorname",
276
- "overbrace",
277
- "overline",
278
- "pmb",
279
- "rm",
280
- "tilde",
281
- "underbrace",
282
- "underline",
283
- "vec",
284
- "widehat",
285
- "widetilde"
286
- ];
287
- var BINARY_COMMANDS = [
288
- "frac",
289
- "tfrac",
290
- "binom",
291
- "dbinom",
292
- "dfrac",
293
- "tbinom"
294
- ];
295
- var EMPTY_NODE = { type: "empty", content: "" };
296
- var LEFT_CURLY_BRACKET = { type: "control", value: "{" };
297
- var RIGHT_CURLY_BRACKET = { type: "control", value: "}" };
298
- var LEFT_SQUARE_BRACKET = { type: "element", value: "[" };
299
- var RIGHT_SQUARE_BRACKET = { type: "element", value: "]" };
300
- var LEFT_COMMAND = { type: "command", value: "\\left" };
301
- var RIGHT_COMMAND = { type: "command", value: "\\right" };
302
- var BEGIN_COMMAND = { type: "command", value: "\\begin" };
303
- var END_COMMAND = { type: "command", value: "\\end" };
304
-
305
- class LatexParserError extends Error {
306
- constructor(message) {
307
- super(message);
308
- this.name = "LatexParserError";
309
- }
310
- }
311
- var SUB_SYMBOL = { type: "control", value: "_" };
312
- var SUP_SYMBOL = { type: "control", value: "^" };
313
-
314
- class LatexParser {
315
- space_sensitive;
316
- newline_sensitive;
317
- constructor(space_sensitive = false, newline_sensitive = true) {
318
- this.space_sensitive = space_sensitive;
319
- this.newline_sensitive = newline_sensitive;
320
- }
321
- parse(tokens) {
322
- const results = [];
323
- let pos = 0;
324
- while (pos < tokens.length) {
325
- const results2 = [];
326
- let pos2 = 0;
327
- while (pos2 < tokens.length) {
328
- const [res, newPos] = this.parseNextExpr(tokens, pos2);
329
- pos2 = newPos;
330
- if (!this.space_sensitive && res.type === "whitespace") {
331
- continue;
332
- }
333
- if (!this.newline_sensitive && res.type === "newline") {
334
- continue;
335
- }
336
- if (res.type === "control" && res.content === "&") {
337
- throw new LatexParserError("Unexpected & outside of an alignment");
338
- }
339
- results2.push(res);
340
- }
341
- if (results2.length === 0) {
342
- return EMPTY_NODE;
343
- } else if (results2.length === 1) {
344
- return results2[0];
345
- } else {
346
- return { type: "ordgroup", content: "", args: results2 };
347
- }
348
- }
349
- if (results.length === 0) {
350
- return EMPTY_NODE;
351
- } else if (results.length === 1) {
352
- return results[0];
353
- } else {
354
- return { type: "ordgroup", content: "", args: results };
355
- }
356
- }
357
- parseNextExpr(tokens, start) {
358
- let [base, pos] = this.parseNextExprWithoutSupSub(tokens, start);
359
- let sub = null;
360
- let sup = null;
361
- let num_prime = 0;
362
- num_prime += eat_primes(tokens, pos);
363
- pos += num_prime;
364
- if (pos < tokens.length && token_eq(tokens[pos], SUB_SYMBOL)) {
365
- [sub, pos] = this.parseNextExprWithoutSupSub(tokens, pos + 1);
366
- num_prime += eat_primes(tokens, pos);
367
- pos += num_prime;
368
- if (pos < tokens.length && token_eq(tokens[pos], SUP_SYMBOL)) {
369
- [sup, pos] = this.parseNextExprWithoutSupSub(tokens, pos + 1);
370
- if (eat_primes(tokens, pos) > 0) {
371
- throw new LatexParserError("Double superscript");
372
- }
373
- }
374
- } else if (pos < tokens.length && token_eq(tokens[pos], SUP_SYMBOL)) {
375
- [sup, pos] = this.parseNextExprWithoutSupSub(tokens, pos + 1);
376
- if (eat_primes(tokens, pos) > 0) {
377
- throw new LatexParserError("Double superscript");
378
- }
379
- if (pos < tokens.length && token_eq(tokens[pos], SUB_SYMBOL)) {
380
- [sub, pos] = this.parseNextExprWithoutSupSub(tokens, pos + 1);
381
- if (eat_primes(tokens, pos) > 0) {
382
- throw new LatexParserError("Double superscript");
383
- }
384
- }
385
- }
386
- if (sub !== null || sup !== null || num_prime > 0) {
387
- const res = { base };
388
- if (sub) {
389
- res.sub = sub;
390
- }
391
- if (num_prime > 0) {
392
- res.sup = { type: "ordgroup", content: "", args: [] };
393
- for (let i = 0;i < num_prime; i++) {
394
- res.sup.args.push({ type: "symbol", content: "\\prime" });
395
- }
396
- if (sup) {
397
- res.sup.args.push(sup);
398
- }
399
- if (res.sup.args.length === 1) {
400
- res.sup = res.sup.args[0];
401
- }
402
- } else if (sup) {
403
- res.sup = sup;
404
- }
405
- return [{ type: "supsub", content: "", data: res }, pos];
406
- } else {
407
- return [base, pos];
408
- }
409
- }
410
- parseNextExprWithoutSupSub(tokens, start) {
411
- const firstToken = tokens[start];
412
- const tokenType = firstToken.type;
413
- switch (tokenType) {
414
- case "element":
415
- case "text":
416
- case "comment":
417
- case "whitespace":
418
- case "newline":
419
- return [{ type: tokenType, content: firstToken.value }, start + 1];
420
- case "command":
421
- if (token_eq(firstToken, BEGIN_COMMAND)) {
422
- return this.parseBeginEndExpr(tokens, start);
423
- } else if (token_eq(firstToken, LEFT_COMMAND)) {
424
- return this.parseLeftRightExpr(tokens, start);
425
- } else {
426
- return this.parseCommandExpr(tokens, start);
427
- }
428
- case "control":
429
- const controlChar = firstToken.value;
430
- switch (controlChar) {
431
- case "{":
432
- const posClosingBracket = find_closing_curly_bracket(tokens, start);
433
- const exprInside = tokens.slice(start + 1, posClosingBracket);
434
- return [this.parse(exprInside), posClosingBracket + 1];
435
- case "}":
436
- throw new LatexParserError("Unmatched '}'");
437
- case "\\\\":
438
- return [{ type: "control", content: "\\\\" }, start + 1];
439
- case "\\,":
440
- return [{ type: "control", content: "\\," }, start + 1];
441
- case "_": {
442
- let [sub, pos] = this.parseNextExpr(tokens, start + 1);
443
- let sup = undefined;
444
- if (pos < tokens.length && token_eq(tokens[pos], SUP_SYMBOL)) {
445
- [sup, pos] = this.parseNextExpr(tokens, pos + 1);
446
- }
447
- const subData = { base: EMPTY_NODE, sub, sup };
448
- return [{ type: "supsub", content: "", data: subData }, pos];
449
- }
450
- case "^": {
451
- let [sup, pos] = this.parseNextExpr(tokens, start + 1);
452
- let sub = undefined;
453
- if (pos < tokens.length && token_eq(tokens[pos], SUB_SYMBOL)) {
454
- [sub, pos] = this.parseNextExpr(tokens, pos + 1);
455
- }
456
- const supData = { base: EMPTY_NODE, sub, sup };
457
- return [{ type: "supsub", content: "", data: supData }, pos];
458
- }
459
- case "&":
460
- return [{ type: "control", content: "&" }, start + 1];
461
- default:
462
- throw new LatexParserError("Unknown control sequence");
463
- }
464
- default:
465
- throw new LatexParserError("Unknown token type");
466
- }
467
- }
468
- parseCommandExpr(tokens, start) {
469
- assert(tokens[start].type === "command");
470
- const command = tokens[start].value;
471
- let pos = start + 1;
472
- if (["left", "right", "begin", "end"].includes(command.slice(1))) {
473
- throw new LatexParserError("Unexpected command: " + command);
474
- }
475
- const paramNum = get_command_param_num(command.slice(1));
476
- if (paramNum === 0) {
477
- return [{ type: "symbol", content: command }, pos];
478
- } else if (paramNum === 1) {
479
- if (command === "\\sqrt" && pos < tokens.length && token_eq(tokens[pos], LEFT_SQUARE_BRACKET)) {
480
- const posLeftSquareBracket = pos;
481
- const posRightSquareBracket = find_closing_square_bracket(tokens, pos);
482
- const exprInside = tokens.slice(posLeftSquareBracket + 1, posRightSquareBracket);
483
- const exponent = this.parse(exprInside);
484
- const [arg12, newPos2] = this.parseNextExprWithoutSupSub(tokens, posRightSquareBracket + 1);
485
- return [{ type: "unaryFunc", content: command, args: [arg12], data: exponent }, newPos2];
486
- } else if (command === "\\text") {
487
- if (pos + 2 >= tokens.length) {
488
- throw new LatexParserError("Expecting content for \\text command");
489
- }
490
- assert(token_eq(tokens[pos], LEFT_CURLY_BRACKET));
491
- assert(tokens[pos + 1].type === "text");
492
- assert(token_eq(tokens[pos + 2], RIGHT_CURLY_BRACKET));
493
- const text = tokens[pos + 1].value;
494
- return [{ type: "text", content: text }, pos + 3];
495
- }
496
- let [arg1, newPos] = this.parseNextExprWithoutSupSub(tokens, pos);
497
- return [{ type: "unaryFunc", content: command, args: [arg1] }, newPos];
498
- } else if (paramNum === 2) {
499
- const [arg1, pos1] = this.parseNextExprWithoutSupSub(tokens, pos);
500
- const [arg2, pos2] = this.parseNextExprWithoutSupSub(tokens, pos1);
501
- return [{ type: "binaryFunc", content: command, args: [arg1, arg2] }, pos2];
502
- } else {
503
- throw new Error("Invalid number of parameters");
504
- }
505
- }
506
- parseLeftRightExpr(tokens, start) {
507
- assert(token_eq(tokens[start], LEFT_COMMAND));
508
- let pos = start + 1;
509
- pos += eat_whitespaces(tokens, pos).length;
510
- if (pos >= tokens.length) {
511
- throw new LatexParserError("Expecting delimiter after \\left");
512
- }
513
- const leftDelimiter = eat_parenthesis(tokens, pos);
514
- if (leftDelimiter === null) {
515
- throw new LatexParserError("Invalid delimiter after \\left");
516
- }
517
- pos++;
518
- const exprInsideStart = pos;
519
- const idx = find_closing_right_command(tokens, pos);
520
- if (idx === -1) {
521
- throw new LatexParserError("No matching \\right");
522
- }
523
- const exprInsideEnd = idx;
524
- pos = idx + 1;
525
- pos += eat_whitespaces(tokens, pos).length;
526
- if (pos >= tokens.length) {
527
- throw new LatexParserError("Expecting \\right after \\left");
528
- }
529
- const rightDelimiter = eat_parenthesis(tokens, pos);
530
- if (rightDelimiter === null) {
531
- throw new LatexParserError("Invalid delimiter after \\right");
532
- }
533
- pos++;
534
- const exprInside = tokens.slice(exprInsideStart, exprInsideEnd);
535
- const body = this.parse(exprInside);
536
- const args = [
537
- { type: "element", content: leftDelimiter.value },
538
- body,
539
- { type: "element", content: rightDelimiter.value }
540
- ];
541
- const res = { type: "leftright", content: "", args };
542
- return [res, pos];
543
- }
544
- parseBeginEndExpr(tokens, start) {
545
- assert(token_eq(tokens[start], BEGIN_COMMAND));
546
- let pos = start + 1;
547
- assert(token_eq(tokens[pos], LEFT_CURLY_BRACKET));
548
- assert(tokens[pos + 1].type === "text");
549
- assert(token_eq(tokens[pos + 2], RIGHT_CURLY_BRACKET));
550
- const envName = tokens[pos + 1].value;
551
- pos += 3;
552
- pos += eat_whitespaces(tokens, pos).length;
553
- const exprInsideStart = pos;
554
- const endIdx = find_closing_end_command(tokens, pos);
555
- if (endIdx === -1) {
556
- throw new LatexParserError("No matching \\end");
557
- }
558
- const exprInsideEnd = endIdx;
559
- pos = endIdx + 1;
560
- assert(token_eq(tokens[pos], LEFT_CURLY_BRACKET));
561
- assert(tokens[pos + 1].type === "text");
562
- assert(token_eq(tokens[pos + 2], RIGHT_CURLY_BRACKET));
563
- if (tokens[pos + 1].value !== envName) {
564
- throw new LatexParserError("Mismatched \\begin and \\end environments");
565
- }
566
- pos += 3;
567
- const exprInside = tokens.slice(exprInsideStart, exprInsideEnd);
568
- while (exprInside.length > 0 && ["whitespace", "newline"].includes(exprInside[exprInside.length - 1].type)) {
569
- exprInside.pop();
570
- }
571
- const body = this.parseAligned(exprInside);
572
- const res = { type: "beginend", content: envName, data: body };
573
- return [res, pos];
574
- }
575
- parseAligned(tokens) {
576
- let pos = 0;
577
- const allRows = [];
578
- let row = [];
579
- allRows.push(row);
580
- let group = { type: "ordgroup", content: "", args: [] };
581
- row.push(group);
582
- while (pos < tokens.length) {
583
- const [res, newPos] = this.parseNextExpr(tokens, pos);
584
- pos = newPos;
585
- if (res.type === "whitespace") {
586
- continue;
587
- } else if (res.type === "newline" && !this.newline_sensitive) {
588
- continue;
589
- } else if (res.type === "control" && res.content === "\\\\") {
590
- row = [];
591
- group = { type: "ordgroup", content: "", args: [] };
592
- row.push(group);
593
- allRows.push(row);
594
- } else if (res.type === "control" && res.content === "&") {
595
- group = { type: "ordgroup", content: "", args: [] };
596
- row.push(group);
597
- } else {
598
- group.args.push(res);
599
- }
600
- }
601
- return allRows;
602
- }
603
- }
604
-
605
1
  // src/map.ts
606
2
  var symbolMap = new Map([
607
3
  ["nonumber", ""],
@@ -902,7 +298,788 @@ var symbolMap = new Map([
902
298
  ["TeX", "#TeX"]
903
299
  ]);
904
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
+
905
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: ${node.type}`, node);
1081
+ }
1082
+ }
906
1083
  function convertToken(token) {
907
1084
  if (/^[a-zA-Z0-9]$/.test(token)) {
908
1085
  return token;
@@ -975,142 +1152,74 @@ class TypstWriter {
975
1152
  this.buffer += str;
976
1153
  }
977
1154
  append(node) {
978
- if (node.type === "empty" || node.type === "whitespace") {
979
- return;
980
- } else if (node.type === "ordgroup") {
981
- node.args.forEach((arg) => this.append(arg));
982
- } else if (node.type === "element") {
983
- let content = node.content;
984
- if (node.content === "," && this.insideFunctionDepth > 0) {
985
- content = "comma";
986
- }
987
- this.queue.push({ type: "symbol", content });
988
- } else if (node.type === "symbol") {
989
- this.queue.push({ type: "symbol", content: node.content });
990
- } else if (node.type === "text") {
991
- this.queue.push(node);
992
- } else if (node.type === "supsub") {
993
- let { base, sup, sub } = node.data;
994
- if (base && base.type === "unaryFunc" && base.content === "\\overbrace" && sup) {
995
- this.append({ type: "binaryFunc", content: "\\overbrace", args: [base.args[0], sup] });
996
- return;
997
- } else if (base && base.type === "unaryFunc" && base.content === "\\underbrace" && sub) {
998
- this.append({ type: "binaryFunc", content: "\\underbrace", args: [base.args[0], sub] });
999
- 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;
1000
1165
  }
1001
- if (base.type === "empty") {
1002
- this.queue.push({ type: "text", content: "" });
1003
- } 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;
1004
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;
1005
1197
  }
1006
- let trailing_space_needed = false;
1007
- const has_prime = sup && sup.type === "symbol" && sup.content === "\\prime";
1008
- if (has_prime) {
1009
- this.queue.push({ type: "atom", content: "\'" });
1010
- trailing_space_needed = false;
1011
- }
1012
- if (sub) {
1013
- this.queue.push({ type: "atom", content: "_" });
1014
- trailing_space_needed = this.appendWithBracketsIfNeeded(sub);
1015
- }
1016
- if (sup && !has_prime) {
1017
- this.queue.push({ type: "atom", content: "^" });
1018
- trailing_space_needed = this.appendWithBracketsIfNeeded(sup);
1019
- }
1020
- if (trailing_space_needed) {
1021
- this.queue.push({ type: "softSpace", content: "" });
1022
- }
1023
- } else if (node.type === "leftright") {
1024
- const [left, body, right] = node.args;
1025
- if (["[]", "()", "\\{\\}", "\\lfloor\\rfloor", "\\lceil\\rceil"].includes(left.content + right.content)) {
1026
- this.append(left);
1027
- this.append(body);
1028
- this.append(right);
1029
- return;
1030
- }
1031
- const func_symbol = { type: "symbol", content: "lr" };
1032
- this.queue.push(func_symbol);
1033
- this.insideFunctionDepth++;
1034
- this.queue.push({ type: "atom", content: "(" });
1035
- this.append(left);
1036
- this.append(body);
1037
- this.append(right);
1038
- this.queue.push({ type: "atom", content: ")" });
1039
- this.insideFunctionDepth--;
1040
- } else if (node.type === "binaryFunc") {
1041
- const func_symbol = { type: "symbol", content: node.content };
1042
- const [arg0, arg1] = node.args;
1043
- this.queue.push(func_symbol);
1044
- this.insideFunctionDepth++;
1045
- this.queue.push({ type: "atom", content: "(" });
1046
- this.append(arg0);
1047
- this.queue.push({ type: "atom", content: "," });
1048
- this.append(arg1);
1049
- this.queue.push({ type: "atom", content: ")" });
1050
- this.insideFunctionDepth--;
1051
- } else if (node.type === "unaryFunc") {
1052
- const func_symbol = { type: "symbol", content: node.content };
1053
- const arg0 = node.args[0];
1054
- if (node.content === "\\sqrt" && node.data) {
1055
- func_symbol.content = "root";
1198
+ case "binaryFunc": {
1199
+ const func_symbol = { type: "symbol", content: node.content };
1200
+ const [arg0, arg1] = node.args;
1056
1201
  this.queue.push(func_symbol);
1057
1202
  this.insideFunctionDepth++;
1058
1203
  this.queue.push({ type: "atom", content: "(" });
1059
- this.append(node.data);
1060
- this.queue.push({ type: "atom", content: "," });
1061
1204
  this.append(arg0);
1205
+ this.queue.push({ type: "atom", content: "," });
1206
+ this.append(arg1);
1062
1207
  this.queue.push({ type: "atom", content: ")" });
1063
1208
  this.insideFunctionDepth--;
1064
- return;
1065
- } else if (node.content === "\\mathbf") {
1066
- this.append({ type: "symbol", content: "upright" });
1067
- this.insideFunctionDepth++;
1068
- 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];
1069
1214
  this.queue.push(func_symbol);
1070
1215
  this.insideFunctionDepth++;
1071
1216
  this.queue.push({ type: "atom", content: "(" });
1072
1217
  this.append(arg0);
1073
1218
  this.queue.push({ type: "atom", content: ")" });
1074
1219
  this.insideFunctionDepth--;
1075
- this.queue.push({ type: "atom", content: ")" });
1076
- this.insideFunctionDepth--;
1077
- return;
1078
- } else if (node.content === "\\mathbb") {
1079
- const body = node.args[0];
1080
- if (body.type === "element" && /^[A-Z]$/.test(body.content)) {
1081
- this.queue.push({ type: "symbol", content: body.content + body.content });
1082
- return;
1083
- }
1084
- } else if (node.content === "\\operatorname") {
1085
- let body = node.args;
1086
- if (body.length === 1 && body[0].type == "ordgroup") {
1087
- body = body[0].args;
1088
- }
1089
- const text = body.reduce((buff, n) => {
1090
- buff += convertToken(n.content);
1091
- return buff;
1092
- }, "");
1093
- if (this.preferTypstIntrinsic && TYPST_INTRINSIC_SYMBOLS.includes(text)) {
1094
- this.queue.push({ type: "symbol", content: text });
1095
- } else {
1096
- this.queue.push({ type: "symbol", content: "op" });
1097
- this.queue.push({ type: "atom", content: "(" });
1098
- this.queue.push({ type: "text", content: text });
1099
- this.queue.push({ type: "atom", content: ")" });
1100
- }
1101
- return;
1220
+ break;
1102
1221
  }
1103
- this.queue.push(func_symbol);
1104
- this.insideFunctionDepth++;
1105
- this.queue.push({ type: "atom", content: "(" });
1106
- this.append(arg0);
1107
- this.queue.push({ type: "atom", content: ")" });
1108
- this.insideFunctionDepth--;
1109
- } else if (node.type === "newline") {
1110
- this.queue.push({ type: "newline", content: "\n" });
1111
- return;
1112
- } else if (node.type === "beginend") {
1113
- if (node.content.startsWith("align")) {
1222
+ case "align": {
1114
1223
  const matrix = node.data;
1115
1224
  matrix.forEach((row, i) => {
1116
1225
  row.forEach((cell, j) => {
@@ -1120,10 +1229,12 @@ class TypstWriter {
1120
1229
  this.append(cell);
1121
1230
  });
1122
1231
  if (i < matrix.length - 1) {
1123
- this.queue.push({ type: "symbol", content: "\\\\" });
1232
+ this.queue.push({ type: "symbol", content: "\\" });
1124
1233
  }
1125
1234
  });
1126
- } else {
1235
+ break;
1236
+ }
1237
+ case "matrix": {
1127
1238
  const matrix = node.data;
1128
1239
  this.queue.push({ type: "symbol", content: "mat" });
1129
1240
  this.insideFunctionDepth++;
@@ -1131,10 +1242,6 @@ class TypstWriter {
1131
1242
  this.queue.push({ type: "symbol", content: "delim: #none, " });
1132
1243
  matrix.forEach((row, i) => {
1133
1244
  row.forEach((cell, j) => {
1134
- if (cell.type === "ordgroup" && cell.args.length === 0) {
1135
- this.queue.push({ type: "atom", content: "," });
1136
- return;
1137
- }
1138
1245
  this.append(cell);
1139
1246
  if (j < row.length - 1) {
1140
1247
  this.queue.push({ type: "atom", content: "," });
@@ -1147,37 +1254,44 @@ class TypstWriter {
1147
1254
  });
1148
1255
  this.queue.push({ type: "atom", content: ")" });
1149
1256
  this.insideFunctionDepth--;
1257
+ break;
1150
1258
  }
1151
- } else if (node.type === "matrix") {
1152
- } else if (node.type === "unknownMacro") {
1153
- if (this.nonStrict) {
1154
- this.queue.push({ type: "symbol", content: node.content });
1155
- } else {
1156
- throw new TypstWriterError(`Unknown macro: ${node.content}`, node);
1157
- }
1158
- } else if (node.type === "control") {
1159
- if (node.content === "\\\\") {
1160
- this.queue.push({ type: "symbol", content: node.content });
1161
- } else if (node.content === "\\,") {
1162
- this.queue.push({ type: "symbol", content: "thin" });
1163
- } else {
1164
- 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;
1165
1266
  }
1166
- } else if (node.type === "comment") {
1167
- 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);
1168
1275
  } else {
1169
- throw new TypstWriterError(`Unimplemented node type to append: ${node.type}`, node);
1276
+ this.queue.push({
1277
+ type: "atom",
1278
+ content: "("
1279
+ });
1280
+ this.append(node);
1281
+ this.queue.push({
1282
+ type: "atom",
1283
+ content: ")"
1284
+ });
1170
1285
  }
1286
+ return is_single;
1171
1287
  }
1172
1288
  flushQueue() {
1173
1289
  this.queue.forEach((node) => {
1174
1290
  let str = "";
1175
1291
  switch (node.type) {
1176
1292
  case "atom":
1177
- str = node.content;
1178
- break;
1179
1293
  case "symbol":
1180
- str = convertToken(node.content);
1294
+ str = node.content;
1181
1295
  break;
1182
1296
  case "text":
1183
1297
  str = `"${node.content}"`;
@@ -1201,23 +1315,6 @@ class TypstWriter {
1201
1315
  });
1202
1316
  this.queue = [];
1203
1317
  }
1204
- appendWithBracketsIfNeeded(node) {
1205
- const is_single = ["symbol", "element", "unaryFunc", "binaryFunc", "leftright"].includes(node.type);
1206
- if (is_single) {
1207
- this.append(node);
1208
- } else {
1209
- this.queue.push({
1210
- type: "atom",
1211
- content: "("
1212
- });
1213
- this.append(node);
1214
- this.queue.push({
1215
- type: "atom",
1216
- content: ")"
1217
- });
1218
- }
1219
- return is_single;
1220
- }
1221
1318
  finalize() {
1222
1319
  this.flushQueue();
1223
1320
  const smartFloorPass = function(input) {
@@ -1239,7 +1336,7 @@ class TypstWriter {
1239
1336
  // src/index.ts
1240
1337
  function tex2typst(tex, options) {
1241
1338
  const opt = {
1242
- nonStrict: false,
1339
+ nonStrict: true,
1243
1340
  preferTypstIntrinsic: true,
1244
1341
  customTexMacros: {}
1245
1342
  };
@@ -1254,9 +1351,10 @@ function tex2typst(tex, options) {
1254
1351
  opt.customTexMacros = options.customTexMacros;
1255
1352
  }
1256
1353
  }
1257
- const t = parseTex(tex, opt.customTexMacros);
1354
+ const texTree = parseTex(tex, opt.customTexMacros);
1355
+ const typstTree = convertTree(texTree);
1258
1356
  const writer2 = new TypstWriter(opt.nonStrict, opt.preferTypstIntrinsic);
1259
- writer2.append(t);
1357
+ writer2.append(typstTree);
1260
1358
  return writer2.finalize();
1261
1359
  }
1262
1360
  export {