webtex-cn 0.1.0

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.
@@ -0,0 +1,228 @@
1
+ /**
2
+ * TeX Tokenizer for luatex-cn compatible files.
3
+ * Splits TeX source into a stream of tokens.
4
+ */
5
+
6
+ export const TokenType = {
7
+ COMMAND: 'COMMAND',
8
+ OPEN_BRACE: 'OPEN_BRACE',
9
+ CLOSE_BRACE: 'CLOSE_BRACE',
10
+ OPEN_BRACKET: 'OPEN_BRACKET',
11
+ CLOSE_BRACKET: 'CLOSE_BRACKET',
12
+ TEXT: 'TEXT',
13
+ NEWLINE: 'NEWLINE', // \\
14
+ COMMENT: 'COMMENT',
15
+ BEGIN: 'BEGIN',
16
+ END: 'END',
17
+ MATH: 'MATH',
18
+ PARAGRAPH_BREAK: 'PARAGRAPH_BREAK',
19
+ EOF: 'EOF',
20
+ };
21
+
22
+ function isLetter(ch) {
23
+ return /[a-zA-Z]/.test(ch);
24
+ }
25
+
26
+ function isCJK(ch) {
27
+ const code = ch.codePointAt(0);
28
+ // CJK Unified Ideographs
29
+ if (code >= 0x4E00 && code <= 0x9FFF) return true;
30
+ // CJK Extension A
31
+ if (code >= 0x3400 && code <= 0x4DBF) return true;
32
+ // CJK Compatibility Ideographs
33
+ if (code >= 0xF900 && code <= 0xFAFF) return true;
34
+ return false;
35
+ }
36
+
37
+ function isCommandChar(ch) {
38
+ return isLetter(ch) || isCJK(ch) || ch === '@' || ch === '*';
39
+ }
40
+
41
+ export class Tokenizer {
42
+ constructor(source) {
43
+ this.source = source;
44
+ this.pos = 0;
45
+ this.tokens = [];
46
+ }
47
+
48
+ peek() {
49
+ if (this.pos >= this.source.length) return null;
50
+ return this.source[this.pos];
51
+ }
52
+
53
+ advance() {
54
+ const ch = this.source[this.pos];
55
+ this.pos++;
56
+ return ch;
57
+ }
58
+
59
+ tokenize() {
60
+ while (this.pos < this.source.length) {
61
+ const ch = this.peek();
62
+
63
+ if (ch === '%') {
64
+ // Comment: skip to end of line
65
+ this.skipComment();
66
+ continue;
67
+ }
68
+
69
+ if (ch === '$') {
70
+ this.readMath();
71
+ continue;
72
+ }
73
+
74
+ if (ch === '\\') {
75
+ this.readCommand();
76
+ continue;
77
+ }
78
+
79
+ if (ch === '{') {
80
+ this.tokens.push({ type: TokenType.OPEN_BRACE, value: '{' });
81
+ this.advance();
82
+ continue;
83
+ }
84
+
85
+ if (ch === '}') {
86
+ this.tokens.push({ type: TokenType.CLOSE_BRACE, value: '}' });
87
+ this.advance();
88
+ continue;
89
+ }
90
+
91
+ if (ch === '[') {
92
+ this.tokens.push({ type: TokenType.OPEN_BRACKET, value: '[' });
93
+ this.advance();
94
+ continue;
95
+ }
96
+
97
+ if (ch === ']') {
98
+ this.tokens.push({ type: TokenType.CLOSE_BRACKET, value: ']' });
99
+ this.advance();
100
+ continue;
101
+ }
102
+
103
+ // Regular text
104
+ this.readText();
105
+ }
106
+
107
+ this.tokens.push({ type: TokenType.EOF, value: '' });
108
+ return this.tokens;
109
+ }
110
+
111
+ skipComment() {
112
+ while (this.pos < this.source.length && this.source[this.pos] !== '\n') {
113
+ this.pos++;
114
+ }
115
+ // Skip the newline too
116
+ if (this.pos < this.source.length) this.pos++;
117
+ }
118
+
119
+ readCommand() {
120
+ this.advance(); // skip '\'
121
+
122
+ if (this.pos >= this.source.length) {
123
+ this.tokens.push({ type: TokenType.TEXT, value: '\\' });
124
+ return;
125
+ }
126
+
127
+ const nextCh = this.peek();
128
+
129
+ // \\ = forced newline
130
+ if (nextCh === '\\') {
131
+ this.advance();
132
+ this.tokens.push({ type: TokenType.NEWLINE, value: '\\\\' });
133
+ return;
134
+ }
135
+
136
+ // \{ \} \[ \] \% \$ \& \# \_ \~ \^ - escaped characters
137
+ if ('{}[]%$&#_~^'.includes(nextCh)) {
138
+ this.advance();
139
+ this.tokens.push({ type: TokenType.TEXT, value: nextCh });
140
+ return;
141
+ }
142
+
143
+ // \ (backslash + space) = control space
144
+ if (nextCh === ' ' || nextCh === '\n') {
145
+ this.advance();
146
+ this.tokens.push({ type: TokenType.TEXT, value: ' ' });
147
+ return;
148
+ }
149
+
150
+ // Read command name: sequence of letters or CJK chars
151
+ if (!isCommandChar(nextCh)) {
152
+ // Single non-letter character command (like \,)
153
+ this.advance();
154
+ this.tokens.push({ type: TokenType.COMMAND, value: nextCh });
155
+ return;
156
+ }
157
+
158
+ // Determine if we're reading ASCII letters or CJK characters
159
+ const isAsciiStart = isLetter(nextCh) || nextCh === '@' || nextCh === '*';
160
+ const isCJKStart = isCJK(nextCh);
161
+
162
+ let name = '';
163
+ if (isAsciiStart) {
164
+ // ASCII command: read letters, @, *
165
+ while (this.pos < this.source.length && (isLetter(this.peek()) || this.peek() === '@' || this.peek() === '*')) {
166
+ name += this.advance();
167
+ }
168
+ // Skip trailing spaces after ASCII commands
169
+ while (this.pos < this.source.length && this.source[this.pos] === ' ') {
170
+ this.pos++;
171
+ }
172
+ } else if (isCJKStart) {
173
+ // CJK command: read CJK characters
174
+ while (this.pos < this.source.length && isCJK(this.peek())) {
175
+ name += this.advance();
176
+ }
177
+ }
178
+
179
+ // Special handling for \begin and \end
180
+ if (name === 'begin') {
181
+ this.tokens.push({ type: TokenType.BEGIN, value: 'begin' });
182
+ } else if (name === 'end') {
183
+ this.tokens.push({ type: TokenType.END, value: 'end' });
184
+ } else {
185
+ this.tokens.push({ type: TokenType.COMMAND, value: name });
186
+ }
187
+ }
188
+
189
+ readMath() {
190
+ this.advance(); // skip opening $
191
+ let content = '';
192
+ while (this.pos < this.source.length) {
193
+ const ch = this.peek();
194
+ if (ch === '$') {
195
+ this.advance(); // skip closing $
196
+ this.tokens.push({ type: TokenType.MATH, value: content });
197
+ return;
198
+ }
199
+ content += this.advance();
200
+ }
201
+ // Unclosed $: treat as text
202
+ this.tokens.push({ type: TokenType.TEXT, value: '$' + content });
203
+ }
204
+
205
+ readText() {
206
+ let text = '';
207
+ while (this.pos < this.source.length) {
208
+ const ch = this.peek();
209
+ if (ch === '\\' || ch === '{' || ch === '}' || ch === '[' || ch === ']' || ch === '%' || ch === '$') {
210
+ break;
211
+ }
212
+ text += this.advance();
213
+ }
214
+ if (text) {
215
+ // Split on blank lines (paragraph breaks): \n followed by whitespace-only line
216
+ const parts = text.split(/\n[ \t]*\n/);
217
+ for (let i = 0; i < parts.length; i++) {
218
+ if (i > 0) {
219
+ this.tokens.push({ type: TokenType.PARAGRAPH_BREAK, value: '' });
220
+ }
221
+ const collapsed = parts[i].replace(/[ \t]+/g, ' ');
222
+ if (collapsed.trim() || collapsed === ' ') {
223
+ this.tokens.push({ type: TokenType.TEXT, value: collapsed });
224
+ }
225
+ }
226
+ }
227
+ }
228
+ }