@ai-sdk-tool/parser 3.2.1 → 3.3.0
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.
- package/README.md +24 -116
- package/dist/{chunk-DCK5APVO.js → chunk-3QJVNEHE.js} +15 -12
- package/dist/chunk-3QJVNEHE.js.map +1 -0
- package/dist/chunk-DZB6Y354.js +240 -0
- package/dist/chunk-DZB6Y354.js.map +1 -0
- package/dist/chunk-IX4FJELL.js +671 -0
- package/dist/chunk-IX4FJELL.js.map +1 -0
- package/dist/chunk-TR2ARLIF.js +2216 -0
- package/dist/chunk-TR2ARLIF.js.map +1 -0
- package/dist/community.cjs +3293 -276
- package/dist/community.cjs.map +1 -1
- package/dist/community.js +4 -1
- package/dist/community.js.map +1 -1
- package/dist/index.cjs +3356 -254
- package/dist/index.cjs.map +1 -1
- package/dist/index.d.cts +1 -1
- package/dist/index.d.ts +1 -1
- package/dist/index.js +11 -1
- package/dist/rjson.cjs +697 -0
- package/dist/rjson.cjs.map +1 -0
- package/dist/rjson.d.cts +108 -0
- package/dist/rjson.d.ts +108 -0
- package/dist/rjson.js +11 -0
- package/dist/rjson.js.map +1 -0
- package/dist/rxml.cjs +2471 -0
- package/dist/rxml.cjs.map +1 -0
- package/dist/rxml.d.cts +90 -0
- package/dist/rxml.d.ts +90 -0
- package/dist/rxml.js +10 -0
- package/dist/rxml.js.map +1 -0
- package/dist/schema-coerce.cjs +266 -0
- package/dist/schema-coerce.cjs.map +1 -0
- package/dist/schema-coerce.d.cts +5 -0
- package/dist/schema-coerce.d.ts +5 -0
- package/dist/schema-coerce.js +11 -0
- package/dist/schema-coerce.js.map +1 -0
- package/package.json +45 -14
- package/dist/chunk-DCK5APVO.js.map +0 -1
package/dist/index.cjs
CHANGED
|
@@ -17,7 +17,6 @@ var __copyProps = (to, from, except, desc) => {
|
|
|
17
17
|
}
|
|
18
18
|
return to;
|
|
19
19
|
};
|
|
20
|
-
var __reExport = (target, mod, secondTarget) => (__copyProps(target, mod, "default"), secondTarget && __copyProps(secondTarget, mod, "default"));
|
|
21
20
|
var __toESM = (mod, isNodeMode, target) => (target = mod != null ? __create(__getProtoOf(mod)) : {}, __copyProps(
|
|
22
21
|
// If the importer is in node compatibility mode or this is not an ESM
|
|
23
22
|
// file that has been converted to a CommonJS file using a Babel-
|
|
@@ -52,7 +51,10 @@ __export(src_exports, {
|
|
|
52
51
|
logParsedSummary: () => logParsedSummary,
|
|
53
52
|
logRawChunk: () => logRawChunk,
|
|
54
53
|
originalToolsSchema: () => originalToolsSchema,
|
|
54
|
+
parse: () => parse,
|
|
55
|
+
stringify: () => stringify,
|
|
55
56
|
toolChoiceStream: () => toolChoiceStream,
|
|
57
|
+
transform: () => transform,
|
|
56
58
|
transformParams: () => transformParams,
|
|
57
59
|
wrapGenerate: () => wrapGenerate,
|
|
58
60
|
wrapStream: () => wrapStream,
|
|
@@ -62,10 +64,671 @@ __export(src_exports, {
|
|
|
62
64
|
yamlToolMiddleware: () => yamlToolMiddleware
|
|
63
65
|
});
|
|
64
66
|
module.exports = __toCommonJS(src_exports);
|
|
65
|
-
__reExport(src_exports, require("@ai-sdk-tool/rjson"), module.exports);
|
|
66
67
|
|
|
67
|
-
// src/
|
|
68
|
-
var
|
|
68
|
+
// src/rjson/index.ts
|
|
69
|
+
var WHITESPACE_TEST_REGEX = /\s/;
|
|
70
|
+
var WHITESPACE_REGEX = /^\s+/;
|
|
71
|
+
var OBJECT_START_REGEX = /^\{/;
|
|
72
|
+
var OBJECT_END_REGEX = /^\}/;
|
|
73
|
+
var ARRAY_START_REGEX = /^\[/;
|
|
74
|
+
var ARRAY_END_REGEX = /^\]/;
|
|
75
|
+
var COMMA_REGEX = /^,/;
|
|
76
|
+
var COLON_REGEX = /^:/;
|
|
77
|
+
var KEYWORD_REGEX = /^(?:true|false|null)/;
|
|
78
|
+
var NUMBER_REGEX = /^-?\d+(?:\.\d+)?(?:[eE][+-]?\d+)?/;
|
|
79
|
+
var STRING_DOUBLE_REGEX = /^"(?:[^"\\]|\\["bnrtf\\/]|\\u[0-9a-fA-F]{4})*"/;
|
|
80
|
+
var STRING_SINGLE_REGEX = /^'((?:[^'\\]|\\['bnrtf\\/]|\\u[0-9a-fA-F]{4})*)'/;
|
|
81
|
+
var COMMENT_SINGLE_REGEX = /^\/\/.*?(?:\r\n|\r|\n)/;
|
|
82
|
+
var COMMENT_MULTI_REGEX = /^\/\*[\s\S]*?\*\//;
|
|
83
|
+
var IDENTIFIER_REGEX = /^[$a-zA-Z0-9_\-+.*?!|&%^/#\\]+/;
|
|
84
|
+
function some(array, f) {
|
|
85
|
+
let acc = false;
|
|
86
|
+
for (let i = 0; i < array.length; i += 1) {
|
|
87
|
+
const result = f(array[i], i, array);
|
|
88
|
+
acc = result === void 0 ? false : result;
|
|
89
|
+
if (acc) {
|
|
90
|
+
return acc;
|
|
91
|
+
}
|
|
92
|
+
}
|
|
93
|
+
return acc;
|
|
94
|
+
}
|
|
95
|
+
function makeLexer(tokenSpecs) {
|
|
96
|
+
return (contents) => {
|
|
97
|
+
const tokens = [];
|
|
98
|
+
let line = 1;
|
|
99
|
+
let remainingContents = contents;
|
|
100
|
+
function findToken() {
|
|
101
|
+
const result = some(tokenSpecs, (tokenSpec) => {
|
|
102
|
+
const m = tokenSpec.re.exec(remainingContents);
|
|
103
|
+
if (m) {
|
|
104
|
+
const raw = m[0];
|
|
105
|
+
remainingContents = remainingContents.slice(raw.length);
|
|
106
|
+
return {
|
|
107
|
+
raw,
|
|
108
|
+
matched: tokenSpec.f(m)
|
|
109
|
+
// Process the match using the spec's function
|
|
110
|
+
};
|
|
111
|
+
}
|
|
112
|
+
return;
|
|
113
|
+
});
|
|
114
|
+
return result === false ? void 0 : result;
|
|
115
|
+
}
|
|
116
|
+
while (remainingContents !== "") {
|
|
117
|
+
const matched = findToken();
|
|
118
|
+
if (!matched) {
|
|
119
|
+
const err = new SyntaxError(
|
|
120
|
+
`Unexpected character: ${remainingContents[0]}; input: ${remainingContents.substr(
|
|
121
|
+
0,
|
|
122
|
+
100
|
|
123
|
+
)}`
|
|
124
|
+
);
|
|
125
|
+
err.line = line;
|
|
126
|
+
throw err;
|
|
127
|
+
}
|
|
128
|
+
const tokenWithLine = matched.matched;
|
|
129
|
+
tokenWithLine.line = line;
|
|
130
|
+
line += matched.raw.replace(/[^\n]/g, "").length;
|
|
131
|
+
tokens.push(tokenWithLine);
|
|
132
|
+
}
|
|
133
|
+
return tokens;
|
|
134
|
+
};
|
|
135
|
+
}
|
|
136
|
+
function fStringSingle(m) {
|
|
137
|
+
const content = m[1].replace(
|
|
138
|
+
/([^'\\]|\\['bnrtf\\]|\\u[0-9a-fA-F]{4})/g,
|
|
139
|
+
(mm) => {
|
|
140
|
+
if (mm === '"') {
|
|
141
|
+
return '\\"';
|
|
142
|
+
}
|
|
143
|
+
if (mm === "\\'") {
|
|
144
|
+
return "'";
|
|
145
|
+
}
|
|
146
|
+
return mm;
|
|
147
|
+
}
|
|
148
|
+
);
|
|
149
|
+
const match = `"${content}"`;
|
|
150
|
+
return {
|
|
151
|
+
type: "string",
|
|
152
|
+
match,
|
|
153
|
+
// The transformed, double-quoted string representation
|
|
154
|
+
// Use JSON.parse on the transformed string to handle escape sequences correctly
|
|
155
|
+
value: JSON.parse(match)
|
|
156
|
+
};
|
|
157
|
+
}
|
|
158
|
+
function fStringDouble(m) {
|
|
159
|
+
return {
|
|
160
|
+
type: "string",
|
|
161
|
+
match: m[0],
|
|
162
|
+
// The raw matched string (including quotes)
|
|
163
|
+
value: JSON.parse(m[0])
|
|
164
|
+
// Use JSON.parse to handle escapes and get the value
|
|
165
|
+
};
|
|
166
|
+
}
|
|
167
|
+
function fIdentifier(m) {
|
|
168
|
+
const value = m[0];
|
|
169
|
+
const match = '"' + value.replace(/\\/g, "\\\\").replace(/"/g, '\\"') + // Escape backslashes and quotes
|
|
170
|
+
'"';
|
|
171
|
+
return {
|
|
172
|
+
type: "string",
|
|
173
|
+
// Treat identifiers as strings
|
|
174
|
+
value,
|
|
175
|
+
// The original identifier name
|
|
176
|
+
match
|
|
177
|
+
// The double-quoted string representation
|
|
178
|
+
};
|
|
179
|
+
}
|
|
180
|
+
function fComment(m) {
|
|
181
|
+
const match = m[0].replace(
|
|
182
|
+
/./g,
|
|
183
|
+
(c) => WHITESPACE_TEST_REGEX.test(c) ? c : " "
|
|
184
|
+
);
|
|
185
|
+
return {
|
|
186
|
+
type: " ",
|
|
187
|
+
// Represent comments as whitespace tokens
|
|
188
|
+
match,
|
|
189
|
+
// String containing original newlines and spaces for other chars
|
|
190
|
+
value: void 0
|
|
191
|
+
// Comments don't have a semantic value
|
|
192
|
+
};
|
|
193
|
+
}
|
|
194
|
+
function fNumber(m) {
|
|
195
|
+
return {
|
|
196
|
+
type: "number",
|
|
197
|
+
match: m[0],
|
|
198
|
+
// The raw matched number string
|
|
199
|
+
value: Number.parseFloat(m[0])
|
|
200
|
+
// Convert string to number
|
|
201
|
+
};
|
|
202
|
+
}
|
|
203
|
+
function fKeyword(m) {
|
|
204
|
+
let value;
|
|
205
|
+
switch (m[0]) {
|
|
206
|
+
case "null":
|
|
207
|
+
value = null;
|
|
208
|
+
break;
|
|
209
|
+
case "true":
|
|
210
|
+
value = true;
|
|
211
|
+
break;
|
|
212
|
+
case "false":
|
|
213
|
+
value = false;
|
|
214
|
+
break;
|
|
215
|
+
default:
|
|
216
|
+
throw new Error(`Unexpected keyword: ${m[0]}`);
|
|
217
|
+
}
|
|
218
|
+
return {
|
|
219
|
+
type: "atom",
|
|
220
|
+
// Use 'atom' type for these literals
|
|
221
|
+
match: m[0],
|
|
222
|
+
// The raw matched keyword
|
|
223
|
+
value
|
|
224
|
+
// The corresponding JavaScript value
|
|
225
|
+
};
|
|
226
|
+
}
|
|
227
|
+
function makeTokenSpecs(relaxed) {
|
|
228
|
+
function f(type) {
|
|
229
|
+
return (m) => {
|
|
230
|
+
return { type, match: m[0], value: void 0 };
|
|
231
|
+
};
|
|
232
|
+
}
|
|
233
|
+
let tokenSpecs = [
|
|
234
|
+
{ re: WHITESPACE_REGEX, f: f(" ") },
|
|
235
|
+
// Whitespace
|
|
236
|
+
{ re: OBJECT_START_REGEX, f: f("{") },
|
|
237
|
+
// Object start
|
|
238
|
+
{ re: OBJECT_END_REGEX, f: f("}") },
|
|
239
|
+
// Object end
|
|
240
|
+
{ re: ARRAY_START_REGEX, f: f("[") },
|
|
241
|
+
// Array start
|
|
242
|
+
{ re: ARRAY_END_REGEX, f: f("]") },
|
|
243
|
+
// Array end
|
|
244
|
+
{ re: COMMA_REGEX, f: f(",") },
|
|
245
|
+
// Comma separator
|
|
246
|
+
{ re: COLON_REGEX, f: f(":") },
|
|
247
|
+
// Key-value separator
|
|
248
|
+
{ re: KEYWORD_REGEX, f: fKeyword },
|
|
249
|
+
// Keywords
|
|
250
|
+
// Number: optional sign, digits, optional decimal part, optional exponent
|
|
251
|
+
{ re: NUMBER_REGEX, f: fNumber },
|
|
252
|
+
// String: double-quoted, handles escapes
|
|
253
|
+
{ re: STRING_DOUBLE_REGEX, f: fStringDouble }
|
|
254
|
+
];
|
|
255
|
+
if (relaxed) {
|
|
256
|
+
tokenSpecs = tokenSpecs.concat([
|
|
257
|
+
// Single-quoted strings
|
|
258
|
+
{
|
|
259
|
+
re: STRING_SINGLE_REGEX,
|
|
260
|
+
f: fStringSingle
|
|
261
|
+
},
|
|
262
|
+
// Single-line comments (// ...)
|
|
263
|
+
{ re: COMMENT_SINGLE_REGEX, f: fComment },
|
|
264
|
+
// Multi-line comments (/* ... */)
|
|
265
|
+
{ re: COMMENT_MULTI_REGEX, f: fComment },
|
|
266
|
+
// Unquoted identifiers (treated as strings)
|
|
267
|
+
// Allows letters, numbers, _, -, +, ., *, ?, !, |, &, %, ^, /, #, \
|
|
268
|
+
{ re: IDENTIFIER_REGEX, f: fIdentifier }
|
|
269
|
+
// Note: The order matters here. Identifiers are checked after keywords/numbers.
|
|
270
|
+
]);
|
|
271
|
+
}
|
|
272
|
+
return tokenSpecs;
|
|
273
|
+
}
|
|
274
|
+
var lexer = makeLexer(makeTokenSpecs(true));
|
|
275
|
+
var strictLexer = makeLexer(makeTokenSpecs(false));
|
|
276
|
+
function previousNWSToken(tokens, index) {
|
|
277
|
+
let currentIndex = index;
|
|
278
|
+
for (; currentIndex >= 0; currentIndex -= 1) {
|
|
279
|
+
if (tokens[currentIndex].type !== " ") {
|
|
280
|
+
return currentIndex;
|
|
281
|
+
}
|
|
282
|
+
}
|
|
283
|
+
return;
|
|
284
|
+
}
|
|
285
|
+
function stripTrailingComma(tokens) {
|
|
286
|
+
const res = [];
|
|
287
|
+
tokens.forEach((token, index) => {
|
|
288
|
+
if (index > 0 && (token.type === "]" || token.type === "}")) {
|
|
289
|
+
const prevNWSTokenIndex = previousNWSToken(res, res.length - 1);
|
|
290
|
+
if (prevNWSTokenIndex !== void 0 && res[prevNWSTokenIndex].type === ",") {
|
|
291
|
+
const preCommaIndex = previousNWSToken(res, prevNWSTokenIndex - 1);
|
|
292
|
+
if (preCommaIndex !== void 0 && res[preCommaIndex].type !== "[" && res[preCommaIndex].type !== "{") {
|
|
293
|
+
res[prevNWSTokenIndex] = {
|
|
294
|
+
type: " ",
|
|
295
|
+
match: " ",
|
|
296
|
+
// Represent as a single space
|
|
297
|
+
value: void 0,
|
|
298
|
+
// Whitespace has no value
|
|
299
|
+
line: res[prevNWSTokenIndex].line
|
|
300
|
+
// Preserve original line number
|
|
301
|
+
};
|
|
302
|
+
}
|
|
303
|
+
}
|
|
304
|
+
}
|
|
305
|
+
res.push(token);
|
|
306
|
+
});
|
|
307
|
+
return res;
|
|
308
|
+
}
|
|
309
|
+
function transform(text) {
|
|
310
|
+
let tokens = lexer(text);
|
|
311
|
+
tokens = stripTrailingComma(tokens);
|
|
312
|
+
return tokens.reduce((str, token) => str + token.match, "");
|
|
313
|
+
}
|
|
314
|
+
function popToken(tokens, state) {
|
|
315
|
+
var _a, _b;
|
|
316
|
+
const token = tokens[state.pos];
|
|
317
|
+
state.pos += 1;
|
|
318
|
+
if (!token) {
|
|
319
|
+
const lastLine = tokens.length !== 0 ? (_b = (_a = tokens.at(-1)) == null ? void 0 : _a.line) != null ? _b : 1 : 1;
|
|
320
|
+
return { type: "eof", match: "", value: void 0, line: lastLine };
|
|
321
|
+
}
|
|
322
|
+
return token;
|
|
323
|
+
}
|
|
324
|
+
function strToken(token) {
|
|
325
|
+
switch (token.type) {
|
|
326
|
+
case "atom":
|
|
327
|
+
case "string":
|
|
328
|
+
case "number":
|
|
329
|
+
return `${token.type} ${token.match}`;
|
|
330
|
+
case "eof":
|
|
331
|
+
return "end-of-file";
|
|
332
|
+
default:
|
|
333
|
+
return `'${token.type}'`;
|
|
334
|
+
}
|
|
335
|
+
}
|
|
336
|
+
function skipColon(tokens, state) {
|
|
337
|
+
const colon = popToken(tokens, state);
|
|
338
|
+
if (colon.type !== ":") {
|
|
339
|
+
const message = `Unexpected token: ${strToken(colon)}, expected ':'`;
|
|
340
|
+
if (state.tolerant) {
|
|
341
|
+
state.warnings.push({
|
|
342
|
+
message,
|
|
343
|
+
line: colon.line
|
|
344
|
+
});
|
|
345
|
+
state.pos -= 1;
|
|
346
|
+
} else {
|
|
347
|
+
const err = new SyntaxError(message);
|
|
348
|
+
err.line = colon.line;
|
|
349
|
+
throw err;
|
|
350
|
+
}
|
|
351
|
+
}
|
|
352
|
+
}
|
|
353
|
+
function skipPunctuation(tokens, state, valid) {
|
|
354
|
+
const punctuation = [",", ":", "]", "}"];
|
|
355
|
+
let token = popToken(tokens, state);
|
|
356
|
+
while (true) {
|
|
357
|
+
if (valid == null ? void 0 : valid.includes(token.type)) {
|
|
358
|
+
return token;
|
|
359
|
+
}
|
|
360
|
+
if (token.type === "eof") {
|
|
361
|
+
return token;
|
|
362
|
+
}
|
|
363
|
+
if (punctuation.includes(token.type)) {
|
|
364
|
+
const message = `Unexpected token: ${strToken(
|
|
365
|
+
token
|
|
366
|
+
)}, expected '[', '{', number, string or atom`;
|
|
367
|
+
if (state.tolerant) {
|
|
368
|
+
state.warnings.push({
|
|
369
|
+
message,
|
|
370
|
+
line: token.line
|
|
371
|
+
});
|
|
372
|
+
token = popToken(tokens, state);
|
|
373
|
+
} else {
|
|
374
|
+
const err = new SyntaxError(message);
|
|
375
|
+
err.line = token.line;
|
|
376
|
+
throw err;
|
|
377
|
+
}
|
|
378
|
+
} else {
|
|
379
|
+
return token;
|
|
380
|
+
}
|
|
381
|
+
}
|
|
382
|
+
}
|
|
383
|
+
function raiseError(state, token, message) {
|
|
384
|
+
if (state.tolerant) {
|
|
385
|
+
state.warnings.push({
|
|
386
|
+
message,
|
|
387
|
+
line: token.line
|
|
388
|
+
});
|
|
389
|
+
} else {
|
|
390
|
+
const err = new SyntaxError(message);
|
|
391
|
+
err.line = token.line;
|
|
392
|
+
throw err;
|
|
393
|
+
}
|
|
394
|
+
}
|
|
395
|
+
function raiseUnexpected(state, token, expected) {
|
|
396
|
+
raiseError(
|
|
397
|
+
state,
|
|
398
|
+
token,
|
|
399
|
+
`Unexpected token: ${strToken(token)}, expected ${expected}`
|
|
400
|
+
);
|
|
401
|
+
}
|
|
402
|
+
function checkDuplicates(state, obj, token) {
|
|
403
|
+
const key = String(token.value);
|
|
404
|
+
if (!state.duplicate && Object.hasOwn(obj, key)) {
|
|
405
|
+
raiseError(state, token, `Duplicate key: ${key}`);
|
|
406
|
+
}
|
|
407
|
+
}
|
|
408
|
+
function appendPair(state, obj, key, value) {
|
|
409
|
+
const finalValue = state.reviver ? state.reviver(key, value) : value;
|
|
410
|
+
if (finalValue !== void 0) {
|
|
411
|
+
obj[key] = finalValue;
|
|
412
|
+
}
|
|
413
|
+
}
|
|
414
|
+
function parsePair(tokens, state, obj) {
|
|
415
|
+
let token = skipPunctuation(tokens, state, [":", "string", "number", "atom"]);
|
|
416
|
+
let value;
|
|
417
|
+
if (token.type !== "string") {
|
|
418
|
+
raiseUnexpected(state, token, "string key");
|
|
419
|
+
if (state.tolerant) {
|
|
420
|
+
switch (token.type) {
|
|
421
|
+
case ":":
|
|
422
|
+
token = {
|
|
423
|
+
type: "string",
|
|
424
|
+
value: "null",
|
|
425
|
+
match: '"null"',
|
|
426
|
+
line: token.line
|
|
427
|
+
};
|
|
428
|
+
state.pos -= 1;
|
|
429
|
+
break;
|
|
430
|
+
case "number":
|
|
431
|
+
// Use number as string key
|
|
432
|
+
case "atom":
|
|
433
|
+
token = {
|
|
434
|
+
type: "string",
|
|
435
|
+
value: String(token.value),
|
|
436
|
+
match: `"${token.value}"`,
|
|
437
|
+
line: token.line
|
|
438
|
+
};
|
|
439
|
+
break;
|
|
440
|
+
case "[":
|
|
441
|
+
// Assume missing key before an array
|
|
442
|
+
case "{":
|
|
443
|
+
state.pos -= 1;
|
|
444
|
+
value = parseAny(tokens, state);
|
|
445
|
+
checkDuplicates(state, obj, {
|
|
446
|
+
type: "string",
|
|
447
|
+
value: "null",
|
|
448
|
+
match: '"null"',
|
|
449
|
+
line: token.line
|
|
450
|
+
});
|
|
451
|
+
appendPair(state, obj, "null", value);
|
|
452
|
+
return;
|
|
453
|
+
// Finished parsing this "pair"
|
|
454
|
+
case "eof":
|
|
455
|
+
return;
|
|
456
|
+
// Cannot recover
|
|
457
|
+
default:
|
|
458
|
+
return;
|
|
459
|
+
}
|
|
460
|
+
} else {
|
|
461
|
+
return;
|
|
462
|
+
}
|
|
463
|
+
}
|
|
464
|
+
checkDuplicates(state, obj, token);
|
|
465
|
+
const key = String(token.value);
|
|
466
|
+
skipColon(tokens, state);
|
|
467
|
+
value = parseAny(tokens, state);
|
|
468
|
+
appendPair(state, obj, key, value);
|
|
469
|
+
}
|
|
470
|
+
function parseElement(tokens, state, arr) {
|
|
471
|
+
const key = arr.length;
|
|
472
|
+
const value = parseAny(tokens, state);
|
|
473
|
+
arr[key] = state.reviver ? state.reviver(String(key), value) : value;
|
|
474
|
+
}
|
|
475
|
+
function parseObject(tokens, state) {
|
|
476
|
+
const obj = {};
|
|
477
|
+
return parseMany(tokens, state, obj, {
|
|
478
|
+
skip: [":", "}"],
|
|
479
|
+
// Initially skip over colon or closing brace (for empty/tolerant cases)
|
|
480
|
+
elementParser: parsePair,
|
|
481
|
+
// Use parsePair to parse each key-value element
|
|
482
|
+
elementName: "string key",
|
|
483
|
+
// Expected element type for errors
|
|
484
|
+
endSymbol: "}"
|
|
485
|
+
// The closing token for an object
|
|
486
|
+
});
|
|
487
|
+
}
|
|
488
|
+
function parseArray(tokens, state) {
|
|
489
|
+
const arr = [];
|
|
490
|
+
return parseMany(tokens, state, arr, {
|
|
491
|
+
skip: ["]"],
|
|
492
|
+
// Initially skip over closing bracket (for empty/tolerant cases)
|
|
493
|
+
elementParser: parseElement,
|
|
494
|
+
// Use parseElement to parse each array item
|
|
495
|
+
elementName: "json value",
|
|
496
|
+
// Expected element type for errors
|
|
497
|
+
endSymbol: "]"
|
|
498
|
+
// The closing token for an array
|
|
499
|
+
});
|
|
500
|
+
}
|
|
501
|
+
function handleInvalidToken(token, state, opts, result) {
|
|
502
|
+
raiseUnexpected(state, token, `',' or '${opts.endSymbol}'`);
|
|
503
|
+
if (state.tolerant) {
|
|
504
|
+
if (token.type === "eof") {
|
|
505
|
+
return result;
|
|
506
|
+
}
|
|
507
|
+
state.pos -= 1;
|
|
508
|
+
return null;
|
|
509
|
+
}
|
|
510
|
+
return result;
|
|
511
|
+
}
|
|
512
|
+
function handleCommaToken(params) {
|
|
513
|
+
const { token, tokens, state, opts, result } = params;
|
|
514
|
+
const nextToken = tokens[state.pos];
|
|
515
|
+
if (state.tolerant && nextToken && nextToken.type === opts.endSymbol) {
|
|
516
|
+
raiseError(state, token, `Trailing comma before '${opts.endSymbol}'`);
|
|
517
|
+
popToken(tokens, state);
|
|
518
|
+
return result;
|
|
519
|
+
}
|
|
520
|
+
opts.elementParser(tokens, state, result);
|
|
521
|
+
return null;
|
|
522
|
+
}
|
|
523
|
+
function parseManyInitialElement(tokens, state, result, opts) {
|
|
524
|
+
const token = skipPunctuation(tokens, state, opts.skip);
|
|
525
|
+
if (token.type === "eof") {
|
|
526
|
+
raiseUnexpected(state, token, `'${opts.endSymbol}' or ${opts.elementName}`);
|
|
527
|
+
return result;
|
|
528
|
+
}
|
|
529
|
+
if (token.type === opts.endSymbol) {
|
|
530
|
+
return result;
|
|
531
|
+
}
|
|
532
|
+
state.pos -= 1;
|
|
533
|
+
opts.elementParser(tokens, state, result);
|
|
534
|
+
return;
|
|
535
|
+
}
|
|
536
|
+
function parseManyProcessToken(params) {
|
|
537
|
+
const { token, tokens, state, opts, result } = params;
|
|
538
|
+
if (token.type !== opts.endSymbol && token.type !== ",") {
|
|
539
|
+
const handledResult = handleInvalidToken(token, state, opts, result);
|
|
540
|
+
if (handledResult !== null) {
|
|
541
|
+
return handledResult;
|
|
542
|
+
}
|
|
543
|
+
}
|
|
544
|
+
if (token.type === opts.endSymbol) {
|
|
545
|
+
return result;
|
|
546
|
+
}
|
|
547
|
+
if (token.type === ",") {
|
|
548
|
+
const handledResult = handleCommaToken({
|
|
549
|
+
token,
|
|
550
|
+
tokens,
|
|
551
|
+
state,
|
|
552
|
+
opts,
|
|
553
|
+
result
|
|
554
|
+
});
|
|
555
|
+
if (handledResult !== null) {
|
|
556
|
+
return handledResult;
|
|
557
|
+
}
|
|
558
|
+
return;
|
|
559
|
+
}
|
|
560
|
+
opts.elementParser(tokens, state, result);
|
|
561
|
+
return;
|
|
562
|
+
}
|
|
563
|
+
function parseMany(tokens, state, result, opts) {
|
|
564
|
+
const initialResult = parseManyInitialElement(tokens, state, result, opts);
|
|
565
|
+
if (initialResult !== void 0) {
|
|
566
|
+
return initialResult;
|
|
567
|
+
}
|
|
568
|
+
while (true) {
|
|
569
|
+
const token = popToken(tokens, state);
|
|
570
|
+
const processedResult = parseManyProcessToken({
|
|
571
|
+
token,
|
|
572
|
+
tokens,
|
|
573
|
+
state,
|
|
574
|
+
opts,
|
|
575
|
+
result
|
|
576
|
+
});
|
|
577
|
+
if (processedResult !== void 0) {
|
|
578
|
+
return processedResult;
|
|
579
|
+
}
|
|
580
|
+
}
|
|
581
|
+
}
|
|
582
|
+
function endChecks(tokens, state, ret) {
|
|
583
|
+
if (state.pos < tokens.length) {
|
|
584
|
+
if (state.tolerant) {
|
|
585
|
+
skipPunctuation(tokens, state);
|
|
586
|
+
}
|
|
587
|
+
if (state.pos < tokens.length) {
|
|
588
|
+
raiseError(
|
|
589
|
+
state,
|
|
590
|
+
tokens[state.pos],
|
|
591
|
+
`Unexpected token: ${strToken(tokens[state.pos])}, expected end-of-input`
|
|
592
|
+
);
|
|
593
|
+
}
|
|
594
|
+
}
|
|
595
|
+
if (state.tolerant && state.warnings.length > 0) {
|
|
596
|
+
const message = state.warnings.length === 1 ? state.warnings[0].message : `${state.warnings.length} parse warnings`;
|
|
597
|
+
const err = new SyntaxError(message);
|
|
598
|
+
err.line = state.warnings[0].line;
|
|
599
|
+
err.warnings = state.warnings;
|
|
600
|
+
err.obj = ret;
|
|
601
|
+
throw err;
|
|
602
|
+
}
|
|
603
|
+
}
|
|
604
|
+
function parseAny(tokens, state, end = false) {
|
|
605
|
+
const token = skipPunctuation(tokens, state);
|
|
606
|
+
let ret;
|
|
607
|
+
if (token.type === "eof") {
|
|
608
|
+
if (end) {
|
|
609
|
+
raiseUnexpected(state, token, "json value");
|
|
610
|
+
}
|
|
611
|
+
raiseUnexpected(state, token, "json value");
|
|
612
|
+
return;
|
|
613
|
+
}
|
|
614
|
+
switch (token.type) {
|
|
615
|
+
case "{":
|
|
616
|
+
ret = parseObject(tokens, state);
|
|
617
|
+
break;
|
|
618
|
+
case "[":
|
|
619
|
+
ret = parseArray(tokens, state);
|
|
620
|
+
break;
|
|
621
|
+
case "string":
|
|
622
|
+
// String literal
|
|
623
|
+
case "number":
|
|
624
|
+
// Number literal
|
|
625
|
+
case "atom":
|
|
626
|
+
ret = token.value;
|
|
627
|
+
break;
|
|
628
|
+
default:
|
|
629
|
+
raiseUnexpected(state, token, "json value");
|
|
630
|
+
if (state.tolerant) {
|
|
631
|
+
ret = null;
|
|
632
|
+
} else {
|
|
633
|
+
return;
|
|
634
|
+
}
|
|
635
|
+
}
|
|
636
|
+
if (end) {
|
|
637
|
+
ret = state.reviver ? state.reviver("", ret) : ret;
|
|
638
|
+
endChecks(tokens, state, ret);
|
|
639
|
+
}
|
|
640
|
+
return ret;
|
|
641
|
+
}
|
|
642
|
+
function normalizeParseOptions(optsOrReviver) {
|
|
643
|
+
var _a;
|
|
644
|
+
let options = {};
|
|
645
|
+
if (typeof optsOrReviver === "function") {
|
|
646
|
+
options.reviver = optsOrReviver;
|
|
647
|
+
} else if (optsOrReviver !== null && typeof optsOrReviver === "object") {
|
|
648
|
+
options = { ...optsOrReviver };
|
|
649
|
+
} else if (optsOrReviver !== void 0) {
|
|
650
|
+
throw new TypeError(
|
|
651
|
+
"Second argument must be a reviver function or an options object."
|
|
652
|
+
);
|
|
653
|
+
}
|
|
654
|
+
if (options.relaxed === void 0) {
|
|
655
|
+
if (options.warnings === true || options.tolerant === true) {
|
|
656
|
+
options.relaxed = true;
|
|
657
|
+
} else if (options.warnings === false && options.tolerant === false) {
|
|
658
|
+
options.relaxed = false;
|
|
659
|
+
} else {
|
|
660
|
+
options.relaxed = true;
|
|
661
|
+
}
|
|
662
|
+
}
|
|
663
|
+
options.tolerant = options.tolerant || options.warnings;
|
|
664
|
+
options.duplicate = (_a = options.duplicate) != null ? _a : false;
|
|
665
|
+
return options;
|
|
666
|
+
}
|
|
667
|
+
function createParseState(options) {
|
|
668
|
+
var _a, _b;
|
|
669
|
+
return {
|
|
670
|
+
pos: 0,
|
|
671
|
+
reviver: options.reviver,
|
|
672
|
+
tolerant: (_a = options.tolerant) != null ? _a : false,
|
|
673
|
+
duplicate: (_b = options.duplicate) != null ? _b : false,
|
|
674
|
+
warnings: []
|
|
675
|
+
};
|
|
676
|
+
}
|
|
677
|
+
function parseWithCustomParser(text, options) {
|
|
678
|
+
const lexerToUse = options.relaxed ? lexer : strictLexer;
|
|
679
|
+
let tokens = lexerToUse(text);
|
|
680
|
+
if (options.relaxed) {
|
|
681
|
+
tokens = stripTrailingComma(tokens);
|
|
682
|
+
}
|
|
683
|
+
tokens = tokens.filter((token) => token.type !== " ");
|
|
684
|
+
const state = createParseState(options);
|
|
685
|
+
return parseAny(tokens, state, true);
|
|
686
|
+
}
|
|
687
|
+
function parseWithTransform(text, options) {
|
|
688
|
+
let tokens = lexer(text);
|
|
689
|
+
tokens = stripTrailingComma(tokens);
|
|
690
|
+
const newtext = tokens.reduce((str, token) => str + token.match, "");
|
|
691
|
+
return JSON.parse(
|
|
692
|
+
newtext,
|
|
693
|
+
options.reviver
|
|
694
|
+
);
|
|
695
|
+
}
|
|
696
|
+
function parse(text, optsOrReviver) {
|
|
697
|
+
const options = normalizeParseOptions(optsOrReviver);
|
|
698
|
+
if (!(options.relaxed || options.warnings || options.tolerant) && options.duplicate) {
|
|
699
|
+
return JSON.parse(
|
|
700
|
+
text,
|
|
701
|
+
options.reviver
|
|
702
|
+
);
|
|
703
|
+
}
|
|
704
|
+
if (options.warnings || options.tolerant || !options.duplicate) {
|
|
705
|
+
return parseWithCustomParser(text, options);
|
|
706
|
+
}
|
|
707
|
+
return parseWithTransform(text, options);
|
|
708
|
+
}
|
|
709
|
+
function stringifyPair(obj, key) {
|
|
710
|
+
return `${JSON.stringify(key)}:${stringify(obj[key])}`;
|
|
711
|
+
}
|
|
712
|
+
function stringify(obj) {
|
|
713
|
+
const type = typeof obj;
|
|
714
|
+
if (type === "string" || type === "number" || type === "boolean" || obj === null) {
|
|
715
|
+
return JSON.stringify(obj);
|
|
716
|
+
}
|
|
717
|
+
if (type === "undefined") {
|
|
718
|
+
return "null";
|
|
719
|
+
}
|
|
720
|
+
if (Array.isArray(obj)) {
|
|
721
|
+
const elements = obj.map(stringify).join(",");
|
|
722
|
+
return `[${elements}]`;
|
|
723
|
+
}
|
|
724
|
+
if (type === "object") {
|
|
725
|
+
const keys = Object.keys(obj);
|
|
726
|
+
keys.sort();
|
|
727
|
+
const pairs = keys.map((key) => stringifyPair(obj, key)).join(",");
|
|
728
|
+
return `{${pairs}}`;
|
|
729
|
+
}
|
|
730
|
+
return "null";
|
|
731
|
+
}
|
|
69
732
|
|
|
70
733
|
// src/core/utils/debug.ts
|
|
71
734
|
var LINE_SPLIT_REGEX = /\r?\n/;
|
|
@@ -254,7 +917,7 @@ function escapeRegExp(literal) {
|
|
|
254
917
|
function processToolCallJson(toolCallJson, fullMatch, processedElements, options) {
|
|
255
918
|
var _a, _b;
|
|
256
919
|
try {
|
|
257
|
-
const parsedToolCall =
|
|
920
|
+
const parsedToolCall = parse(toolCallJson);
|
|
258
921
|
processedElements.push({
|
|
259
922
|
type: "tool-call",
|
|
260
923
|
toolCallId: generateId(),
|
|
@@ -324,260 +987,2698 @@ function closeTextBlock(state, controller) {
|
|
|
324
987
|
state.hasEmittedTextStart = false;
|
|
325
988
|
}
|
|
326
989
|
}
|
|
327
|
-
function emitIncompleteToolCall(state, controller, toolCallStart) {
|
|
328
|
-
if (!state.currentToolCallJson) {
|
|
329
|
-
return;
|
|
990
|
+
function emitIncompleteToolCall(state, controller, toolCallStart) {
|
|
991
|
+
if (!state.currentToolCallJson) {
|
|
992
|
+
return;
|
|
993
|
+
}
|
|
994
|
+
logParseFailure({
|
|
995
|
+
phase: "stream",
|
|
996
|
+
reason: "Incomplete streaming tool call segment emitted as text",
|
|
997
|
+
snippet: `${toolCallStart}${state.currentToolCallJson}`
|
|
998
|
+
});
|
|
999
|
+
const errorId = generateId();
|
|
1000
|
+
const errorContent = `${toolCallStart}${state.currentToolCallJson}`;
|
|
1001
|
+
controller.enqueue({
|
|
1002
|
+
type: "text-start",
|
|
1003
|
+
id: errorId
|
|
1004
|
+
});
|
|
1005
|
+
controller.enqueue({
|
|
1006
|
+
type: "text-delta",
|
|
1007
|
+
id: errorId,
|
|
1008
|
+
delta: errorContent
|
|
1009
|
+
});
|
|
1010
|
+
controller.enqueue({
|
|
1011
|
+
type: "text-end",
|
|
1012
|
+
id: errorId
|
|
1013
|
+
});
|
|
1014
|
+
state.currentToolCallJson = "";
|
|
1015
|
+
}
|
|
1016
|
+
function handleFinishChunk(state, controller, toolCallStart, chunk) {
|
|
1017
|
+
if (state.buffer.length > 0) {
|
|
1018
|
+
flushBuffer(state, controller, toolCallStart);
|
|
1019
|
+
}
|
|
1020
|
+
closeTextBlock(state, controller);
|
|
1021
|
+
emitIncompleteToolCall(state, controller, toolCallStart);
|
|
1022
|
+
controller.enqueue(chunk);
|
|
1023
|
+
}
|
|
1024
|
+
function publishText(text, state, controller) {
|
|
1025
|
+
if (state.isInsideToolCall) {
|
|
1026
|
+
closeTextBlock(state, controller);
|
|
1027
|
+
state.currentToolCallJson += text;
|
|
1028
|
+
} else if (text.length > 0) {
|
|
1029
|
+
if (!state.currentTextId) {
|
|
1030
|
+
state.currentTextId = generateId();
|
|
1031
|
+
controller.enqueue({
|
|
1032
|
+
type: "text-start",
|
|
1033
|
+
id: state.currentTextId
|
|
1034
|
+
});
|
|
1035
|
+
state.hasEmittedTextStart = true;
|
|
1036
|
+
}
|
|
1037
|
+
controller.enqueue({
|
|
1038
|
+
type: "text-delta",
|
|
1039
|
+
id: state.currentTextId,
|
|
1040
|
+
delta: text
|
|
1041
|
+
});
|
|
1042
|
+
}
|
|
1043
|
+
}
|
|
1044
|
+
function emitToolCall(context) {
|
|
1045
|
+
var _a, _b;
|
|
1046
|
+
const { state, controller, toolCallStart, toolCallEnd, options } = context;
|
|
1047
|
+
try {
|
|
1048
|
+
const parsedToolCall = parse(state.currentToolCallJson);
|
|
1049
|
+
closeTextBlock(state, controller);
|
|
1050
|
+
controller.enqueue({
|
|
1051
|
+
type: "tool-call",
|
|
1052
|
+
toolCallId: generateId(),
|
|
1053
|
+
toolName: parsedToolCall.name,
|
|
1054
|
+
input: JSON.stringify((_a = parsedToolCall.arguments) != null ? _a : {})
|
|
1055
|
+
});
|
|
1056
|
+
} catch (error) {
|
|
1057
|
+
logParseFailure({
|
|
1058
|
+
phase: "stream",
|
|
1059
|
+
reason: "Failed to parse streaming tool call JSON segment",
|
|
1060
|
+
snippet: `${toolCallStart}${state.currentToolCallJson}${toolCallEnd}`,
|
|
1061
|
+
error
|
|
1062
|
+
});
|
|
1063
|
+
const errorId = generateId();
|
|
1064
|
+
const errorContent = `${toolCallStart}${state.currentToolCallJson}${toolCallEnd}`;
|
|
1065
|
+
controller.enqueue({
|
|
1066
|
+
type: "text-start",
|
|
1067
|
+
id: errorId
|
|
1068
|
+
});
|
|
1069
|
+
controller.enqueue({
|
|
1070
|
+
type: "text-delta",
|
|
1071
|
+
id: errorId,
|
|
1072
|
+
delta: errorContent
|
|
1073
|
+
});
|
|
1074
|
+
controller.enqueue({
|
|
1075
|
+
type: "text-end",
|
|
1076
|
+
id: errorId
|
|
1077
|
+
});
|
|
1078
|
+
(_b = options == null ? void 0 : options.onError) == null ? void 0 : _b.call(
|
|
1079
|
+
options,
|
|
1080
|
+
"Could not process streaming JSON tool call; emitting original text.",
|
|
1081
|
+
{
|
|
1082
|
+
toolCall: errorContent
|
|
1083
|
+
}
|
|
1084
|
+
);
|
|
1085
|
+
}
|
|
1086
|
+
}
|
|
1087
|
+
function processTagMatch(context) {
|
|
1088
|
+
const { state } = context;
|
|
1089
|
+
if (state.isInsideToolCall) {
|
|
1090
|
+
emitToolCall(context);
|
|
1091
|
+
state.currentToolCallJson = "";
|
|
1092
|
+
state.isInsideToolCall = false;
|
|
1093
|
+
} else {
|
|
1094
|
+
state.currentToolCallJson = "";
|
|
1095
|
+
state.isInsideToolCall = true;
|
|
1096
|
+
}
|
|
1097
|
+
}
|
|
1098
|
+
function processBufferTags(context) {
|
|
1099
|
+
const { state, controller, toolCallStart, toolCallEnd } = context;
|
|
1100
|
+
let startIndex = getPotentialStartIndex(
|
|
1101
|
+
state.buffer,
|
|
1102
|
+
state.isInsideToolCall ? toolCallEnd : toolCallStart
|
|
1103
|
+
);
|
|
1104
|
+
while (startIndex != null) {
|
|
1105
|
+
const tag = state.isInsideToolCall ? toolCallEnd : toolCallStart;
|
|
1106
|
+
if (startIndex + tag.length > state.buffer.length) {
|
|
1107
|
+
break;
|
|
1108
|
+
}
|
|
1109
|
+
publishText(state.buffer.slice(0, startIndex), state, controller);
|
|
1110
|
+
state.buffer = state.buffer.slice(startIndex + tag.length);
|
|
1111
|
+
processTagMatch(context);
|
|
1112
|
+
startIndex = getPotentialStartIndex(
|
|
1113
|
+
state.buffer,
|
|
1114
|
+
state.isInsideToolCall ? toolCallEnd : toolCallStart
|
|
1115
|
+
);
|
|
1116
|
+
}
|
|
1117
|
+
}
|
|
1118
|
+
function handlePartialTag(state, controller, toolCallStart) {
|
|
1119
|
+
if (state.isInsideToolCall) {
|
|
1120
|
+
return;
|
|
1121
|
+
}
|
|
1122
|
+
const potentialIndex = getPotentialStartIndex(state.buffer, toolCallStart);
|
|
1123
|
+
if (potentialIndex != null && potentialIndex + toolCallStart.length > state.buffer.length) {
|
|
1124
|
+
publishText(state.buffer.slice(0, potentialIndex), state, controller);
|
|
1125
|
+
state.buffer = state.buffer.slice(potentialIndex);
|
|
1126
|
+
} else {
|
|
1127
|
+
publishText(state.buffer, state, controller);
|
|
1128
|
+
state.buffer = "";
|
|
1129
|
+
}
|
|
1130
|
+
}
|
|
1131
|
+
var jsonProtocol = ({
|
|
1132
|
+
toolCallStart = "<tool_call>",
|
|
1133
|
+
toolCallEnd = "</tool_call>"
|
|
1134
|
+
} = {}) => ({
|
|
1135
|
+
formatTools({
|
|
1136
|
+
tools,
|
|
1137
|
+
toolSystemPromptTemplate
|
|
1138
|
+
}) {
|
|
1139
|
+
return toolSystemPromptTemplate(tools || []);
|
|
1140
|
+
},
|
|
1141
|
+
formatToolCall(toolCall) {
|
|
1142
|
+
let args = {};
|
|
1143
|
+
if (toolCall.input != null) {
|
|
1144
|
+
try {
|
|
1145
|
+
args = JSON.parse(toolCall.input);
|
|
1146
|
+
} catch (e) {
|
|
1147
|
+
args = toolCall.input;
|
|
1148
|
+
}
|
|
1149
|
+
}
|
|
1150
|
+
return `${toolCallStart}${JSON.stringify({
|
|
1151
|
+
name: toolCall.toolName,
|
|
1152
|
+
arguments: args
|
|
1153
|
+
})}${toolCallEnd}`;
|
|
1154
|
+
},
|
|
1155
|
+
parseGeneratedText({
|
|
1156
|
+
text,
|
|
1157
|
+
options
|
|
1158
|
+
}) {
|
|
1159
|
+
const startEsc = escapeRegExp(toolCallStart);
|
|
1160
|
+
const endEsc = escapeRegExp(toolCallEnd);
|
|
1161
|
+
const toolCallRegex = new RegExp(
|
|
1162
|
+
`${startEsc}([\0-\uFFFF]*?)${endEsc}`,
|
|
1163
|
+
"gs"
|
|
1164
|
+
);
|
|
1165
|
+
const processedElements = [];
|
|
1166
|
+
let currentIndex = 0;
|
|
1167
|
+
let match = toolCallRegex.exec(text);
|
|
1168
|
+
while (match !== null) {
|
|
1169
|
+
currentIndex = processMatchedToolCall({
|
|
1170
|
+
match,
|
|
1171
|
+
text,
|
|
1172
|
+
currentIndex,
|
|
1173
|
+
processedElements,
|
|
1174
|
+
options
|
|
1175
|
+
});
|
|
1176
|
+
match = toolCallRegex.exec(text);
|
|
1177
|
+
}
|
|
1178
|
+
if (currentIndex < text.length) {
|
|
1179
|
+
const remainingText = text.substring(currentIndex);
|
|
1180
|
+
addTextSegment(remainingText, processedElements);
|
|
1181
|
+
}
|
|
1182
|
+
return processedElements;
|
|
1183
|
+
},
|
|
1184
|
+
createStreamParser({
|
|
1185
|
+
options
|
|
1186
|
+
}) {
|
|
1187
|
+
const state = {
|
|
1188
|
+
isInsideToolCall: false,
|
|
1189
|
+
buffer: "",
|
|
1190
|
+
currentToolCallJson: "",
|
|
1191
|
+
currentTextId: null,
|
|
1192
|
+
hasEmittedTextStart: false
|
|
1193
|
+
};
|
|
1194
|
+
return new TransformStream({
|
|
1195
|
+
transform(chunk, controller) {
|
|
1196
|
+
var _a;
|
|
1197
|
+
if (chunk.type === "finish") {
|
|
1198
|
+
handleFinishChunk(state, controller, toolCallStart, chunk);
|
|
1199
|
+
return;
|
|
1200
|
+
}
|
|
1201
|
+
if (chunk.type !== "text-delta") {
|
|
1202
|
+
controller.enqueue(chunk);
|
|
1203
|
+
return;
|
|
1204
|
+
}
|
|
1205
|
+
const textContent = (_a = chunk.delta) != null ? _a : "";
|
|
1206
|
+
state.buffer += textContent;
|
|
1207
|
+
processBufferTags({
|
|
1208
|
+
state,
|
|
1209
|
+
controller,
|
|
1210
|
+
toolCallStart,
|
|
1211
|
+
toolCallEnd,
|
|
1212
|
+
options
|
|
1213
|
+
});
|
|
1214
|
+
handlePartialTag(state, controller, toolCallStart);
|
|
1215
|
+
}
|
|
1216
|
+
});
|
|
1217
|
+
},
|
|
1218
|
+
extractToolCallSegments({ text }) {
|
|
1219
|
+
const startEsc = escapeRegExp(toolCallStart);
|
|
1220
|
+
const endEsc = escapeRegExp(toolCallEnd);
|
|
1221
|
+
const regex = new RegExp(`${startEsc}([\0-\uFFFF]*?)${endEsc}`, "gs");
|
|
1222
|
+
const segments = [];
|
|
1223
|
+
let m = regex.exec(text);
|
|
1224
|
+
while (m != null) {
|
|
1225
|
+
segments.push(m[0]);
|
|
1226
|
+
m = regex.exec(text);
|
|
1227
|
+
}
|
|
1228
|
+
return segments;
|
|
1229
|
+
}
|
|
1230
|
+
});
|
|
1231
|
+
|
|
1232
|
+
// src/core/protocols/protocol-interface.ts
|
|
1233
|
+
function isProtocolFactory(protocol) {
|
|
1234
|
+
return typeof protocol === "function";
|
|
1235
|
+
}
|
|
1236
|
+
function isTCMProtocolFactory(protocol) {
|
|
1237
|
+
return typeof protocol === "function";
|
|
1238
|
+
}
|
|
1239
|
+
|
|
1240
|
+
// src/rxml/errors/types.ts
|
|
1241
|
+
var RXMLParseError = class extends Error {
|
|
1242
|
+
constructor(message, cause, line, column) {
|
|
1243
|
+
super(message);
|
|
1244
|
+
this.name = "RXMLParseError";
|
|
1245
|
+
this.cause = cause;
|
|
1246
|
+
this.line = line;
|
|
1247
|
+
this.column = column;
|
|
1248
|
+
}
|
|
1249
|
+
};
|
|
1250
|
+
var RXMLDuplicateStringTagError = class extends Error {
|
|
1251
|
+
constructor(message) {
|
|
1252
|
+
super(message);
|
|
1253
|
+
this.name = "RXMLDuplicateStringTagError";
|
|
1254
|
+
}
|
|
1255
|
+
};
|
|
1256
|
+
var RXMLCoercionError = class extends Error {
|
|
1257
|
+
constructor(message, cause) {
|
|
1258
|
+
super(message);
|
|
1259
|
+
this.name = "RXMLCoercionError";
|
|
1260
|
+
this.cause = cause;
|
|
1261
|
+
}
|
|
1262
|
+
};
|
|
1263
|
+
var RXMLStringifyError = class extends Error {
|
|
1264
|
+
constructor(message, cause) {
|
|
1265
|
+
super(message);
|
|
1266
|
+
this.name = "RXMLStringifyError";
|
|
1267
|
+
this.cause = cause;
|
|
1268
|
+
}
|
|
1269
|
+
};
|
|
1270
|
+
|
|
1271
|
+
// src/rxml/core/types.ts
|
|
1272
|
+
var CharCodes = {
|
|
1273
|
+
OPEN_BRACKET: "<".charCodeAt(0),
|
|
1274
|
+
CLOSE_BRACKET: ">".charCodeAt(0),
|
|
1275
|
+
MINUS: "-".charCodeAt(0),
|
|
1276
|
+
SLASH: "/".charCodeAt(0),
|
|
1277
|
+
EXCLAMATION: "!".charCodeAt(0),
|
|
1278
|
+
QUESTION: "?".charCodeAt(0),
|
|
1279
|
+
SINGLE_QUOTE: "'".charCodeAt(0),
|
|
1280
|
+
DOUBLE_QUOTE: '"'.charCodeAt(0),
|
|
1281
|
+
OPEN_CORNER_BRACKET: "[".charCodeAt(0),
|
|
1282
|
+
CLOSE_CORNER_BRACKET: "]".charCodeAt(0),
|
|
1283
|
+
SPACE: " ".charCodeAt(0),
|
|
1284
|
+
TAB: " ".charCodeAt(0),
|
|
1285
|
+
NEWLINE: "\n".charCodeAt(0),
|
|
1286
|
+
CARRIAGE_RETURN: "\r".charCodeAt(0)
|
|
1287
|
+
};
|
|
1288
|
+
var DEFAULT_NO_CHILD_NODES = [
|
|
1289
|
+
"img",
|
|
1290
|
+
"br",
|
|
1291
|
+
"input",
|
|
1292
|
+
"meta",
|
|
1293
|
+
"link",
|
|
1294
|
+
"hr",
|
|
1295
|
+
"area",
|
|
1296
|
+
"base",
|
|
1297
|
+
"col",
|
|
1298
|
+
"embed",
|
|
1299
|
+
"param",
|
|
1300
|
+
"source",
|
|
1301
|
+
"track",
|
|
1302
|
+
"wbr"
|
|
1303
|
+
];
|
|
1304
|
+
var NAME_SPACER = "\r\n >/= ";
|
|
1305
|
+
|
|
1306
|
+
// src/rxml/utils/helpers.ts
|
|
1307
|
+
var NAME_START_CHAR_REGEX = /[A-Za-z_:]/;
|
|
1308
|
+
var NAME_CHAR_REGEX = /[A-Za-z0-9_.:-]/;
|
|
1309
|
+
function isNameStartChar(ch) {
|
|
1310
|
+
return NAME_START_CHAR_REGEX.test(ch);
|
|
1311
|
+
}
|
|
1312
|
+
function isNameChar(ch) {
|
|
1313
|
+
return NAME_CHAR_REGEX.test(ch);
|
|
1314
|
+
}
|
|
1315
|
+
function skipQuoted(s, i) {
|
|
1316
|
+
const quote = s[i];
|
|
1317
|
+
let pos = i + 1;
|
|
1318
|
+
while (pos < s.length) {
|
|
1319
|
+
const ch = s[pos];
|
|
1320
|
+
if (ch === "\\") {
|
|
1321
|
+
pos += 2;
|
|
1322
|
+
continue;
|
|
1323
|
+
}
|
|
1324
|
+
if (ch === quote) {
|
|
1325
|
+
return pos + 1;
|
|
1326
|
+
}
|
|
1327
|
+
pos += 1;
|
|
1328
|
+
}
|
|
1329
|
+
return pos;
|
|
1330
|
+
}
|
|
1331
|
+
function parseName(s, pos) {
|
|
1332
|
+
const start = pos;
|
|
1333
|
+
let currentPos = pos;
|
|
1334
|
+
while (NAME_SPACER.indexOf(s[currentPos]) === -1 && s[currentPos]) {
|
|
1335
|
+
currentPos += 1;
|
|
1336
|
+
}
|
|
1337
|
+
return { name: s.slice(start, currentPos), newPos: currentPos };
|
|
1338
|
+
}
|
|
1339
|
+
function parseString(s, pos) {
|
|
1340
|
+
const startChar = s[pos];
|
|
1341
|
+
const startPos = pos + 1;
|
|
1342
|
+
const endPos = s.indexOf(startChar, startPos);
|
|
1343
|
+
if (endPos === -1) {
|
|
1344
|
+
const tagEnd = s.indexOf(">", startPos);
|
|
1345
|
+
if (tagEnd !== -1) {
|
|
1346
|
+
return { value: s.slice(startPos, tagEnd), newPos: tagEnd };
|
|
1347
|
+
}
|
|
1348
|
+
return { value: s.slice(startPos), newPos: s.length };
|
|
1349
|
+
}
|
|
1350
|
+
return { value: s.slice(startPos, endPos), newPos: endPos + 1 };
|
|
1351
|
+
}
|
|
1352
|
+
function getLineColumn(s, pos) {
|
|
1353
|
+
let line = 1;
|
|
1354
|
+
let column = 1;
|
|
1355
|
+
for (let i = 0; i < pos && i < s.length; i += 1) {
|
|
1356
|
+
if (s[i] === "\n") {
|
|
1357
|
+
line += 1;
|
|
1358
|
+
column = 1;
|
|
1359
|
+
} else {
|
|
1360
|
+
column += 1;
|
|
1361
|
+
}
|
|
1362
|
+
}
|
|
1363
|
+
return { line, column };
|
|
1364
|
+
}
|
|
1365
|
+
function escapeXml(text) {
|
|
1366
|
+
return text.replace(/&/g, "&").replace(/</g, "<").replace(/>/g, ">").replace(/"/g, """).replace(/'/g, "'");
|
|
1367
|
+
}
|
|
1368
|
+
function escapeXmlMinimalText(text) {
|
|
1369
|
+
return text.replace(/&/g, "&").replace(/</g, "<").replace(/]]>/g, "]]>");
|
|
1370
|
+
}
|
|
1371
|
+
function escapeXmlMinimalAttr(value, wrapper = '"') {
|
|
1372
|
+
let escaped = value.replace(/&/g, "&").replace(/</g, "<");
|
|
1373
|
+
if (wrapper === '"') {
|
|
1374
|
+
escaped = escaped.replace(/"/g, """);
|
|
1375
|
+
} else {
|
|
1376
|
+
escaped = escaped.replace(/'/g, "'");
|
|
1377
|
+
}
|
|
1378
|
+
return escaped;
|
|
1379
|
+
}
|
|
1380
|
+
function unescapeXml(text) {
|
|
1381
|
+
return text.replace(/</g, "<").replace(/>/g, ">").replace(/"/g, '"').replace(/'/g, "'").replace(/&/g, "&");
|
|
1382
|
+
}
|
|
1383
|
+
|
|
1384
|
+
// src/rxml/builders/stringify.ts
|
|
1385
|
+
function stringify2(rootTag, obj, options = {}) {
|
|
1386
|
+
var _a, _b, _c, _d, _e;
|
|
1387
|
+
try {
|
|
1388
|
+
const format = (_a = options.format) != null ? _a : true;
|
|
1389
|
+
const declaration = (_b = options.declaration) != null ? _b : false;
|
|
1390
|
+
const minimalEscaping = (_c = options.minimalEscaping) != null ? _c : false;
|
|
1391
|
+
const suppressEmptyNode = (_d = options.suppressEmptyNode) != null ? _d : false;
|
|
1392
|
+
const strictBooleanAttributes = (_e = options.strictBooleanAttributes) != null ? _e : false;
|
|
1393
|
+
let result = "";
|
|
1394
|
+
if (declaration) {
|
|
1395
|
+
result += '<?xml version="1.0" encoding="UTF-8"?>\n';
|
|
1396
|
+
}
|
|
1397
|
+
result += stringifyValue(rootTag, obj, {
|
|
1398
|
+
depth: 0,
|
|
1399
|
+
format,
|
|
1400
|
+
suppressEmptyNode,
|
|
1401
|
+
minimalEscaping,
|
|
1402
|
+
strictBooleanAttributes
|
|
1403
|
+
});
|
|
1404
|
+
if (result.endsWith("\n")) {
|
|
1405
|
+
return result.slice(0, -1);
|
|
1406
|
+
}
|
|
1407
|
+
return result;
|
|
1408
|
+
} catch (error) {
|
|
1409
|
+
throw new RXMLStringifyError("Failed to stringify XML", error);
|
|
1410
|
+
}
|
|
1411
|
+
}
|
|
1412
|
+
function escapeContent(content, minimalEscaping) {
|
|
1413
|
+
return minimalEscaping ? escapeXmlMinimalText(content) : escapeXml(content);
|
|
1414
|
+
}
|
|
1415
|
+
function createSelfClosingTag(tagName, indent, newline) {
|
|
1416
|
+
return `${indent}<${tagName}/>${newline}`;
|
|
1417
|
+
}
|
|
1418
|
+
function createTextElement(tagName, content, indent, newline) {
|
|
1419
|
+
return `${indent}<${tagName}>${content}</${tagName}>${newline}`;
|
|
1420
|
+
}
|
|
1421
|
+
function isPrimitive(value) {
|
|
1422
|
+
return typeof value === "string" || typeof value === "number" || typeof value === "boolean";
|
|
1423
|
+
}
|
|
1424
|
+
function stringifyPrimitive(tagName, value, context, format) {
|
|
1425
|
+
const { minimalEscaping, suppressEmptyNode } = context;
|
|
1426
|
+
const content = escapeContent(String(value), minimalEscaping);
|
|
1427
|
+
if (content === "" && suppressEmptyNode) {
|
|
1428
|
+
return "";
|
|
1429
|
+
}
|
|
1430
|
+
return createTextElement(tagName, content, format.indent, format.newline);
|
|
1431
|
+
}
|
|
1432
|
+
function stringifyArray(tagName, value, context) {
|
|
1433
|
+
let result = "";
|
|
1434
|
+
for (const item of value) {
|
|
1435
|
+
result += stringifyValue(tagName, item, context);
|
|
1436
|
+
}
|
|
1437
|
+
return result;
|
|
1438
|
+
}
|
|
1439
|
+
function stringifyValue(tagName, value, context) {
|
|
1440
|
+
const { format, suppressEmptyNode, minimalEscaping } = context;
|
|
1441
|
+
const indent = format ? " ".repeat(context.depth) : "";
|
|
1442
|
+
const newline = format ? "\n" : "";
|
|
1443
|
+
if (value === null || value === void 0) {
|
|
1444
|
+
if (suppressEmptyNode) {
|
|
1445
|
+
return "";
|
|
1446
|
+
}
|
|
1447
|
+
return createSelfClosingTag(tagName, indent, newline);
|
|
1448
|
+
}
|
|
1449
|
+
if (isPrimitive(value)) {
|
|
1450
|
+
return stringifyPrimitive(tagName, value, context, { indent, newline });
|
|
1451
|
+
}
|
|
1452
|
+
if (Array.isArray(value)) {
|
|
1453
|
+
return stringifyArray(tagName, value, context);
|
|
1454
|
+
}
|
|
1455
|
+
if (typeof value === "object") {
|
|
1456
|
+
return stringifyObject(tagName, value, context);
|
|
1457
|
+
}
|
|
1458
|
+
const content = escapeContent(String(value), minimalEscaping);
|
|
1459
|
+
if (content === "" && suppressEmptyNode) {
|
|
1460
|
+
return "";
|
|
1461
|
+
}
|
|
1462
|
+
return createTextElement(tagName, content, indent, newline);
|
|
1463
|
+
}
|
|
1464
|
+
function extractObjectParts(obj) {
|
|
1465
|
+
const attributes = {};
|
|
1466
|
+
const elements = {};
|
|
1467
|
+
let textContent;
|
|
1468
|
+
for (const [key, value] of Object.entries(obj)) {
|
|
1469
|
+
if (key.startsWith("@")) {
|
|
1470
|
+
attributes[key.substring(1)] = value;
|
|
1471
|
+
} else if (key === "#text" || key === "_text") {
|
|
1472
|
+
textContent = String(value);
|
|
1473
|
+
} else if (key === "_attributes") {
|
|
1474
|
+
if (typeof value === "object" && value !== null) {
|
|
1475
|
+
Object.assign(attributes, value);
|
|
1476
|
+
}
|
|
1477
|
+
} else {
|
|
1478
|
+
elements[key] = value;
|
|
1479
|
+
}
|
|
1480
|
+
}
|
|
1481
|
+
return { attributes, elements, textContent };
|
|
1482
|
+
}
|
|
1483
|
+
function formatAttribute(attrName, attrValue, minimalEscaping, strictBooleanAttributes) {
|
|
1484
|
+
if (attrValue === null) {
|
|
1485
|
+
return strictBooleanAttributes ? ` ${attrName}="${attrName}"` : ` ${attrName}`;
|
|
1486
|
+
}
|
|
1487
|
+
const valueStr = String(attrValue);
|
|
1488
|
+
if (valueStr.indexOf('"') === -1) {
|
|
1489
|
+
const escaped2 = minimalEscaping ? escapeXmlMinimalAttr(valueStr, '"') : escapeXml(valueStr);
|
|
1490
|
+
return ` ${attrName}="${escaped2}"`;
|
|
1491
|
+
}
|
|
1492
|
+
const escaped = minimalEscaping ? escapeXmlMinimalAttr(valueStr, "'") : escapeXml(valueStr);
|
|
1493
|
+
return ` ${attrName}='${escaped}'`;
|
|
1494
|
+
}
|
|
1495
|
+
function buildOpeningTag(tagName, attributes, context) {
|
|
1496
|
+
let openTag = `<${tagName}`;
|
|
1497
|
+
const { minimalEscaping, strictBooleanAttributes } = context;
|
|
1498
|
+
for (const [attrName, attrValue] of Object.entries(attributes)) {
|
|
1499
|
+
openTag += formatAttribute(
|
|
1500
|
+
attrName,
|
|
1501
|
+
attrValue,
|
|
1502
|
+
minimalEscaping,
|
|
1503
|
+
strictBooleanAttributes
|
|
1504
|
+
);
|
|
1505
|
+
}
|
|
1506
|
+
return openTag;
|
|
1507
|
+
}
|
|
1508
|
+
function stringifyTextOnlyContent(options) {
|
|
1509
|
+
const { tagName, textContent, openTag, format, minimalEscaping } = options;
|
|
1510
|
+
const content = escapeContent(textContent, minimalEscaping);
|
|
1511
|
+
return `${format.indent}${openTag}${content}</${tagName}>${format.newline}`;
|
|
1512
|
+
}
|
|
1513
|
+
function stringifyComplexContent(tagName, parts, context, options) {
|
|
1514
|
+
const { format, minimalEscaping, depth } = context;
|
|
1515
|
+
const { textContent, elements } = parts;
|
|
1516
|
+
const hasElements = Object.keys(elements).length > 0;
|
|
1517
|
+
let result = `${options.indent}${options.openTag}`;
|
|
1518
|
+
if (textContent) {
|
|
1519
|
+
const content = escapeContent(textContent, minimalEscaping);
|
|
1520
|
+
result += format ? `${options.newline}${options.childIndent}${content}` : content;
|
|
1521
|
+
}
|
|
1522
|
+
if (hasElements) {
|
|
1523
|
+
if (format) {
|
|
1524
|
+
result += options.newline;
|
|
1525
|
+
}
|
|
1526
|
+
for (const [elementName, elementValue] of Object.entries(elements)) {
|
|
1527
|
+
result += stringifyValue(elementName, elementValue, {
|
|
1528
|
+
...context,
|
|
1529
|
+
depth: depth + 1
|
|
1530
|
+
});
|
|
1531
|
+
}
|
|
1532
|
+
if (format) {
|
|
1533
|
+
result += options.indent;
|
|
1534
|
+
}
|
|
1535
|
+
}
|
|
1536
|
+
result += `</${tagName}>${options.newline}`;
|
|
1537
|
+
return result;
|
|
1538
|
+
}
|
|
1539
|
+
function stringifyObject(tagName, obj, context) {
|
|
1540
|
+
const { depth, format, suppressEmptyNode } = context;
|
|
1541
|
+
const indent = format ? " ".repeat(depth) : "";
|
|
1542
|
+
const newline = format ? "\n" : "";
|
|
1543
|
+
const childIndent = format ? " ".repeat(depth + 1) : "";
|
|
1544
|
+
const parts = extractObjectParts(obj);
|
|
1545
|
+
const openTag = buildOpeningTag(tagName, parts.attributes, context);
|
|
1546
|
+
const hasElements = Object.keys(parts.elements).length > 0;
|
|
1547
|
+
const hasTextContent = parts.textContent !== void 0 && parts.textContent !== "";
|
|
1548
|
+
if (!(hasElements || hasTextContent)) {
|
|
1549
|
+
if (suppressEmptyNode) {
|
|
1550
|
+
return "";
|
|
1551
|
+
}
|
|
1552
|
+
return `${indent}${openTag}/>${newline}`;
|
|
1553
|
+
}
|
|
1554
|
+
const fullOpenTag = `${openTag}>`;
|
|
1555
|
+
if (!hasElements && hasTextContent && parts.textContent) {
|
|
1556
|
+
return stringifyTextOnlyContent({
|
|
1557
|
+
tagName,
|
|
1558
|
+
textContent: parts.textContent,
|
|
1559
|
+
openTag: fullOpenTag,
|
|
1560
|
+
format: { indent, newline },
|
|
1561
|
+
minimalEscaping: context.minimalEscaping
|
|
1562
|
+
});
|
|
1563
|
+
}
|
|
1564
|
+
return stringifyComplexContent(tagName, parts, context, {
|
|
1565
|
+
indent,
|
|
1566
|
+
newline,
|
|
1567
|
+
childIndent,
|
|
1568
|
+
openTag: fullOpenTag
|
|
1569
|
+
});
|
|
1570
|
+
}
|
|
1571
|
+
|
|
1572
|
+
// src/schema-coerce/index.ts
|
|
1573
|
+
var NUMERIC_REGEX = /^-?\d+(?:\.\d+)?(?:[eE][+-]?\d+)?$/;
|
|
1574
|
+
var EMPTY_OBJECT_REGEX = /^\{\s*\}$/s;
|
|
1575
|
+
var NEWLINE_SPLIT_REGEX = /\n+/;
|
|
1576
|
+
var COMMA_SPLIT_REGEX = /,\s*/;
|
|
1577
|
+
var DIGIT_KEY_REGEX = /^\d+$/;
|
|
1578
|
+
function unwrapJsonSchema(schema) {
|
|
1579
|
+
if (!schema || typeof schema !== "object") {
|
|
1580
|
+
return schema;
|
|
1581
|
+
}
|
|
1582
|
+
const s = schema;
|
|
1583
|
+
if (s.jsonSchema && typeof s.jsonSchema === "object") {
|
|
1584
|
+
return unwrapJsonSchema(s.jsonSchema);
|
|
1585
|
+
}
|
|
1586
|
+
return schema;
|
|
1587
|
+
}
|
|
1588
|
+
function getSchemaType(schema) {
|
|
1589
|
+
const unwrapped = unwrapJsonSchema(schema);
|
|
1590
|
+
if (!unwrapped || typeof unwrapped !== "object") {
|
|
1591
|
+
return;
|
|
1592
|
+
}
|
|
1593
|
+
const t = unwrapped.type;
|
|
1594
|
+
if (typeof t === "string") {
|
|
1595
|
+
return t;
|
|
1596
|
+
}
|
|
1597
|
+
if (Array.isArray(t)) {
|
|
1598
|
+
const preferred = [
|
|
1599
|
+
"object",
|
|
1600
|
+
"array",
|
|
1601
|
+
"boolean",
|
|
1602
|
+
"number",
|
|
1603
|
+
"integer",
|
|
1604
|
+
"string"
|
|
1605
|
+
];
|
|
1606
|
+
for (const p of preferred) {
|
|
1607
|
+
if (t.includes(p)) {
|
|
1608
|
+
return p;
|
|
1609
|
+
}
|
|
1610
|
+
}
|
|
1611
|
+
}
|
|
1612
|
+
const s = unwrapped;
|
|
1613
|
+
if (s && typeof s === "object" && (s.properties || s.additionalProperties)) {
|
|
1614
|
+
return "object";
|
|
1615
|
+
}
|
|
1616
|
+
if (s && typeof s === "object" && (s.items || s.prefixItems)) {
|
|
1617
|
+
return "array";
|
|
1618
|
+
}
|
|
1619
|
+
return;
|
|
1620
|
+
}
|
|
1621
|
+
function coerceStringWithoutSchema(value) {
|
|
1622
|
+
const s = value.trim();
|
|
1623
|
+
const lower = s.toLowerCase();
|
|
1624
|
+
if (lower === "true") {
|
|
1625
|
+
return true;
|
|
1626
|
+
}
|
|
1627
|
+
if (lower === "false") {
|
|
1628
|
+
return false;
|
|
1629
|
+
}
|
|
1630
|
+
if (NUMERIC_REGEX.test(s)) {
|
|
1631
|
+
const num = Number(s);
|
|
1632
|
+
if (Number.isFinite(num)) {
|
|
1633
|
+
return num;
|
|
1634
|
+
}
|
|
1635
|
+
}
|
|
1636
|
+
if (s.startsWith("{") && s.endsWith("}") || s.startsWith("[") && s.endsWith("]")) {
|
|
1637
|
+
try {
|
|
1638
|
+
const parsed = JSON.parse(s);
|
|
1639
|
+
return coerceBySchema(parsed, void 0);
|
|
1640
|
+
} catch (e) {
|
|
1641
|
+
}
|
|
1642
|
+
}
|
|
1643
|
+
return value;
|
|
1644
|
+
}
|
|
1645
|
+
function coerceStringToObject(s, unwrapped) {
|
|
1646
|
+
try {
|
|
1647
|
+
let normalized = s.replace(/'/g, '"');
|
|
1648
|
+
normalized = normalized.replace(EMPTY_OBJECT_REGEX, "{}");
|
|
1649
|
+
const obj = JSON.parse(normalized);
|
|
1650
|
+
if (obj && typeof obj === "object" && !Array.isArray(obj)) {
|
|
1651
|
+
const props = unwrapped.properties;
|
|
1652
|
+
const out = {};
|
|
1653
|
+
for (const [k, v] of Object.entries(obj)) {
|
|
1654
|
+
const propSchema = props ? props[k] : void 0;
|
|
1655
|
+
out[k] = typeof propSchema === "boolean" ? v : coerceBySchema(v, propSchema);
|
|
1656
|
+
}
|
|
1657
|
+
return out;
|
|
1658
|
+
}
|
|
1659
|
+
} catch (e) {
|
|
1660
|
+
}
|
|
1661
|
+
return null;
|
|
1662
|
+
}
|
|
1663
|
+
function coerceStringToArray(s, unwrapped) {
|
|
1664
|
+
const prefixItems = Array.isArray(unwrapped.prefixItems) ? unwrapped.prefixItems : void 0;
|
|
1665
|
+
const itemsSchema = unwrapped.items;
|
|
1666
|
+
try {
|
|
1667
|
+
const normalized = s.replace(/'/g, '"');
|
|
1668
|
+
const arr = JSON.parse(normalized);
|
|
1669
|
+
if (Array.isArray(arr)) {
|
|
1670
|
+
if (prefixItems && arr.length === prefixItems.length) {
|
|
1671
|
+
return arr.map((v, i) => coerceBySchema(v, prefixItems[i]));
|
|
1672
|
+
}
|
|
1673
|
+
return arr.map((v) => coerceBySchema(v, itemsSchema));
|
|
1674
|
+
}
|
|
1675
|
+
} catch (e) {
|
|
1676
|
+
const csv = s.includes("\n") ? s.split(NEWLINE_SPLIT_REGEX) : s.split(COMMA_SPLIT_REGEX);
|
|
1677
|
+
const trimmed = csv.map((x) => x.trim()).filter((x) => x.length > 0);
|
|
1678
|
+
if (prefixItems && trimmed.length === prefixItems.length) {
|
|
1679
|
+
return trimmed.map((x, i) => coerceBySchema(x, prefixItems[i]));
|
|
1680
|
+
}
|
|
1681
|
+
return trimmed.map((x) => coerceBySchema(x, itemsSchema));
|
|
1682
|
+
}
|
|
1683
|
+
return null;
|
|
1684
|
+
}
|
|
1685
|
+
function coerceObjectToObject(value, unwrapped) {
|
|
1686
|
+
const out = {};
|
|
1687
|
+
const props = unwrapped.properties;
|
|
1688
|
+
for (const [k, v] of Object.entries(value)) {
|
|
1689
|
+
const propSchema = props ? props[k] : void 0;
|
|
1690
|
+
out[k] = typeof propSchema === "boolean" ? v : coerceBySchema(v, propSchema);
|
|
1691
|
+
}
|
|
1692
|
+
return out;
|
|
1693
|
+
}
|
|
1694
|
+
function coerceArrayToArray(value, prefixItems, itemsSchema) {
|
|
1695
|
+
if (prefixItems && value.length === prefixItems.length) {
|
|
1696
|
+
return value.map((v, i) => coerceBySchema(v, prefixItems[i]));
|
|
1697
|
+
}
|
|
1698
|
+
return value.map((v) => coerceBySchema(v, itemsSchema));
|
|
1699
|
+
}
|
|
1700
|
+
function coerceObjectToArray(maybe, prefixItems, itemsSchema) {
|
|
1701
|
+
if (Object.hasOwn(maybe, "item")) {
|
|
1702
|
+
const items = maybe.item;
|
|
1703
|
+
const arr = Array.isArray(items) ? items : [items];
|
|
1704
|
+
return coerceArrayToArray(arr, prefixItems, itemsSchema);
|
|
1705
|
+
}
|
|
1706
|
+
const keys = Object.keys(maybe);
|
|
1707
|
+
if (keys.length === 1) {
|
|
1708
|
+
const singleValue = maybe[keys[0]];
|
|
1709
|
+
if (Array.isArray(singleValue)) {
|
|
1710
|
+
return singleValue.map((v) => coerceBySchema(v, itemsSchema));
|
|
1711
|
+
}
|
|
1712
|
+
}
|
|
1713
|
+
if (keys.length > 0 && keys.every((k) => DIGIT_KEY_REGEX.test(k))) {
|
|
1714
|
+
const arr = keys.sort((a, b) => Number(a) - Number(b)).map((k) => maybe[k]);
|
|
1715
|
+
return coerceArrayToArray(arr, prefixItems, itemsSchema);
|
|
1716
|
+
}
|
|
1717
|
+
return null;
|
|
1718
|
+
}
|
|
1719
|
+
function coercePrimitiveToArray(value, prefixItems, itemsSchema) {
|
|
1720
|
+
if (prefixItems && prefixItems.length > 0) {
|
|
1721
|
+
return [coerceBySchema(value, prefixItems[0])];
|
|
1722
|
+
}
|
|
1723
|
+
return [coerceBySchema(value, itemsSchema)];
|
|
1724
|
+
}
|
|
1725
|
+
function coerceStringToPrimitive(s, schemaType) {
|
|
1726
|
+
if (schemaType === "boolean") {
|
|
1727
|
+
const lower = s.toLowerCase();
|
|
1728
|
+
if (lower === "true") {
|
|
1729
|
+
return true;
|
|
1730
|
+
}
|
|
1731
|
+
if (lower === "false") {
|
|
1732
|
+
return false;
|
|
1733
|
+
}
|
|
1734
|
+
}
|
|
1735
|
+
if ((schemaType === "number" || schemaType === "integer") && NUMERIC_REGEX.test(s)) {
|
|
1736
|
+
const num = Number(s);
|
|
1737
|
+
if (Number.isFinite(num)) {
|
|
1738
|
+
return num;
|
|
1739
|
+
}
|
|
1740
|
+
}
|
|
1741
|
+
return null;
|
|
1742
|
+
}
|
|
1743
|
+
function coerceStringValue(value, schemaType, u) {
|
|
1744
|
+
const s = value.trim();
|
|
1745
|
+
if (schemaType === "object") {
|
|
1746
|
+
const result = coerceStringToObject(s, u);
|
|
1747
|
+
if (result !== null) {
|
|
1748
|
+
return result;
|
|
1749
|
+
}
|
|
1750
|
+
}
|
|
1751
|
+
if (schemaType === "array") {
|
|
1752
|
+
const result = coerceStringToArray(s, u);
|
|
1753
|
+
if (result !== null) {
|
|
1754
|
+
return result;
|
|
1755
|
+
}
|
|
1756
|
+
}
|
|
1757
|
+
const primitiveResult = coerceStringToPrimitive(s, schemaType);
|
|
1758
|
+
if (primitiveResult !== null) {
|
|
1759
|
+
return primitiveResult;
|
|
1760
|
+
}
|
|
1761
|
+
return value;
|
|
1762
|
+
}
|
|
1763
|
+
function coerceArrayValue(value, prefixItems, itemsSchema) {
|
|
1764
|
+
if (Array.isArray(value)) {
|
|
1765
|
+
return coerceArrayToArray(value, prefixItems, itemsSchema);
|
|
1766
|
+
}
|
|
1767
|
+
if (value && typeof value === "object") {
|
|
1768
|
+
const result = coerceObjectToArray(
|
|
1769
|
+
value,
|
|
1770
|
+
prefixItems,
|
|
1771
|
+
itemsSchema
|
|
1772
|
+
);
|
|
1773
|
+
if (result !== null) {
|
|
1774
|
+
return result;
|
|
1775
|
+
}
|
|
1776
|
+
}
|
|
1777
|
+
if (value == null || typeof value === "string" || typeof value === "number" || typeof value === "boolean") {
|
|
1778
|
+
return coercePrimitiveToArray(value, prefixItems, itemsSchema);
|
|
1779
|
+
}
|
|
1780
|
+
return value;
|
|
1781
|
+
}
|
|
1782
|
+
function coerceBySchema(value, schema) {
|
|
1783
|
+
const unwrapped = unwrapJsonSchema(schema);
|
|
1784
|
+
if (!unwrapped || typeof unwrapped !== "object") {
|
|
1785
|
+
if (typeof value === "string") {
|
|
1786
|
+
return coerceStringWithoutSchema(value);
|
|
1787
|
+
}
|
|
1788
|
+
return value;
|
|
1789
|
+
}
|
|
1790
|
+
const schemaType = getSchemaType(unwrapped);
|
|
1791
|
+
const u = unwrapped;
|
|
1792
|
+
if (typeof value === "string") {
|
|
1793
|
+
return coerceStringValue(value, schemaType, u);
|
|
1794
|
+
}
|
|
1795
|
+
if (schemaType === "object" && value && typeof value === "object" && !Array.isArray(value)) {
|
|
1796
|
+
return coerceObjectToObject(value, u);
|
|
1797
|
+
}
|
|
1798
|
+
if (schemaType === "array") {
|
|
1799
|
+
const prefixItems = Array.isArray(u.prefixItems) ? u.prefixItems : void 0;
|
|
1800
|
+
const itemsSchema = u.items;
|
|
1801
|
+
return coerceArrayValue(value, prefixItems, itemsSchema);
|
|
1802
|
+
}
|
|
1803
|
+
return value;
|
|
1804
|
+
}
|
|
1805
|
+
|
|
1806
|
+
// src/rxml/schema/coercion.ts
|
|
1807
|
+
function getPropertySchema(toolSchema, key) {
|
|
1808
|
+
const unwrapped = unwrapJsonSchema(toolSchema);
|
|
1809
|
+
if (!unwrapped || typeof unwrapped !== "object") {
|
|
1810
|
+
return;
|
|
1811
|
+
}
|
|
1812
|
+
const u = unwrapped;
|
|
1813
|
+
const props = u.properties;
|
|
1814
|
+
if (props && Object.hasOwn(props, key)) {
|
|
1815
|
+
return props[key];
|
|
1816
|
+
}
|
|
1817
|
+
return;
|
|
1818
|
+
}
|
|
1819
|
+
function getNodeValue(children, schema, tagName, textNodeName) {
|
|
1820
|
+
if (children.length === 0) {
|
|
1821
|
+
return "";
|
|
1822
|
+
}
|
|
1823
|
+
if (children.length === 1 && typeof children[0] === "string") {
|
|
1824
|
+
return children[0];
|
|
1825
|
+
}
|
|
1826
|
+
return processComplexContent(
|
|
1827
|
+
children,
|
|
1828
|
+
getPropertySchema(schema, tagName),
|
|
1829
|
+
textNodeName
|
|
1830
|
+
);
|
|
1831
|
+
}
|
|
1832
|
+
function addAttributesToValue(value, attributes, textNodeName) {
|
|
1833
|
+
if (Object.keys(attributes).length === 0) {
|
|
1834
|
+
return value;
|
|
1835
|
+
}
|
|
1836
|
+
if (typeof value === "string") {
|
|
1837
|
+
const valueResult = { [textNodeName]: value };
|
|
1838
|
+
for (const [attrName, attrValue] of Object.entries(attributes)) {
|
|
1839
|
+
valueResult[`@_${attrName}`] = attrValue;
|
|
1840
|
+
}
|
|
1841
|
+
return valueResult;
|
|
1842
|
+
}
|
|
1843
|
+
if (value && typeof value === "object" && !Array.isArray(value)) {
|
|
1844
|
+
for (const [attrName, attrValue] of Object.entries(attributes)) {
|
|
1845
|
+
value[`@_${attrName}`] = attrValue;
|
|
1846
|
+
}
|
|
1847
|
+
}
|
|
1848
|
+
return value;
|
|
1849
|
+
}
|
|
1850
|
+
function addToResult(result, tagName, value) {
|
|
1851
|
+
if (result[tagName]) {
|
|
1852
|
+
if (!Array.isArray(result[tagName])) {
|
|
1853
|
+
result[tagName] = [result[tagName]];
|
|
1854
|
+
}
|
|
1855
|
+
result[tagName].push(value);
|
|
1856
|
+
} else {
|
|
1857
|
+
result[tagName] = value;
|
|
1858
|
+
}
|
|
1859
|
+
}
|
|
1860
|
+
function domToObject(nodes, schema, textNodeName = "#text") {
|
|
1861
|
+
const result = {};
|
|
1862
|
+
for (const node of nodes) {
|
|
1863
|
+
if (typeof node === "string") {
|
|
1864
|
+
continue;
|
|
1865
|
+
}
|
|
1866
|
+
const { tagName, children, attributes } = node;
|
|
1867
|
+
let value = getNodeValue(children, schema, tagName, textNodeName);
|
|
1868
|
+
value = addAttributesToValue(value, attributes, textNodeName);
|
|
1869
|
+
addToResult(result, tagName, value);
|
|
1870
|
+
}
|
|
1871
|
+
return result;
|
|
1872
|
+
}
|
|
1873
|
+
function processChildElement(child, schema, textNodeName) {
|
|
1874
|
+
let childValue;
|
|
1875
|
+
if (child.children.length === 0) {
|
|
1876
|
+
childValue = "";
|
|
1877
|
+
} else if (child.children.length === 1 && typeof child.children[0] === "string") {
|
|
1878
|
+
childValue = child.children[0];
|
|
1879
|
+
} else {
|
|
1880
|
+
childValue = processComplexContent(
|
|
1881
|
+
child.children,
|
|
1882
|
+
getPropertySchema(schema, child.tagName),
|
|
1883
|
+
textNodeName
|
|
1884
|
+
);
|
|
1885
|
+
}
|
|
1886
|
+
return addAttributesToValue(childValue, child.attributes, textNodeName);
|
|
1887
|
+
}
|
|
1888
|
+
function combineContent(textContent, elements, textNodeName) {
|
|
1889
|
+
const hasText = textContent.length > 0;
|
|
1890
|
+
const hasElements = Object.keys(elements).length > 0;
|
|
1891
|
+
if (hasText && hasElements) {
|
|
1892
|
+
return {
|
|
1893
|
+
[textNodeName]: textContent.join("").trim(),
|
|
1894
|
+
...elements
|
|
1895
|
+
};
|
|
1896
|
+
}
|
|
1897
|
+
if (hasText) {
|
|
1898
|
+
return textContent.join("").trim();
|
|
1899
|
+
}
|
|
1900
|
+
if (hasElements) {
|
|
1901
|
+
return elements;
|
|
1902
|
+
}
|
|
1903
|
+
return "";
|
|
1904
|
+
}
|
|
1905
|
+
function processComplexContent(children, schema, textNodeName) {
|
|
1906
|
+
const textContent = [];
|
|
1907
|
+
const elements = {};
|
|
1908
|
+
for (const child of children) {
|
|
1909
|
+
if (typeof child === "string") {
|
|
1910
|
+
textContent.push(child);
|
|
1911
|
+
} else {
|
|
1912
|
+
const childValue = processChildElement(child, schema, textNodeName);
|
|
1913
|
+
addToResult(elements, child.tagName, childValue);
|
|
1914
|
+
}
|
|
1915
|
+
}
|
|
1916
|
+
return combineContent(textContent, elements, textNodeName);
|
|
1917
|
+
}
|
|
1918
|
+
function coerceDomBySchema(domObject, schema) {
|
|
1919
|
+
try {
|
|
1920
|
+
return coerceBySchema(domObject, schema);
|
|
1921
|
+
} catch (error) {
|
|
1922
|
+
throw new RXMLCoercionError("Failed to coerce DOM object by schema", error);
|
|
1923
|
+
}
|
|
1924
|
+
}
|
|
1925
|
+
function visitObjectProperties(props, collected, visit) {
|
|
1926
|
+
for (const [key, propSchema] of Object.entries(props)) {
|
|
1927
|
+
const t = getSchemaType(propSchema);
|
|
1928
|
+
if (t === "string") {
|
|
1929
|
+
collected.add(key);
|
|
1930
|
+
} else if (t === "object" || t === "array") {
|
|
1931
|
+
visit(propSchema);
|
|
1932
|
+
}
|
|
1933
|
+
}
|
|
1934
|
+
}
|
|
1935
|
+
function visitArrayItems(u, visit) {
|
|
1936
|
+
const items = u.items;
|
|
1937
|
+
if (items) {
|
|
1938
|
+
visit(items);
|
|
1939
|
+
}
|
|
1940
|
+
const prefix = u.prefixItems;
|
|
1941
|
+
if (Array.isArray(prefix)) {
|
|
1942
|
+
for (const item of prefix) {
|
|
1943
|
+
visit(item);
|
|
1944
|
+
}
|
|
1945
|
+
}
|
|
1946
|
+
}
|
|
1947
|
+
function getStringTypedProperties(schema) {
|
|
1948
|
+
const collected = /* @__PURE__ */ new Set();
|
|
1949
|
+
const visit = (s) => {
|
|
1950
|
+
const unwrapped = unwrapJsonSchema(s);
|
|
1951
|
+
if (!unwrapped || typeof unwrapped !== "object") {
|
|
1952
|
+
return;
|
|
1953
|
+
}
|
|
1954
|
+
const u = unwrapped;
|
|
1955
|
+
const type = getSchemaType(unwrapped);
|
|
1956
|
+
if (type === "object") {
|
|
1957
|
+
const props = u.properties;
|
|
1958
|
+
if (props && typeof props === "object") {
|
|
1959
|
+
visitObjectProperties(props, collected, visit);
|
|
1960
|
+
}
|
|
1961
|
+
} else if (type === "array") {
|
|
1962
|
+
visitArrayItems(u, visit);
|
|
1963
|
+
}
|
|
1964
|
+
};
|
|
1965
|
+
visit(schema);
|
|
1966
|
+
return collected;
|
|
1967
|
+
}
|
|
1968
|
+
function processArrayContent(value, schema, textNodeName) {
|
|
1969
|
+
if (!Array.isArray(value)) {
|
|
1970
|
+
return value;
|
|
1971
|
+
}
|
|
1972
|
+
const schemaType = getSchemaType(schema);
|
|
1973
|
+
if (schemaType === "string") {
|
|
1974
|
+
return value.map((item) => {
|
|
1975
|
+
if (typeof item === "string") {
|
|
1976
|
+
return item.trim();
|
|
1977
|
+
}
|
|
1978
|
+
if (item && typeof item === "object" && textNodeName in item) {
|
|
1979
|
+
const textVal = item[textNodeName];
|
|
1980
|
+
return typeof textVal === "string" ? textVal.trim() : String(textVal);
|
|
1981
|
+
}
|
|
1982
|
+
return String(item);
|
|
1983
|
+
});
|
|
1984
|
+
}
|
|
1985
|
+
return value.map((item) => {
|
|
1986
|
+
if (typeof item === "string") {
|
|
1987
|
+
return item.trim();
|
|
1988
|
+
}
|
|
1989
|
+
if (item && typeof item === "object" && textNodeName in item) {
|
|
1990
|
+
const textVal = item[textNodeName];
|
|
1991
|
+
return typeof textVal === "string" ? textVal.trim() : textVal;
|
|
1992
|
+
}
|
|
1993
|
+
return item;
|
|
1994
|
+
});
|
|
1995
|
+
}
|
|
1996
|
+
function processIndexedTuple(obj, textNodeName) {
|
|
1997
|
+
const keys = Object.keys(obj);
|
|
1998
|
+
const indices = keys.map((k) => Number.parseInt(k, 10)).sort((a, b) => a - b);
|
|
1999
|
+
const isValidTuple = indices[0] === 0 && indices.every((val, idx) => val === idx);
|
|
2000
|
+
if (!isValidTuple) {
|
|
2001
|
+
return [obj];
|
|
2002
|
+
}
|
|
2003
|
+
const sortedKeys = keys.sort(
|
|
2004
|
+
(a, b) => Number.parseInt(a, 10) - Number.parseInt(b, 10)
|
|
2005
|
+
);
|
|
2006
|
+
return sortedKeys.map((key) => {
|
|
2007
|
+
const item = obj[key];
|
|
2008
|
+
if (item && typeof item === "object" && textNodeName in item) {
|
|
2009
|
+
const textVal = item[textNodeName];
|
|
2010
|
+
return typeof textVal === "string" ? textVal.trim() : textVal;
|
|
2011
|
+
}
|
|
2012
|
+
return typeof item === "string" ? item.trim() : item;
|
|
2013
|
+
});
|
|
2014
|
+
}
|
|
2015
|
+
|
|
2016
|
+
// src/rxml/schema/extraction.ts
|
|
2017
|
+
function skipDoctype(xmlContent, i, len) {
|
|
2018
|
+
const gt = xmlContent.indexOf(">", i + 1);
|
|
2019
|
+
return gt === -1 ? len : gt + 1;
|
|
2020
|
+
}
|
|
2021
|
+
function skipComment(xmlContent, i, len) {
|
|
2022
|
+
const close = xmlContent.indexOf("-->", i + 4);
|
|
2023
|
+
return close === -1 ? len : close + 3;
|
|
2024
|
+
}
|
|
2025
|
+
function skipCdata(xmlContent, i, len) {
|
|
2026
|
+
const close = xmlContent.indexOf("]]>", i + 9);
|
|
2027
|
+
return close === -1 ? len : close + 3;
|
|
2028
|
+
}
|
|
2029
|
+
function skipProcessingInstruction(xmlContent, i, len) {
|
|
2030
|
+
const close = xmlContent.indexOf("?>", i + 1);
|
|
2031
|
+
return close === -1 ? len : close + 2;
|
|
2032
|
+
}
|
|
2033
|
+
function skipSpecialConstruct(xmlContent, i, len) {
|
|
2034
|
+
const ch = xmlContent[i];
|
|
2035
|
+
if (ch === "!") {
|
|
2036
|
+
if (xmlContent.startsWith("!DOCTYPE", i + 1)) {
|
|
2037
|
+
return skipDoctype(xmlContent, i, len);
|
|
2038
|
+
}
|
|
2039
|
+
if (xmlContent.startsWith("!--", i + 1)) {
|
|
2040
|
+
return skipComment(xmlContent, i, len);
|
|
2041
|
+
}
|
|
2042
|
+
if (xmlContent.startsWith("![CDATA[", i + 1)) {
|
|
2043
|
+
return skipCdata(xmlContent, i, len);
|
|
2044
|
+
}
|
|
2045
|
+
const gt = xmlContent.indexOf(">", i + 1);
|
|
2046
|
+
return gt === -1 ? len : gt + 1;
|
|
2047
|
+
}
|
|
2048
|
+
if (ch === "?") {
|
|
2049
|
+
return skipProcessingInstruction(xmlContent, i, len);
|
|
2050
|
+
}
|
|
2051
|
+
return -1;
|
|
2052
|
+
}
|
|
2053
|
+
function parseTagName(xmlContent, i, len) {
|
|
2054
|
+
let j = i;
|
|
2055
|
+
if (j < len && isNameStartChar(xmlContent[j])) {
|
|
2056
|
+
j += 1;
|
|
2057
|
+
while (j < len && isNameChar(xmlContent[j])) {
|
|
2058
|
+
j += 1;
|
|
2059
|
+
}
|
|
2060
|
+
}
|
|
2061
|
+
return { name: xmlContent.slice(i, j), pos: j };
|
|
2062
|
+
}
|
|
2063
|
+
function skipToTagEnd(xmlContent, start, len) {
|
|
2064
|
+
let k = start;
|
|
2065
|
+
let isSelfClosing = false;
|
|
2066
|
+
while (k < len) {
|
|
2067
|
+
const c = xmlContent[k];
|
|
2068
|
+
if (c === '"' || c === "'") {
|
|
2069
|
+
k = skipQuoted(xmlContent, k);
|
|
2070
|
+
continue;
|
|
2071
|
+
}
|
|
2072
|
+
if (c === ">") {
|
|
2073
|
+
break;
|
|
2074
|
+
}
|
|
2075
|
+
if (c === "/" && xmlContent[k + 1] === ">") {
|
|
2076
|
+
isSelfClosing = true;
|
|
2077
|
+
k += 1;
|
|
2078
|
+
break;
|
|
2079
|
+
}
|
|
2080
|
+
k += 1;
|
|
2081
|
+
}
|
|
2082
|
+
return { pos: k, isSelfClosing };
|
|
2083
|
+
}
|
|
2084
|
+
function processClosingTagMatch(options) {
|
|
2085
|
+
const { xmlContent, nx, len, tagName, depth, nextLt } = options;
|
|
2086
|
+
const tagInfo = parseTagName(xmlContent, nx + 1, len);
|
|
2087
|
+
const gt = xmlContent.indexOf(">", tagInfo.pos);
|
|
2088
|
+
if (tagInfo.name === tagName) {
|
|
2089
|
+
const newDepth = depth - 1;
|
|
2090
|
+
if (newDepth === 0) {
|
|
2091
|
+
return { newPos: nextLt, newDepth, found: true };
|
|
2092
|
+
}
|
|
2093
|
+
return { newPos: gt === -1 ? len : gt + 1, newDepth, found: false };
|
|
2094
|
+
}
|
|
2095
|
+
return { newPos: gt === -1 ? len : gt + 1, newDepth: depth, found: false };
|
|
2096
|
+
}
|
|
2097
|
+
function processOpeningTagMatch(options) {
|
|
2098
|
+
const { xmlContent, nx, len, tagName, depth } = options;
|
|
2099
|
+
const tagInfo = parseTagName(xmlContent, nx, len);
|
|
2100
|
+
const tagEndInfo = skipToTagEnd(xmlContent, tagInfo.pos, len);
|
|
2101
|
+
const newDepth = tagInfo.name === tagName && !tagEndInfo.isSelfClosing ? depth + 1 : depth;
|
|
2102
|
+
const newPos = xmlContent[tagEndInfo.pos] === ">" ? tagEndInfo.pos + 1 : tagEndInfo.pos + 1;
|
|
2103
|
+
return { newPos, newDepth };
|
|
2104
|
+
}
|
|
2105
|
+
function findMatchingCloseTag(xmlContent, startPos, tagName, len) {
|
|
2106
|
+
let pos = startPos;
|
|
2107
|
+
let depth = 1;
|
|
2108
|
+
while (pos < len) {
|
|
2109
|
+
const nextLt = xmlContent.indexOf("<", pos);
|
|
2110
|
+
if (nextLt === -1 || nextLt + 1 >= len) {
|
|
2111
|
+
break;
|
|
2112
|
+
}
|
|
2113
|
+
const nx = nextLt + 1;
|
|
2114
|
+
const h = xmlContent[nx];
|
|
2115
|
+
const specialPos = skipSpecialConstruct(xmlContent, nx, len);
|
|
2116
|
+
if (specialPos !== -1) {
|
|
2117
|
+
pos = specialPos;
|
|
2118
|
+
continue;
|
|
2119
|
+
}
|
|
2120
|
+
if (h === "/") {
|
|
2121
|
+
const result = processClosingTagMatch({
|
|
2122
|
+
xmlContent,
|
|
2123
|
+
nx,
|
|
2124
|
+
len,
|
|
2125
|
+
tagName,
|
|
2126
|
+
depth,
|
|
2127
|
+
nextLt
|
|
2128
|
+
});
|
|
2129
|
+
if (result.found) {
|
|
2130
|
+
return result.newPos;
|
|
2131
|
+
}
|
|
2132
|
+
pos = result.newPos;
|
|
2133
|
+
depth = result.newDepth;
|
|
2134
|
+
} else {
|
|
2135
|
+
const result = processOpeningTagMatch({
|
|
2136
|
+
xmlContent,
|
|
2137
|
+
nx,
|
|
2138
|
+
len,
|
|
2139
|
+
tagName,
|
|
2140
|
+
depth
|
|
2141
|
+
});
|
|
2142
|
+
pos = result.newPos;
|
|
2143
|
+
depth = result.newDepth;
|
|
2144
|
+
}
|
|
2145
|
+
}
|
|
2146
|
+
return -1;
|
|
2147
|
+
}
|
|
2148
|
+
function updateBestMatch(depth, bestDepth, contentStart, contentEnd) {
|
|
2149
|
+
if (depth < bestDepth) {
|
|
2150
|
+
return { start: contentStart, end: contentEnd, depth };
|
|
2151
|
+
}
|
|
2152
|
+
return null;
|
|
2153
|
+
}
|
|
2154
|
+
function processTargetTag(options) {
|
|
2155
|
+
const { xmlContent, tagEnd, isSelfClosing, target, len, depth, bestDepth } = options;
|
|
2156
|
+
const contentStart = xmlContent[tagEnd] === ">" ? tagEnd + 1 : tagEnd + 1;
|
|
2157
|
+
if (isSelfClosing) {
|
|
2158
|
+
return updateBestMatch(depth, bestDepth, contentStart, contentStart);
|
|
2159
|
+
}
|
|
2160
|
+
const closePos = findMatchingCloseTag(xmlContent, contentStart, target, len);
|
|
2161
|
+
if (closePos !== -1) {
|
|
2162
|
+
return updateBestMatch(depth, bestDepth, contentStart, closePos);
|
|
2163
|
+
}
|
|
2164
|
+
return null;
|
|
2165
|
+
}
|
|
2166
|
+
function handleClosingTagInExtract(xmlContent, i, len, depth) {
|
|
2167
|
+
const gt = xmlContent.indexOf(">", i + 1);
|
|
2168
|
+
return {
|
|
2169
|
+
newPos: gt === -1 ? len : gt + 1,
|
|
2170
|
+
newDepth: Math.max(0, depth - 1)
|
|
2171
|
+
};
|
|
2172
|
+
}
|
|
2173
|
+
function processOpeningTagInExtract(options) {
|
|
2174
|
+
const { xmlContent, i, len, target, depth, bestDepth } = options;
|
|
2175
|
+
const tagInfo = parseTagName(xmlContent, i, len);
|
|
2176
|
+
const tagEndInfo = skipToTagEnd(xmlContent, tagInfo.pos, len);
|
|
2177
|
+
const tagEnd = tagEndInfo.pos;
|
|
2178
|
+
const isSelfClosing = tagEndInfo.isSelfClosing;
|
|
2179
|
+
let bestMatch = null;
|
|
2180
|
+
if (tagInfo.name === target) {
|
|
2181
|
+
bestMatch = processTargetTag({
|
|
2182
|
+
xmlContent,
|
|
2183
|
+
tagEnd,
|
|
2184
|
+
isSelfClosing,
|
|
2185
|
+
target,
|
|
2186
|
+
len,
|
|
2187
|
+
depth,
|
|
2188
|
+
bestDepth
|
|
2189
|
+
});
|
|
2190
|
+
}
|
|
2191
|
+
return {
|
|
2192
|
+
newPos: xmlContent[tagEnd] === ">" ? tagEnd + 1 : tagEnd + 1,
|
|
2193
|
+
newDepth: depth + (isSelfClosing ? 0 : 1),
|
|
2194
|
+
bestMatch
|
|
2195
|
+
};
|
|
2196
|
+
}
|
|
2197
|
+
function extractRawInner(xmlContent, tagName) {
|
|
2198
|
+
const len = xmlContent.length;
|
|
2199
|
+
const target = tagName;
|
|
2200
|
+
let bestStart = -1;
|
|
2201
|
+
let bestEnd = -1;
|
|
2202
|
+
let bestDepth = Number.POSITIVE_INFINITY;
|
|
2203
|
+
let i = 0;
|
|
2204
|
+
let depth = 0;
|
|
2205
|
+
while (i < len) {
|
|
2206
|
+
const lt = xmlContent.indexOf("<", i);
|
|
2207
|
+
if (lt === -1 || lt + 1 >= len) {
|
|
2208
|
+
return;
|
|
2209
|
+
}
|
|
2210
|
+
i = lt + 1;
|
|
2211
|
+
const ch = xmlContent[i];
|
|
2212
|
+
const specialPos = skipSpecialConstruct(xmlContent, i, len);
|
|
2213
|
+
if (specialPos !== -1) {
|
|
2214
|
+
i = specialPos;
|
|
2215
|
+
continue;
|
|
2216
|
+
}
|
|
2217
|
+
if (ch === "/") {
|
|
2218
|
+
const result2 = handleClosingTagInExtract(xmlContent, i, len, depth);
|
|
2219
|
+
i = result2.newPos;
|
|
2220
|
+
depth = result2.newDepth;
|
|
2221
|
+
continue;
|
|
2222
|
+
}
|
|
2223
|
+
const result = processOpeningTagInExtract({
|
|
2224
|
+
xmlContent,
|
|
2225
|
+
i,
|
|
2226
|
+
len,
|
|
2227
|
+
target,
|
|
2228
|
+
depth,
|
|
2229
|
+
bestDepth
|
|
2230
|
+
});
|
|
2231
|
+
if (result.bestMatch) {
|
|
2232
|
+
bestStart = result.bestMatch.start;
|
|
2233
|
+
bestEnd = result.bestMatch.end;
|
|
2234
|
+
bestDepth = result.bestMatch.depth;
|
|
2235
|
+
}
|
|
2236
|
+
i = result.newPos;
|
|
2237
|
+
depth = result.newDepth;
|
|
2238
|
+
}
|
|
2239
|
+
if (bestStart !== -1) {
|
|
2240
|
+
return xmlContent.slice(bestStart, bestEnd);
|
|
2241
|
+
}
|
|
2242
|
+
return;
|
|
2243
|
+
}
|
|
2244
|
+
function processOpeningTag(options) {
|
|
2245
|
+
const { xmlContent, tagEnd, isSelfClosing, target, len, ranges } = options;
|
|
2246
|
+
const contentStart = xmlContent[tagEnd] === ">" ? tagEnd + 1 : tagEnd + 1;
|
|
2247
|
+
if (isSelfClosing) {
|
|
2248
|
+
ranges.push({ start: contentStart, end: contentStart });
|
|
2249
|
+
return contentStart;
|
|
2250
|
+
}
|
|
2251
|
+
const closePos = findMatchingCloseTag(xmlContent, contentStart, target, len);
|
|
2252
|
+
if (closePos !== -1) {
|
|
2253
|
+
ranges.push({ start: contentStart, end: closePos });
|
|
2254
|
+
const gt = xmlContent.indexOf(">", closePos);
|
|
2255
|
+
return gt === -1 ? len : gt + 1;
|
|
2256
|
+
}
|
|
2257
|
+
return -1;
|
|
2258
|
+
}
|
|
2259
|
+
function handleClosingTagInFindAll(xmlContent, i, len) {
|
|
2260
|
+
const gt = xmlContent.indexOf(">", i + 1);
|
|
2261
|
+
return gt === -1 ? len : gt + 1;
|
|
2262
|
+
}
|
|
2263
|
+
function findAllInnerRanges(xmlContent, tagName) {
|
|
2264
|
+
const len = xmlContent.length;
|
|
2265
|
+
const target = tagName;
|
|
2266
|
+
const ranges = [];
|
|
2267
|
+
let i = 0;
|
|
2268
|
+
while (i < len) {
|
|
2269
|
+
const lt = xmlContent.indexOf("<", i);
|
|
2270
|
+
if (lt === -1 || lt + 1 >= len) {
|
|
2271
|
+
break;
|
|
2272
|
+
}
|
|
2273
|
+
i = lt + 1;
|
|
2274
|
+
const ch = xmlContent[i];
|
|
2275
|
+
const specialPos = skipSpecialConstruct(xmlContent, i, len);
|
|
2276
|
+
if (specialPos !== -1) {
|
|
2277
|
+
i = specialPos;
|
|
2278
|
+
continue;
|
|
2279
|
+
}
|
|
2280
|
+
if (ch === "/") {
|
|
2281
|
+
i = handleClosingTagInFindAll(xmlContent, i, len);
|
|
2282
|
+
continue;
|
|
2283
|
+
}
|
|
2284
|
+
const tagInfo = parseTagName(xmlContent, i, len);
|
|
2285
|
+
const tagEndInfo = skipToTagEnd(xmlContent, tagInfo.pos, len);
|
|
2286
|
+
const tagEnd = tagEndInfo.pos;
|
|
2287
|
+
const isSelfClosing = tagEndInfo.isSelfClosing;
|
|
2288
|
+
if (tagInfo.name !== target) {
|
|
2289
|
+
i = xmlContent[tagEnd] === ">" ? tagEnd + 1 : tagEnd + 1;
|
|
2290
|
+
continue;
|
|
2291
|
+
}
|
|
2292
|
+
const nextPos = processOpeningTag({
|
|
2293
|
+
xmlContent,
|
|
2294
|
+
tagEnd,
|
|
2295
|
+
isSelfClosing,
|
|
2296
|
+
target,
|
|
2297
|
+
len,
|
|
2298
|
+
ranges
|
|
2299
|
+
});
|
|
2300
|
+
if (nextPos === -1) {
|
|
2301
|
+
break;
|
|
2302
|
+
}
|
|
2303
|
+
i = nextPos;
|
|
2304
|
+
}
|
|
2305
|
+
return ranges;
|
|
2306
|
+
}
|
|
2307
|
+
function findTopLevelTargetRange(options) {
|
|
2308
|
+
const { xmlContent, tagEnd, isSelfClosing, target, len } = options;
|
|
2309
|
+
const contentStart = xmlContent[tagEnd] === ">" ? tagEnd + 1 : tagEnd + 1;
|
|
2310
|
+
if (isSelfClosing) {
|
|
2311
|
+
return { start: contentStart, end: contentStart };
|
|
2312
|
+
}
|
|
2313
|
+
const closePos = findMatchingCloseTag(xmlContent, contentStart, target, len);
|
|
2314
|
+
if (closePos !== -1) {
|
|
2315
|
+
return { start: contentStart, end: closePos };
|
|
2316
|
+
}
|
|
2317
|
+
return;
|
|
2318
|
+
}
|
|
2319
|
+
function handleClosingTagInFindFirst(xmlContent, i, len, depth) {
|
|
2320
|
+
const gt = xmlContent.indexOf(">", i + 1);
|
|
2321
|
+
return {
|
|
2322
|
+
newPos: gt === -1 ? len : gt + 1,
|
|
2323
|
+
newDepth: Math.max(0, depth - 1)
|
|
2324
|
+
};
|
|
2325
|
+
}
|
|
2326
|
+
function findFirstTopLevelRange(xmlContent, tagName) {
|
|
2327
|
+
const len = xmlContent.length;
|
|
2328
|
+
const target = tagName;
|
|
2329
|
+
let i = 0;
|
|
2330
|
+
let depth = 0;
|
|
2331
|
+
while (i < len) {
|
|
2332
|
+
const lt = xmlContent.indexOf("<", i);
|
|
2333
|
+
if (lt === -1 || lt + 1 >= len) {
|
|
2334
|
+
return;
|
|
2335
|
+
}
|
|
2336
|
+
i = lt + 1;
|
|
2337
|
+
const ch = xmlContent[i];
|
|
2338
|
+
const specialPos = skipSpecialConstruct(xmlContent, i, len);
|
|
2339
|
+
if (specialPos !== -1) {
|
|
2340
|
+
i = specialPos;
|
|
2341
|
+
continue;
|
|
2342
|
+
}
|
|
2343
|
+
if (ch === "/") {
|
|
2344
|
+
const result = handleClosingTagInFindFirst(xmlContent, i, len, depth);
|
|
2345
|
+
i = result.newPos;
|
|
2346
|
+
depth = result.newDepth;
|
|
2347
|
+
continue;
|
|
2348
|
+
}
|
|
2349
|
+
const tagInfo = parseTagName(xmlContent, i, len);
|
|
2350
|
+
const tagEndInfo = skipToTagEnd(xmlContent, tagInfo.pos, len);
|
|
2351
|
+
const tagEnd = tagEndInfo.pos;
|
|
2352
|
+
const isSelfClosing = tagEndInfo.isSelfClosing;
|
|
2353
|
+
if (depth === 0 && tagInfo.name === target) {
|
|
2354
|
+
return findTopLevelTargetRange({
|
|
2355
|
+
xmlContent,
|
|
2356
|
+
tagEnd,
|
|
2357
|
+
isSelfClosing,
|
|
2358
|
+
target,
|
|
2359
|
+
len
|
|
2360
|
+
});
|
|
2361
|
+
}
|
|
2362
|
+
i = xmlContent[tagEnd] === ">" ? tagEnd + 1 : tagEnd + 1;
|
|
2363
|
+
depth += isSelfClosing ? 0 : 1;
|
|
2364
|
+
}
|
|
2365
|
+
return;
|
|
2366
|
+
}
|
|
2367
|
+
function isPositionExcluded(pos, excludeRanges) {
|
|
2368
|
+
if (!excludeRanges || excludeRanges.length === 0) {
|
|
2369
|
+
return false;
|
|
2370
|
+
}
|
|
2371
|
+
for (const r of excludeRanges) {
|
|
2372
|
+
if (pos >= r.start && pos < r.end) {
|
|
2373
|
+
return true;
|
|
2374
|
+
}
|
|
2375
|
+
}
|
|
2376
|
+
return false;
|
|
2377
|
+
}
|
|
2378
|
+
function skipCommentInCounting(xmlContent, i, len) {
|
|
2379
|
+
const close = xmlContent.indexOf("-->", i + 4);
|
|
2380
|
+
return close === -1 ? len : close + 3;
|
|
2381
|
+
}
|
|
2382
|
+
function skipCdataInCounting(xmlContent, i, len) {
|
|
2383
|
+
const close = xmlContent.indexOf("]]>", i + 9);
|
|
2384
|
+
return close === -1 ? len : close + 3;
|
|
2385
|
+
}
|
|
2386
|
+
function skipSpecialInCounting(xmlContent, ch, i, len) {
|
|
2387
|
+
if (ch === "!") {
|
|
2388
|
+
if (xmlContent.startsWith("!--", i + 1)) {
|
|
2389
|
+
return skipCommentInCounting(xmlContent, i, len);
|
|
2390
|
+
}
|
|
2391
|
+
if (xmlContent.startsWith("![CDATA[", i + 1)) {
|
|
2392
|
+
return skipCdataInCounting(xmlContent, i, len);
|
|
2393
|
+
}
|
|
2394
|
+
const gt = xmlContent.indexOf(">", i + 1);
|
|
2395
|
+
return gt === -1 ? len : gt + 1;
|
|
2396
|
+
}
|
|
2397
|
+
if (ch === "?") {
|
|
2398
|
+
const close = xmlContent.indexOf("?>", i + 1);
|
|
2399
|
+
return close === -1 ? len : close + 2;
|
|
2400
|
+
}
|
|
2401
|
+
if (ch === "/") {
|
|
2402
|
+
const gt = xmlContent.indexOf(">", i + 1);
|
|
2403
|
+
return gt === -1 ? len : gt + 1;
|
|
2404
|
+
}
|
|
2405
|
+
return -1;
|
|
2406
|
+
}
|
|
2407
|
+
function parseAndCountTag(options) {
|
|
2408
|
+
const { xmlContent, i, len, target, lt, excludeRanges } = options;
|
|
2409
|
+
let j = i;
|
|
2410
|
+
if (j < len && isNameStartChar(xmlContent[j])) {
|
|
2411
|
+
j += 1;
|
|
2412
|
+
while (j < len && isNameChar(xmlContent[j])) {
|
|
2413
|
+
j += 1;
|
|
2414
|
+
}
|
|
2415
|
+
}
|
|
2416
|
+
const name = xmlContent.slice(i, j);
|
|
2417
|
+
let k = j;
|
|
2418
|
+
while (k < len) {
|
|
2419
|
+
const c = xmlContent[k];
|
|
2420
|
+
if (c === '"' || c === "'") {
|
|
2421
|
+
k = skipQuoted(xmlContent, k);
|
|
2422
|
+
continue;
|
|
2423
|
+
}
|
|
2424
|
+
if (c === ">") {
|
|
2425
|
+
break;
|
|
2426
|
+
}
|
|
2427
|
+
if (c === "/" && xmlContent[k + 1] === ">") {
|
|
2428
|
+
k += 1;
|
|
2429
|
+
break;
|
|
2430
|
+
}
|
|
2431
|
+
k += 1;
|
|
2432
|
+
}
|
|
2433
|
+
const shouldCount = name === target && !isPositionExcluded(lt, excludeRanges);
|
|
2434
|
+
return { nextPos: k + 1, shouldCount };
|
|
2435
|
+
}
|
|
2436
|
+
function countTagOccurrences(xmlContent, tagName, excludeRanges, shouldSkipFirst = true) {
|
|
2437
|
+
const len = xmlContent.length;
|
|
2438
|
+
const target = tagName;
|
|
2439
|
+
let i = 0;
|
|
2440
|
+
let count = 0;
|
|
2441
|
+
let skipFirstLocal = shouldSkipFirst;
|
|
2442
|
+
while (i < len) {
|
|
2443
|
+
const lt = xmlContent.indexOf("<", i);
|
|
2444
|
+
if (lt === -1) {
|
|
2445
|
+
break;
|
|
2446
|
+
}
|
|
2447
|
+
i = lt + 1;
|
|
2448
|
+
if (i >= len) {
|
|
2449
|
+
break;
|
|
2450
|
+
}
|
|
2451
|
+
const ch = xmlContent[i];
|
|
2452
|
+
const skipPos = skipSpecialInCounting(xmlContent, ch, i, len);
|
|
2453
|
+
if (skipPos !== -1) {
|
|
2454
|
+
i = skipPos;
|
|
2455
|
+
continue;
|
|
2456
|
+
}
|
|
2457
|
+
const result = parseAndCountTag({
|
|
2458
|
+
xmlContent,
|
|
2459
|
+
i,
|
|
2460
|
+
len,
|
|
2461
|
+
target,
|
|
2462
|
+
lt,
|
|
2463
|
+
excludeRanges
|
|
2464
|
+
});
|
|
2465
|
+
if (result.shouldCount) {
|
|
2466
|
+
if (skipFirstLocal) {
|
|
2467
|
+
skipFirstLocal = false;
|
|
2468
|
+
} else {
|
|
2469
|
+
count += 1;
|
|
2470
|
+
}
|
|
2471
|
+
}
|
|
2472
|
+
i = result.nextPos;
|
|
2473
|
+
}
|
|
2474
|
+
return count;
|
|
2475
|
+
}
|
|
2476
|
+
|
|
2477
|
+
// src/rxml/core/tokenizer.ts
|
|
2478
|
+
var XMLTokenizer = class {
|
|
2479
|
+
constructor(xmlString, options = {}) {
|
|
2480
|
+
this.pos = 0;
|
|
2481
|
+
this.xmlString = xmlString;
|
|
2482
|
+
this.options = {
|
|
2483
|
+
keepComments: false,
|
|
2484
|
+
keepWhitespace: false,
|
|
2485
|
+
noChildNodes: DEFAULT_NO_CHILD_NODES.slice(),
|
|
2486
|
+
textNodeName: "#text",
|
|
2487
|
+
throwOnDuplicateStringTags: true,
|
|
2488
|
+
...options
|
|
2489
|
+
};
|
|
2490
|
+
this.pos = options.pos || 0;
|
|
2491
|
+
}
|
|
2492
|
+
/**
|
|
2493
|
+
* Handle closing tag parsing
|
|
2494
|
+
*/
|
|
2495
|
+
handleClosingTag(tagName, children) {
|
|
2496
|
+
const closeStart = this.pos + 2;
|
|
2497
|
+
this.pos = this.xmlString.indexOf(">", this.pos);
|
|
2498
|
+
const closeTag = this.xmlString.substring(closeStart, this.pos);
|
|
2499
|
+
if (tagName && closeTag.trim() !== tagName) {
|
|
2500
|
+
const { line, column } = getLineColumn(this.xmlString, this.pos);
|
|
2501
|
+
throw new RXMLParseError(
|
|
2502
|
+
`Unexpected close tag at line ${line}, column ${column}. Expected </${tagName}>, found </${closeTag}>`,
|
|
2503
|
+
void 0,
|
|
2504
|
+
line,
|
|
2505
|
+
column
|
|
2506
|
+
);
|
|
2507
|
+
}
|
|
2508
|
+
if (this.pos !== -1) {
|
|
2509
|
+
this.pos += 1;
|
|
2510
|
+
}
|
|
2511
|
+
return children;
|
|
2512
|
+
}
|
|
2513
|
+
/**
|
|
2514
|
+
* Check if we're at end of string and should throw unclosed tag error
|
|
2515
|
+
*/
|
|
2516
|
+
checkUnclosedTag(tagName, consumedToEnd) {
|
|
2517
|
+
if (tagName && this.pos >= this.xmlString.length && !consumedToEnd) {
|
|
2518
|
+
const { line, column } = getLineColumn(this.xmlString, this.pos - 1);
|
|
2519
|
+
throw new RXMLParseError(
|
|
2520
|
+
`Unclosed tag at line ${line}, column ${column}. Expected closing tag </${tagName}>`,
|
|
2521
|
+
void 0,
|
|
2522
|
+
line,
|
|
2523
|
+
column
|
|
2524
|
+
);
|
|
2525
|
+
}
|
|
2526
|
+
}
|
|
2527
|
+
/**
|
|
2528
|
+
* Process special content (comments, CDATA, DOCTYPE) and track if we consumed to end
|
|
2529
|
+
*/
|
|
2530
|
+
processSpecialContent(children) {
|
|
2531
|
+
const prevPos = this.pos;
|
|
2532
|
+
this.handleSpecialContent(children);
|
|
2533
|
+
return this.pos >= this.xmlString.length && prevPos < this.xmlString.length;
|
|
2534
|
+
}
|
|
2535
|
+
/**
|
|
2536
|
+
* Handle text content parsing
|
|
2537
|
+
*/
|
|
2538
|
+
handleTextContent(children) {
|
|
2539
|
+
const text = this.parseText();
|
|
2540
|
+
if (this.options.keepWhitespace) {
|
|
2541
|
+
if (text.length > 0) {
|
|
2542
|
+
children.push(text);
|
|
2543
|
+
}
|
|
2544
|
+
} else {
|
|
2545
|
+
const trimmed = text.trim();
|
|
2546
|
+
if (trimmed.length > 0) {
|
|
2547
|
+
children.push(trimmed);
|
|
2548
|
+
}
|
|
2549
|
+
}
|
|
2550
|
+
this.pos += 1;
|
|
2551
|
+
}
|
|
2552
|
+
/**
|
|
2553
|
+
* Handle regular element parsing
|
|
2554
|
+
*/
|
|
2555
|
+
handleRegularElement(children) {
|
|
2556
|
+
const node = this.parseNode();
|
|
2557
|
+
children.push(node);
|
|
2558
|
+
if (node.tagName[0] === "?") {
|
|
2559
|
+
children.push(...node.children);
|
|
2560
|
+
node.children = [];
|
|
2561
|
+
}
|
|
2562
|
+
}
|
|
2563
|
+
/**
|
|
2564
|
+
* Process a single child element based on the current character
|
|
2565
|
+
*/
|
|
2566
|
+
processSingleChild(children, tagName) {
|
|
2567
|
+
if (this.xmlString.charCodeAt(this.pos) !== CharCodes.OPEN_BRACKET) {
|
|
2568
|
+
this.handleTextContent(children);
|
|
2569
|
+
return { shouldReturn: false, consumedToEnd: false };
|
|
2570
|
+
}
|
|
2571
|
+
const nextChar = this.xmlString.charCodeAt(this.pos + 1);
|
|
2572
|
+
if (nextChar === CharCodes.SLASH) {
|
|
2573
|
+
const result = this.handleClosingTag(tagName, children);
|
|
2574
|
+
if (result !== null) {
|
|
2575
|
+
return { shouldReturn: true, consumedToEnd: false };
|
|
2576
|
+
}
|
|
2577
|
+
return { shouldReturn: false, consumedToEnd: false };
|
|
2578
|
+
}
|
|
2579
|
+
if (nextChar === CharCodes.EXCLAMATION) {
|
|
2580
|
+
const wasConsumedToEnd = this.processSpecialContent(children);
|
|
2581
|
+
return { shouldReturn: false, consumedToEnd: wasConsumedToEnd };
|
|
2582
|
+
}
|
|
2583
|
+
this.handleRegularElement(children);
|
|
2584
|
+
return { shouldReturn: false, consumedToEnd: false };
|
|
2585
|
+
}
|
|
2586
|
+
/**
|
|
2587
|
+
* Parse XML children recursively
|
|
2588
|
+
*/
|
|
2589
|
+
parseChildren(tagName) {
|
|
2590
|
+
const children = [];
|
|
2591
|
+
let consumedToEnd = false;
|
|
2592
|
+
while (this.xmlString[this.pos]) {
|
|
2593
|
+
const result = this.processSingleChild(children, tagName);
|
|
2594
|
+
if (result.shouldReturn) {
|
|
2595
|
+
return children;
|
|
2596
|
+
}
|
|
2597
|
+
if (result.consumedToEnd) {
|
|
2598
|
+
consumedToEnd = true;
|
|
2599
|
+
}
|
|
2600
|
+
}
|
|
2601
|
+
this.checkUnclosedTag(tagName, consumedToEnd);
|
|
2602
|
+
return children;
|
|
2603
|
+
}
|
|
2604
|
+
/**
|
|
2605
|
+
* Check if character is whitespace
|
|
2606
|
+
*/
|
|
2607
|
+
isWhitespace(code) {
|
|
2608
|
+
return code === CharCodes.SPACE || code === CharCodes.TAB || code === CharCodes.NEWLINE || code === CharCodes.CARRIAGE_RETURN;
|
|
2609
|
+
}
|
|
2610
|
+
/**
|
|
2611
|
+
* Skip whitespace characters
|
|
2612
|
+
*/
|
|
2613
|
+
skipWhitespace() {
|
|
2614
|
+
while (this.pos < this.xmlString.length && this.isWhitespace(this.xmlString.charCodeAt(this.pos))) {
|
|
2615
|
+
this.pos += 1;
|
|
2616
|
+
}
|
|
2617
|
+
}
|
|
2618
|
+
/**
|
|
2619
|
+
* Parse attribute value
|
|
2620
|
+
*/
|
|
2621
|
+
parseAttributeValue() {
|
|
2622
|
+
if (this.pos >= this.xmlString.length || this.xmlString[this.pos] !== "=") {
|
|
2623
|
+
return null;
|
|
2624
|
+
}
|
|
2625
|
+
this.pos += 1;
|
|
2626
|
+
this.skipWhitespace();
|
|
2627
|
+
const code = this.xmlString.charCodeAt(this.pos);
|
|
2628
|
+
if (code === CharCodes.SINGLE_QUOTE || code === CharCodes.DOUBLE_QUOTE) {
|
|
2629
|
+
const { value: parsedValue, newPos: valueEnd } = parseString(
|
|
2630
|
+
this.xmlString,
|
|
2631
|
+
this.pos
|
|
2632
|
+
);
|
|
2633
|
+
this.pos = valueEnd;
|
|
2634
|
+
return parsedValue;
|
|
2635
|
+
}
|
|
2636
|
+
return null;
|
|
2637
|
+
}
|
|
2638
|
+
/**
|
|
2639
|
+
* Parse single attribute
|
|
2640
|
+
*/
|
|
2641
|
+
parseAttribute(attributes) {
|
|
2642
|
+
const { name: attrName, newPos: nameEnd } = parseName(
|
|
2643
|
+
this.xmlString,
|
|
2644
|
+
this.pos
|
|
2645
|
+
);
|
|
2646
|
+
this.pos = nameEnd;
|
|
2647
|
+
this.skipWhitespace();
|
|
2648
|
+
const value = this.parseAttributeValue();
|
|
2649
|
+
attributes[attrName] = value;
|
|
2650
|
+
}
|
|
2651
|
+
/**
|
|
2652
|
+
* Parse all attributes
|
|
2653
|
+
*/
|
|
2654
|
+
parseAttributes() {
|
|
2655
|
+
const attributes = {};
|
|
2656
|
+
while (this.xmlString.charCodeAt(this.pos) !== CharCodes.CLOSE_BRACKET && this.xmlString[this.pos]) {
|
|
2657
|
+
const c = this.xmlString.charCodeAt(this.pos);
|
|
2658
|
+
if (this.isWhitespace(c)) {
|
|
2659
|
+
this.pos += 1;
|
|
2660
|
+
continue;
|
|
2661
|
+
}
|
|
2662
|
+
if (c > 64 && c < 91 || c > 96 && c < 123) {
|
|
2663
|
+
this.parseAttribute(attributes);
|
|
2664
|
+
} else {
|
|
2665
|
+
this.pos += 1;
|
|
2666
|
+
}
|
|
2667
|
+
}
|
|
2668
|
+
return attributes;
|
|
2669
|
+
}
|
|
2670
|
+
/**
|
|
2671
|
+
* Parse special tag content (script, style)
|
|
2672
|
+
*/
|
|
2673
|
+
parseSpecialTagContent(_tagName, closingTag) {
|
|
2674
|
+
const start = this.pos + 1;
|
|
2675
|
+
this.pos = this.xmlString.indexOf(closingTag, this.pos);
|
|
2676
|
+
if (this.pos === -1) {
|
|
2677
|
+
const children2 = [this.xmlString.slice(start)];
|
|
2678
|
+
this.pos = this.xmlString.length;
|
|
2679
|
+
return children2;
|
|
2680
|
+
}
|
|
2681
|
+
const children = [this.xmlString.slice(start, this.pos)];
|
|
2682
|
+
this.pos += closingTag.length;
|
|
2683
|
+
return children;
|
|
2684
|
+
}
|
|
2685
|
+
/**
|
|
2686
|
+
* Parse node children based on tag type
|
|
2687
|
+
*/
|
|
2688
|
+
parseNodeChildren(tagName, isSelfClosing) {
|
|
2689
|
+
var _a;
|
|
2690
|
+
if (isSelfClosing) {
|
|
2691
|
+
this.pos += 1;
|
|
2692
|
+
return [];
|
|
2693
|
+
}
|
|
2694
|
+
if (tagName === "script") {
|
|
2695
|
+
return this.parseSpecialTagContent(tagName, "</script>");
|
|
2696
|
+
}
|
|
2697
|
+
if (tagName === "style") {
|
|
2698
|
+
return this.parseSpecialTagContent(tagName, "</style>");
|
|
2699
|
+
}
|
|
2700
|
+
if (((_a = this.options.noChildNodes) == null ? void 0 : _a.indexOf(tagName)) === -1) {
|
|
2701
|
+
this.pos += 1;
|
|
2702
|
+
return this.parseChildren(tagName);
|
|
2703
|
+
}
|
|
2704
|
+
this.pos += 1;
|
|
2705
|
+
if (DEFAULT_NO_CHILD_NODES.includes(tagName)) {
|
|
2706
|
+
return [];
|
|
2707
|
+
}
|
|
2708
|
+
const closingTag = `</${tagName}>`;
|
|
2709
|
+
const closingPos = this.xmlString.indexOf(closingTag, this.pos);
|
|
2710
|
+
if (closingPos !== -1) {
|
|
2711
|
+
this.pos = closingPos + closingTag.length;
|
|
2712
|
+
}
|
|
2713
|
+
return [];
|
|
2714
|
+
}
|
|
2715
|
+
/**
|
|
2716
|
+
* Parse a single XML node
|
|
2717
|
+
*/
|
|
2718
|
+
parseNode() {
|
|
2719
|
+
this.pos += 1;
|
|
2720
|
+
const { name: tagName, newPos } = parseName(this.xmlString, this.pos);
|
|
2721
|
+
this.pos = newPos;
|
|
2722
|
+
const attributes = this.parseAttributes();
|
|
2723
|
+
const isSelfClosing = this.xmlString.charCodeAt(this.pos - 1) === CharCodes.SLASH || tagName[0] === "?" && this.xmlString.charCodeAt(this.pos - 1) === CharCodes.QUESTION;
|
|
2724
|
+
const children = this.parseNodeChildren(tagName, isSelfClosing);
|
|
2725
|
+
return { tagName, attributes, children };
|
|
2726
|
+
}
|
|
2727
|
+
/**
|
|
2728
|
+
* Parse text content until next tag
|
|
2729
|
+
*/
|
|
2730
|
+
parseText() {
|
|
2731
|
+
const start = this.pos;
|
|
2732
|
+
this.pos = this.xmlString.indexOf("<", this.pos) - 1;
|
|
2733
|
+
if (this.pos === -2) {
|
|
2734
|
+
this.pos = this.xmlString.length;
|
|
2735
|
+
}
|
|
2736
|
+
return this.xmlString.slice(start, this.pos + 1);
|
|
2737
|
+
}
|
|
2738
|
+
/**
|
|
2739
|
+
* Handle comments, CDATA, and DOCTYPE declarations
|
|
2740
|
+
*/
|
|
2741
|
+
handleSpecialContent(children) {
|
|
2742
|
+
if (this.xmlString.charCodeAt(this.pos + 2) === CharCodes.MINUS) {
|
|
2743
|
+
this.handleComment(children);
|
|
2744
|
+
} else if (this.xmlString.charCodeAt(this.pos + 2) === CharCodes.OPEN_CORNER_BRACKET && this.xmlString.charCodeAt(this.pos + 8) === CharCodes.OPEN_CORNER_BRACKET && this.xmlString.substr(this.pos + 3, 5).toLowerCase() === "cdata") {
|
|
2745
|
+
this.handleCData(children);
|
|
2746
|
+
} else {
|
|
2747
|
+
this.handleDoctype(children);
|
|
2748
|
+
}
|
|
2749
|
+
}
|
|
2750
|
+
/**
|
|
2751
|
+
* Handle XML comments
|
|
2752
|
+
*/
|
|
2753
|
+
handleComment(children) {
|
|
2754
|
+
const startCommentPos = this.pos;
|
|
2755
|
+
while (this.pos !== -1 && !(this.xmlString.charCodeAt(this.pos) === CharCodes.CLOSE_BRACKET && this.xmlString.charCodeAt(this.pos - 1) === CharCodes.MINUS && this.xmlString.charCodeAt(this.pos - 2) === CharCodes.MINUS)) {
|
|
2756
|
+
this.pos = this.xmlString.indexOf(">", this.pos + 1);
|
|
2757
|
+
}
|
|
2758
|
+
if (this.pos === -1) {
|
|
2759
|
+
this.pos = this.xmlString.length;
|
|
2760
|
+
}
|
|
2761
|
+
if (this.options.keepComments) {
|
|
2762
|
+
children.push(this.xmlString.substring(startCommentPos, this.pos + 1));
|
|
2763
|
+
}
|
|
2764
|
+
this.pos += 1;
|
|
2765
|
+
}
|
|
2766
|
+
/**
|
|
2767
|
+
* Handle CDATA sections
|
|
2768
|
+
*/
|
|
2769
|
+
handleCData(children) {
|
|
2770
|
+
const cdataEndIndex = this.xmlString.indexOf("]]>", this.pos);
|
|
2771
|
+
if (cdataEndIndex === -1) {
|
|
2772
|
+
children.push(this.xmlString.substr(this.pos + 9));
|
|
2773
|
+
this.pos = this.xmlString.length;
|
|
2774
|
+
} else {
|
|
2775
|
+
children.push(this.xmlString.substring(this.pos + 9, cdataEndIndex));
|
|
2776
|
+
this.pos = cdataEndIndex + 3;
|
|
2777
|
+
}
|
|
2778
|
+
}
|
|
2779
|
+
/**
|
|
2780
|
+
* Handle DOCTYPE declarations
|
|
2781
|
+
*/
|
|
2782
|
+
handleDoctype(children) {
|
|
2783
|
+
const startDoctype = this.pos + 1;
|
|
2784
|
+
this.pos += 2;
|
|
2785
|
+
let encapsulated = false;
|
|
2786
|
+
while ((this.xmlString.charCodeAt(this.pos) !== CharCodes.CLOSE_BRACKET || encapsulated) && this.xmlString[this.pos]) {
|
|
2787
|
+
if (this.xmlString.charCodeAt(this.pos) === CharCodes.OPEN_CORNER_BRACKET) {
|
|
2788
|
+
encapsulated = true;
|
|
2789
|
+
} else if (encapsulated && this.xmlString.charCodeAt(this.pos) === CharCodes.CLOSE_CORNER_BRACKET) {
|
|
2790
|
+
encapsulated = false;
|
|
2791
|
+
}
|
|
2792
|
+
this.pos += 1;
|
|
2793
|
+
}
|
|
2794
|
+
children.push(this.xmlString.substring(startDoctype, this.pos));
|
|
2795
|
+
this.pos += 1;
|
|
2796
|
+
}
|
|
2797
|
+
/**
|
|
2798
|
+
* Get current position
|
|
2799
|
+
*/
|
|
2800
|
+
getPosition() {
|
|
2801
|
+
return this.pos;
|
|
2802
|
+
}
|
|
2803
|
+
/**
|
|
2804
|
+
* Set position
|
|
2805
|
+
*/
|
|
2806
|
+
setPosition(pos) {
|
|
2807
|
+
this.pos = pos;
|
|
2808
|
+
}
|
|
2809
|
+
};
|
|
2810
|
+
|
|
2811
|
+
// src/rxml/core/parser.ts
|
|
2812
|
+
var WHITESPACE_REGEX2 = /\s/;
|
|
2813
|
+
var NUMERIC_STRING_REGEX = /^-?\d+(?:\.\d+)?(?:[eE][+-]?\d+)?$/;
|
|
2814
|
+
var DIGIT_KEY_REGEX2 = /^\d+$/;
|
|
2815
|
+
function getTopLevelStringProps(s) {
|
|
2816
|
+
const set = /* @__PURE__ */ new Set();
|
|
2817
|
+
const unwrapped = unwrapJsonSchema(s);
|
|
2818
|
+
if (unwrapped && typeof unwrapped === "object") {
|
|
2819
|
+
const props = unwrapped.properties;
|
|
2820
|
+
if (props && typeof props === "object") {
|
|
2821
|
+
for (const [k, v] of Object.entries(props)) {
|
|
2822
|
+
if (getSchemaType(v) === "string") {
|
|
2823
|
+
set.add(k);
|
|
2824
|
+
}
|
|
2825
|
+
}
|
|
2826
|
+
}
|
|
2827
|
+
}
|
|
2828
|
+
return set;
|
|
2829
|
+
}
|
|
2830
|
+
function restorePlaceholderString(val, placeholderMap) {
|
|
2831
|
+
if (val.startsWith("__RXML_PLACEHOLDER_")) {
|
|
2832
|
+
const orig = placeholderMap.get(val);
|
|
2833
|
+
return orig !== void 0 ? orig : val;
|
|
2834
|
+
}
|
|
2835
|
+
return val;
|
|
2836
|
+
}
|
|
2837
|
+
function restorePlaceholdersInObject(obj, _placeholderMap, textNodeName, restorer) {
|
|
2838
|
+
const out = {};
|
|
2839
|
+
for (const [k, v] of Object.entries(obj)) {
|
|
2840
|
+
const restored = restorer(v);
|
|
2841
|
+
if (k === textNodeName && typeof restored === "string") {
|
|
2842
|
+
out[k] = restored.trim();
|
|
2843
|
+
} else {
|
|
2844
|
+
out[k] = restored;
|
|
2845
|
+
}
|
|
2846
|
+
}
|
|
2847
|
+
return out;
|
|
2848
|
+
}
|
|
2849
|
+
function createPlaceholderRestorer(placeholderMap, textNodeName) {
|
|
2850
|
+
const restorer = (val) => {
|
|
2851
|
+
if (val == null) {
|
|
2852
|
+
return val;
|
|
2853
|
+
}
|
|
2854
|
+
if (typeof val === "string") {
|
|
2855
|
+
return restorePlaceholderString(val, placeholderMap);
|
|
2856
|
+
}
|
|
2857
|
+
if (Array.isArray(val)) {
|
|
2858
|
+
return val.map(restorer);
|
|
2859
|
+
}
|
|
2860
|
+
if (typeof val === "object") {
|
|
2861
|
+
return restorePlaceholdersInObject(
|
|
2862
|
+
val,
|
|
2863
|
+
placeholderMap,
|
|
2864
|
+
textNodeName,
|
|
2865
|
+
restorer
|
|
2866
|
+
);
|
|
2867
|
+
}
|
|
2868
|
+
return val;
|
|
2869
|
+
};
|
|
2870
|
+
return restorer;
|
|
2871
|
+
}
|
|
2872
|
+
function tryConvertToNumber(val) {
|
|
2873
|
+
if (typeof val !== "string") {
|
|
2874
|
+
return val;
|
|
2875
|
+
}
|
|
2876
|
+
const trimmed = val.trim();
|
|
2877
|
+
if (NUMERIC_STRING_REGEX.test(trimmed)) {
|
|
2878
|
+
const num = Number(trimmed);
|
|
2879
|
+
if (Number.isFinite(num)) {
|
|
2880
|
+
return num;
|
|
2881
|
+
}
|
|
2882
|
+
}
|
|
2883
|
+
return trimmed;
|
|
2884
|
+
}
|
|
2885
|
+
function processItemValue(item, textNodeName) {
|
|
2886
|
+
let currentVal = item;
|
|
2887
|
+
if (item && typeof item === "object" && Object.hasOwn(item, textNodeName)) {
|
|
2888
|
+
currentVal = item[textNodeName];
|
|
2889
|
+
}
|
|
2890
|
+
const trimmed = typeof currentVal === "string" ? currentVal.trim() : currentVal;
|
|
2891
|
+
return tryConvertToNumber(trimmed);
|
|
2892
|
+
}
|
|
2893
|
+
function processItemWrapper(itemValue, textNodeName) {
|
|
2894
|
+
if (Array.isArray(itemValue)) {
|
|
2895
|
+
return itemValue.map((item) => processItemValue(item, textNodeName));
|
|
2896
|
+
}
|
|
2897
|
+
const trimmed = typeof itemValue === "string" ? itemValue.trim() : itemValue;
|
|
2898
|
+
return tryConvertToNumber(trimmed);
|
|
2899
|
+
}
|
|
2900
|
+
function deepDecodeStringsBySchema(input, schema) {
|
|
2901
|
+
var _a;
|
|
2902
|
+
if (input == null || schema == null) {
|
|
2903
|
+
return input;
|
|
2904
|
+
}
|
|
2905
|
+
const type = getSchemaType(schema);
|
|
2906
|
+
if (type === "string" && typeof input === "string") {
|
|
2907
|
+
return unescapeXml(input);
|
|
2908
|
+
}
|
|
2909
|
+
if (type === "array" && Array.isArray(input)) {
|
|
2910
|
+
const unwrapped = unwrapJsonSchema(schema);
|
|
2911
|
+
const itemSchema = (_a = unwrapped == null ? void 0 : unwrapped.items) != null ? _a : {};
|
|
2912
|
+
return input.map((item) => deepDecodeStringsBySchema(item, itemSchema));
|
|
2913
|
+
}
|
|
2914
|
+
if (type === "object" && input && typeof input === "object") {
|
|
2915
|
+
const obj = input;
|
|
2916
|
+
const out = {};
|
|
2917
|
+
for (const key of Object.keys(obj)) {
|
|
2918
|
+
const childSchema = getPropertySchema(schema, key);
|
|
2919
|
+
out[key] = deepDecodeStringsBySchema(obj[key], childSchema);
|
|
2920
|
+
}
|
|
2921
|
+
return out;
|
|
2922
|
+
}
|
|
2923
|
+
if (typeof input === "string") {
|
|
2924
|
+
return unescapeXml(input);
|
|
2925
|
+
}
|
|
2926
|
+
return input;
|
|
2927
|
+
}
|
|
2928
|
+
function parse2(xmlInner, schema, options = {}) {
|
|
2929
|
+
var _a, _b, _c;
|
|
2930
|
+
const textNodeName = (_a = options.textNodeName) != null ? _a : "#text";
|
|
2931
|
+
const throwDup = (_b = options.throwOnDuplicateStringTags) != null ? _b : true;
|
|
2932
|
+
let actualXmlInner = xmlInner.trim();
|
|
2933
|
+
if (actualXmlInner.startsWith("<") && actualXmlInner.endsWith(">")) {
|
|
2934
|
+
const s = actualXmlInner;
|
|
2935
|
+
let i = 0;
|
|
2936
|
+
let rootStart = -1;
|
|
2937
|
+
let rootName = "";
|
|
2938
|
+
while (i < s.length) {
|
|
2939
|
+
const lt = s.indexOf("<", i);
|
|
2940
|
+
if (lt === -1) {
|
|
2941
|
+
break;
|
|
2942
|
+
}
|
|
2943
|
+
const next = s[lt + 1];
|
|
2944
|
+
if (next === "?") {
|
|
2945
|
+
const end = s.indexOf("?>", lt + 2);
|
|
2946
|
+
i = end === -1 ? s.length : end + 2;
|
|
2947
|
+
continue;
|
|
2948
|
+
}
|
|
2949
|
+
if (next === "!") {
|
|
2950
|
+
if (s.startsWith("!--", lt + 2)) {
|
|
2951
|
+
const end2 = s.indexOf("-->", lt + 5);
|
|
2952
|
+
i = end2 === -1 ? s.length : end2 + 3;
|
|
2953
|
+
continue;
|
|
2954
|
+
}
|
|
2955
|
+
if (s.startsWith("![CDATA[", lt + 2)) {
|
|
2956
|
+
const end2 = s.indexOf("]]>", lt + 9);
|
|
2957
|
+
i = end2 === -1 ? s.length : end2 + 3;
|
|
2958
|
+
continue;
|
|
2959
|
+
}
|
|
2960
|
+
const end = s.indexOf(">", lt + 2);
|
|
2961
|
+
i = end === -1 ? s.length : end + 1;
|
|
2962
|
+
continue;
|
|
2963
|
+
}
|
|
2964
|
+
if (next === "/") {
|
|
2965
|
+
break;
|
|
2966
|
+
}
|
|
2967
|
+
let j = lt + 1;
|
|
2968
|
+
while (j < s.length && s[j] !== " " && s[j] !== "\n" && s[j] !== "\r" && s[j] !== " " && s[j] !== "/" && s[j] !== ">") {
|
|
2969
|
+
j += 1;
|
|
2970
|
+
}
|
|
2971
|
+
rootStart = lt;
|
|
2972
|
+
rootName = s.slice(lt + 1, j);
|
|
2973
|
+
break;
|
|
2974
|
+
}
|
|
2975
|
+
if (rootStart === 0 && rootName) {
|
|
2976
|
+
const range = findFirstTopLevelRange(s, rootName);
|
|
2977
|
+
if (range) {
|
|
2978
|
+
let fullEnd = range.end + `</${rootName}>`.length;
|
|
2979
|
+
const closeHead = s.indexOf(`</${rootName}`, range.end);
|
|
2980
|
+
if (closeHead === range.end) {
|
|
2981
|
+
let p = closeHead + 2 + rootName.length;
|
|
2982
|
+
while (p < s.length && WHITESPACE_REGEX2.test(s[p])) {
|
|
2983
|
+
p += 1;
|
|
2984
|
+
}
|
|
2985
|
+
if (s[p] === ">") {
|
|
2986
|
+
fullEnd = p + 1;
|
|
2987
|
+
}
|
|
2988
|
+
}
|
|
2989
|
+
if (fullEnd === s.length) {
|
|
2990
|
+
const unwrapped = unwrapJsonSchema(schema);
|
|
2991
|
+
const schemaProps = unwrapped && typeof unwrapped === "object" ? unwrapped.properties : void 0;
|
|
2992
|
+
if (schemaProps && !Object.hasOwn(schemaProps, rootName)) {
|
|
2993
|
+
actualXmlInner = s.slice(range.start, range.end);
|
|
2994
|
+
}
|
|
2995
|
+
}
|
|
2996
|
+
}
|
|
2997
|
+
}
|
|
2998
|
+
}
|
|
2999
|
+
const topLevelStringProps = getTopLevelStringProps(schema);
|
|
3000
|
+
const deepStringTypedProps = getStringTypedProperties(schema);
|
|
3001
|
+
const duplicateKeys = /* @__PURE__ */ new Set();
|
|
3002
|
+
for (const key of topLevelStringProps) {
|
|
3003
|
+
const excludeRanges = [];
|
|
3004
|
+
for (const other of topLevelStringProps) {
|
|
3005
|
+
if (other === key) {
|
|
3006
|
+
continue;
|
|
3007
|
+
}
|
|
3008
|
+
const range = findFirstTopLevelRange(actualXmlInner, other);
|
|
3009
|
+
if (range) {
|
|
3010
|
+
excludeRanges.push(range);
|
|
3011
|
+
}
|
|
3012
|
+
}
|
|
3013
|
+
const occurrences = countTagOccurrences(
|
|
3014
|
+
actualXmlInner,
|
|
3015
|
+
key,
|
|
3016
|
+
excludeRanges,
|
|
3017
|
+
true
|
|
3018
|
+
);
|
|
3019
|
+
if (occurrences > 0 && throwDup) {
|
|
3020
|
+
throw new RXMLDuplicateStringTagError(
|
|
3021
|
+
`Duplicate string tags for <${key}> detected`
|
|
3022
|
+
);
|
|
3023
|
+
}
|
|
3024
|
+
if (occurrences > 0 && !throwDup) {
|
|
3025
|
+
duplicateKeys.add(key);
|
|
3026
|
+
if (options.onError) {
|
|
3027
|
+
options.onError(
|
|
3028
|
+
`RXML: Duplicate string tags for <${key}> detected; using first occurrence.`,
|
|
3029
|
+
{ tag: key, occurrences }
|
|
3030
|
+
);
|
|
3031
|
+
}
|
|
3032
|
+
}
|
|
3033
|
+
}
|
|
3034
|
+
let xmlInnerForParsing = actualXmlInner;
|
|
3035
|
+
const originalContentMap = /* @__PURE__ */ new Map();
|
|
3036
|
+
try {
|
|
3037
|
+
const ranges = [];
|
|
3038
|
+
for (const key of deepStringTypedProps) {
|
|
3039
|
+
const innerRanges = findAllInnerRanges(actualXmlInner, key);
|
|
3040
|
+
for (const r of innerRanges) {
|
|
3041
|
+
if (r.end > r.start) {
|
|
3042
|
+
ranges.push({ ...r, key });
|
|
3043
|
+
}
|
|
3044
|
+
}
|
|
3045
|
+
}
|
|
3046
|
+
if (ranges.length > 0) {
|
|
3047
|
+
const sorted = [...ranges].sort((a, b) => a.start - b.start);
|
|
3048
|
+
let rebuilt = "";
|
|
3049
|
+
let cursor = 0;
|
|
3050
|
+
for (const r of sorted) {
|
|
3051
|
+
if (r.start < cursor) {
|
|
3052
|
+
continue;
|
|
3053
|
+
}
|
|
3054
|
+
if (cursor < r.start) {
|
|
3055
|
+
rebuilt += actualXmlInner.slice(cursor, r.start);
|
|
3056
|
+
}
|
|
3057
|
+
const placeholder = `__RXML_PLACEHOLDER_${r.key}_${r.start}_${r.end}__`;
|
|
3058
|
+
const originalContent = actualXmlInner.slice(r.start, r.end);
|
|
3059
|
+
originalContentMap.set(placeholder, originalContent);
|
|
3060
|
+
rebuilt += placeholder;
|
|
3061
|
+
cursor = r.end;
|
|
3062
|
+
}
|
|
3063
|
+
if (cursor < actualXmlInner.length) {
|
|
3064
|
+
rebuilt += actualXmlInner.slice(cursor);
|
|
3065
|
+
}
|
|
3066
|
+
xmlInnerForParsing = rebuilt;
|
|
3067
|
+
}
|
|
3068
|
+
} catch (error) {
|
|
3069
|
+
if (options.onError) {
|
|
3070
|
+
options.onError(
|
|
3071
|
+
"RXML: Failed to replace string placeholders, falling back to original XML.",
|
|
3072
|
+
{ error }
|
|
3073
|
+
);
|
|
3074
|
+
}
|
|
3075
|
+
xmlInnerForParsing = actualXmlInner;
|
|
3076
|
+
}
|
|
3077
|
+
let parsedNodes;
|
|
3078
|
+
try {
|
|
3079
|
+
const wrappedXml = `<root>${xmlInnerForParsing}</root>`;
|
|
3080
|
+
const tokenizer = new XMLTokenizer(wrappedXml, {
|
|
3081
|
+
...options,
|
|
3082
|
+
textNodeName
|
|
3083
|
+
});
|
|
3084
|
+
const rootNode = tokenizer.parseNode();
|
|
3085
|
+
parsedNodes = rootNode.children;
|
|
3086
|
+
} catch (cause) {
|
|
3087
|
+
throw new RXMLParseError("Failed to parse XML", cause);
|
|
3088
|
+
}
|
|
3089
|
+
const parsedArgs = domToObject(parsedNodes, schema, textNodeName);
|
|
3090
|
+
const restorePlaceholdersDeep = createPlaceholderRestorer(
|
|
3091
|
+
originalContentMap,
|
|
3092
|
+
textNodeName
|
|
3093
|
+
);
|
|
3094
|
+
const parsedArgsRestored = restorePlaceholdersDeep(parsedArgs);
|
|
3095
|
+
const args = {};
|
|
3096
|
+
for (const k of Object.keys(parsedArgsRestored || {})) {
|
|
3097
|
+
const v = parsedArgsRestored[k];
|
|
3098
|
+
let val = v;
|
|
3099
|
+
const propSchema = getPropertySchema(schema, k);
|
|
3100
|
+
const propType = getSchemaType(propSchema);
|
|
3101
|
+
if (propType === "string" && duplicateKeys.has(k) && Array.isArray(v)) {
|
|
3102
|
+
const firstValue = v[0];
|
|
3103
|
+
if (typeof firstValue === "string" && firstValue.startsWith("__RXML_PLACEHOLDER_")) {
|
|
3104
|
+
const originalContent = originalContentMap.get(firstValue);
|
|
3105
|
+
if (originalContent !== void 0) {
|
|
3106
|
+
args[k] = originalContent;
|
|
3107
|
+
continue;
|
|
3108
|
+
}
|
|
3109
|
+
} else {
|
|
3110
|
+
args[k] = firstValue;
|
|
3111
|
+
continue;
|
|
3112
|
+
}
|
|
3113
|
+
}
|
|
3114
|
+
if (propType === "string" && !Array.isArray(v)) {
|
|
3115
|
+
const placeholderUsed = typeof v === "string" && v.startsWith("__RXML_PLACEHOLDER_") || v && typeof v === "object" && Object.hasOwn(v, textNodeName) && typeof v[textNodeName] === "string" && v[textNodeName].startsWith(
|
|
3116
|
+
"__RXML_PLACEHOLDER_"
|
|
3117
|
+
);
|
|
3118
|
+
if (placeholderUsed) {
|
|
3119
|
+
let placeholderKey;
|
|
3120
|
+
if (typeof v === "string") {
|
|
3121
|
+
placeholderKey = v;
|
|
3122
|
+
} else {
|
|
3123
|
+
placeholderKey = v[textNodeName];
|
|
3124
|
+
}
|
|
3125
|
+
const originalContent = originalContentMap.get(placeholderKey);
|
|
3126
|
+
if (originalContent !== void 0) {
|
|
3127
|
+
args[k] = originalContent;
|
|
3128
|
+
continue;
|
|
3129
|
+
}
|
|
3130
|
+
}
|
|
3131
|
+
const raw = extractRawInner(actualXmlInner, k);
|
|
3132
|
+
if (typeof raw === "string") {
|
|
3133
|
+
args[k] = raw;
|
|
3134
|
+
continue;
|
|
3135
|
+
}
|
|
3136
|
+
}
|
|
3137
|
+
if (v && typeof v === "object" && Object.hasOwn(v, textNodeName)) {
|
|
3138
|
+
val = v[textNodeName];
|
|
3139
|
+
}
|
|
3140
|
+
if (Array.isArray(v)) {
|
|
3141
|
+
if (propType === "string") {
|
|
3142
|
+
const mapped = v.map((item) => {
|
|
3143
|
+
if (item && typeof item === "object" && Object.hasOwn(item, textNodeName)) {
|
|
3144
|
+
const textVal = item[textNodeName];
|
|
3145
|
+
return typeof textVal === "string" ? textVal : String(textVal);
|
|
3146
|
+
}
|
|
3147
|
+
return typeof item === "string" ? item : String(item);
|
|
3148
|
+
});
|
|
3149
|
+
if (mapped.length > 1 && throwDup) {
|
|
3150
|
+
throw new RXMLDuplicateStringTagError(
|
|
3151
|
+
`Duplicate string tags for <${k}> detected`
|
|
3152
|
+
);
|
|
3153
|
+
}
|
|
3154
|
+
if (mapped.length > 1 && !throwDup && options.onError) {
|
|
3155
|
+
options.onError(
|
|
3156
|
+
`RXML: Duplicate string tags for <${k}> detected; using first occurrence.`,
|
|
3157
|
+
{ tag: k, occurrences: mapped.length }
|
|
3158
|
+
);
|
|
3159
|
+
}
|
|
3160
|
+
args[k] = (_c = mapped[0]) != null ? _c : "";
|
|
3161
|
+
continue;
|
|
3162
|
+
}
|
|
3163
|
+
val = processArrayContent(v, propSchema, textNodeName);
|
|
3164
|
+
} else if (v && typeof v === "object" && !Object.hasOwn(v, textNodeName)) {
|
|
3165
|
+
const obj = v;
|
|
3166
|
+
const keys2 = Object.keys(obj);
|
|
3167
|
+
if (keys2.length === 1 && keys2[0] === "item") {
|
|
3168
|
+
val = processItemWrapper(obj.item, textNodeName);
|
|
3169
|
+
} else {
|
|
3170
|
+
let isIndexedTuple = false;
|
|
3171
|
+
if (keys2.length > 0 && keys2.every((key) => DIGIT_KEY_REGEX2.test(key))) {
|
|
3172
|
+
const indices = keys2.map((keyStr) => Number.parseInt(keyStr, 10)).sort((a, b) => a - b);
|
|
3173
|
+
isIndexedTuple = indices[0] === 0 && indices.every((indexVal, idx) => indexVal === idx);
|
|
3174
|
+
}
|
|
3175
|
+
if (isIndexedTuple) {
|
|
3176
|
+
val = processIndexedTuple(obj, textNodeName);
|
|
3177
|
+
} else {
|
|
3178
|
+
val = v;
|
|
3179
|
+
}
|
|
3180
|
+
}
|
|
3181
|
+
}
|
|
3182
|
+
args[k] = typeof val === "string" ? val.trim() : val;
|
|
3183
|
+
}
|
|
3184
|
+
for (const key of topLevelStringProps) {
|
|
3185
|
+
if (!Object.hasOwn(args, key)) {
|
|
3186
|
+
const raw = extractRawInner(actualXmlInner, key);
|
|
3187
|
+
if (typeof raw === "string") {
|
|
3188
|
+
args[key] = raw;
|
|
3189
|
+
}
|
|
3190
|
+
}
|
|
3191
|
+
}
|
|
3192
|
+
let dataToCoerce = args;
|
|
3193
|
+
const keys = Object.keys(args);
|
|
3194
|
+
if (keys.length === 1) {
|
|
3195
|
+
const rootKey = keys[0];
|
|
3196
|
+
const rootValue = args[rootKey];
|
|
3197
|
+
const unwrapped = unwrapJsonSchema(schema);
|
|
3198
|
+
if (unwrapped && typeof unwrapped === "object") {
|
|
3199
|
+
const schemaProps = unwrapped.properties;
|
|
3200
|
+
if (schemaProps && !Object.hasOwn(schemaProps, rootKey)) {
|
|
3201
|
+
dataToCoerce = rootValue;
|
|
3202
|
+
}
|
|
3203
|
+
}
|
|
3204
|
+
}
|
|
3205
|
+
try {
|
|
3206
|
+
const coerced = coerceDomBySchema(dataToCoerce, schema);
|
|
3207
|
+
const decoded = deepDecodeStringsBySchema(coerced, schema);
|
|
3208
|
+
return decoded;
|
|
3209
|
+
} catch (error) {
|
|
3210
|
+
throw new RXMLCoercionError("Failed to coerce by schema", error);
|
|
3211
|
+
}
|
|
3212
|
+
}
|
|
3213
|
+
|
|
3214
|
+
// src/rxml/heuristics/engine.ts
|
|
3215
|
+
function applyRawSegmentUpdate(current, result) {
|
|
3216
|
+
if (result.rawSegment !== void 0) {
|
|
3217
|
+
return { ...current, rawSegment: result.rawSegment };
|
|
3218
|
+
}
|
|
3219
|
+
return current;
|
|
3220
|
+
}
|
|
3221
|
+
function applyParsedUpdate(current, result) {
|
|
3222
|
+
if (result.parsed !== void 0) {
|
|
3223
|
+
return { ...current, parsed: result.parsed };
|
|
3224
|
+
}
|
|
3225
|
+
return current;
|
|
3226
|
+
}
|
|
3227
|
+
function applyWarningsUpdate(current, result) {
|
|
3228
|
+
var _a, _b;
|
|
3229
|
+
if (result.warnings && result.warnings.length > 0) {
|
|
3230
|
+
const meta = (_a = current.meta) != null ? _a : {};
|
|
3231
|
+
const existingWarnings = (_b = meta.warnings) != null ? _b : [];
|
|
3232
|
+
return {
|
|
3233
|
+
...current,
|
|
3234
|
+
meta: { ...meta, warnings: [...existingWarnings, ...result.warnings] }
|
|
3235
|
+
};
|
|
3236
|
+
}
|
|
3237
|
+
return current;
|
|
3238
|
+
}
|
|
3239
|
+
function attemptReparse(current, result, reparseCount, maxReparses, parse4) {
|
|
3240
|
+
if (!result.reparse || result.rawSegment === void 0 || reparseCount >= maxReparses) {
|
|
3241
|
+
return { state: current, newCount: reparseCount };
|
|
3242
|
+
}
|
|
3243
|
+
try {
|
|
3244
|
+
const reparsed = parse4(result.rawSegment, current.schema);
|
|
3245
|
+
return {
|
|
3246
|
+
state: { ...current, parsed: reparsed, errors: [] },
|
|
3247
|
+
newCount: reparseCount + 1
|
|
3248
|
+
};
|
|
3249
|
+
} catch (error) {
|
|
3250
|
+
return {
|
|
3251
|
+
state: { ...current, errors: [...current.errors, error] },
|
|
3252
|
+
newCount: reparseCount + 1
|
|
3253
|
+
};
|
|
3254
|
+
}
|
|
3255
|
+
}
|
|
3256
|
+
function executePhase(ctx, heuristics, options) {
|
|
3257
|
+
var _a;
|
|
3258
|
+
let current = ctx;
|
|
3259
|
+
let reparseCount = 0;
|
|
3260
|
+
const maxReparses = (_a = options.maxReparses) != null ? _a : 2;
|
|
3261
|
+
for (const heuristic of heuristics) {
|
|
3262
|
+
if (!heuristic.applies(current)) {
|
|
3263
|
+
continue;
|
|
3264
|
+
}
|
|
3265
|
+
const result = heuristic.run(current);
|
|
3266
|
+
current = applyRawSegmentUpdate(current, result);
|
|
3267
|
+
current = applyParsedUpdate(current, result);
|
|
3268
|
+
current = applyWarningsUpdate(current, result);
|
|
3269
|
+
const reparseResult = attemptReparse(
|
|
3270
|
+
current,
|
|
3271
|
+
result,
|
|
3272
|
+
reparseCount,
|
|
3273
|
+
maxReparses,
|
|
3274
|
+
options.parse
|
|
3275
|
+
);
|
|
3276
|
+
current = reparseResult.state;
|
|
3277
|
+
reparseCount = reparseResult.newCount;
|
|
3278
|
+
if (result.stop) {
|
|
3279
|
+
break;
|
|
3280
|
+
}
|
|
3281
|
+
}
|
|
3282
|
+
return current;
|
|
3283
|
+
}
|
|
3284
|
+
function applyHeuristicPipeline(ctx, config, options) {
|
|
3285
|
+
let current = ctx;
|
|
3286
|
+
if (config.preParse && config.preParse.length > 0) {
|
|
3287
|
+
current = executePhase(current, config.preParse, options);
|
|
3288
|
+
}
|
|
3289
|
+
if (current.parsed === null && current.errors.length === 0) {
|
|
3290
|
+
try {
|
|
3291
|
+
const parsed = options.parse(current.rawSegment, current.schema);
|
|
3292
|
+
current = { ...current, parsed, errors: [] };
|
|
3293
|
+
} catch (error) {
|
|
3294
|
+
current = { ...current, errors: [error] };
|
|
3295
|
+
}
|
|
3296
|
+
}
|
|
3297
|
+
if (current.errors.length > 0 && config.fallbackReparse && config.fallbackReparse.length > 0) {
|
|
3298
|
+
current = executePhase(current, config.fallbackReparse, options);
|
|
3299
|
+
}
|
|
3300
|
+
if (current.parsed !== null && config.postParse && config.postParse.length > 0) {
|
|
3301
|
+
current = executePhase(current, config.postParse, options);
|
|
3302
|
+
}
|
|
3303
|
+
return current;
|
|
3304
|
+
}
|
|
3305
|
+
function createIntermediateCall(toolName, rawSegment, schema) {
|
|
3306
|
+
return {
|
|
3307
|
+
toolName,
|
|
3308
|
+
schema,
|
|
3309
|
+
rawSegment,
|
|
3310
|
+
parsed: null,
|
|
3311
|
+
errors: [],
|
|
3312
|
+
meta: { originalContent: rawSegment }
|
|
3313
|
+
};
|
|
3314
|
+
}
|
|
3315
|
+
|
|
3316
|
+
// src/rxml/heuristics/xml-defaults.ts
|
|
3317
|
+
var MALFORMED_CLOSE_RE_G = /<\/\s+([A-Za-z0-9_:-]+)\s*>/g;
|
|
3318
|
+
var MALFORMED_CLOSE_RE = /<\/\s+([A-Za-z0-9_:-]+)\s*>/;
|
|
3319
|
+
var STATUS_TO_STEP_BOUNDARY_RE = /<\/status>\s*<step>/g;
|
|
3320
|
+
var WHITESPACE_REGEX3 = /\s/;
|
|
3321
|
+
var NAME_CHAR_RE = /[A-Za-z0-9_:-]/;
|
|
3322
|
+
var NAME_START_CHAR_RE = /[A-Za-z_:]/;
|
|
3323
|
+
var STEP_TAG_RE = /<step>([\s\S]*?)<\/step>/i;
|
|
3324
|
+
var STATUS_TAG_RE = /<status>([\s\S]*?)<\/status>/i;
|
|
3325
|
+
var normalizeCloseTagsHeuristic = {
|
|
3326
|
+
id: "normalize-close-tags",
|
|
3327
|
+
phase: "pre-parse",
|
|
3328
|
+
applies: () => true,
|
|
3329
|
+
run: (ctx) => {
|
|
3330
|
+
const normalized = ctx.rawSegment.replace(MALFORMED_CLOSE_RE_G, "</$1>");
|
|
3331
|
+
if (normalized !== ctx.rawSegment) {
|
|
3332
|
+
return { rawSegment: normalized };
|
|
3333
|
+
}
|
|
3334
|
+
return {};
|
|
3335
|
+
}
|
|
3336
|
+
};
|
|
3337
|
+
var escapeInvalidLtHeuristic = {
|
|
3338
|
+
id: "escape-invalid-lt",
|
|
3339
|
+
phase: "pre-parse",
|
|
3340
|
+
applies: () => true,
|
|
3341
|
+
run: (ctx) => {
|
|
3342
|
+
const escaped = escapeInvalidLt(ctx.rawSegment);
|
|
3343
|
+
if (escaped !== ctx.rawSegment) {
|
|
3344
|
+
return { rawSegment: escaped };
|
|
3345
|
+
}
|
|
3346
|
+
return {};
|
|
3347
|
+
}
|
|
3348
|
+
};
|
|
3349
|
+
var balanceTagsHeuristic = {
|
|
3350
|
+
id: "balance-tags",
|
|
3351
|
+
phase: "fallback-reparse",
|
|
3352
|
+
applies: (ctx) => {
|
|
3353
|
+
var _a;
|
|
3354
|
+
const original = ((_a = ctx.meta) == null ? void 0 : _a.originalContent) || ctx.rawSegment;
|
|
3355
|
+
const normalized = original.replace(MALFORMED_CLOSE_RE_G, "</$1>");
|
|
3356
|
+
const balanced = balanceTags(original);
|
|
3357
|
+
const hasMalformedClose = MALFORMED_CLOSE_RE.test(original);
|
|
3358
|
+
if (!hasMalformedClose && balanced.length > normalized.length && ctx.errors.length === 0) {
|
|
3359
|
+
return false;
|
|
3360
|
+
}
|
|
3361
|
+
return balanced !== normalized;
|
|
3362
|
+
},
|
|
3363
|
+
run: (ctx) => {
|
|
3364
|
+
var _a;
|
|
3365
|
+
const original = ((_a = ctx.meta) == null ? void 0 : _a.originalContent) || ctx.rawSegment;
|
|
3366
|
+
const balanced = balanceTags(original);
|
|
3367
|
+
const escaped = escapeInvalidLt(balanced);
|
|
3368
|
+
return { rawSegment: escaped, reparse: true };
|
|
3369
|
+
}
|
|
3370
|
+
};
|
|
3371
|
+
var dedupeShellStringTagsHeuristic = {
|
|
3372
|
+
id: "dedupe-shell-string-tags",
|
|
3373
|
+
phase: "fallback-reparse",
|
|
3374
|
+
applies: (ctx) => shouldDeduplicateStringTags(ctx.schema),
|
|
3375
|
+
run: (ctx) => {
|
|
3376
|
+
const names = getStringPropertyNames(ctx.schema);
|
|
3377
|
+
let deduped = ctx.rawSegment;
|
|
3378
|
+
for (const key of names) {
|
|
3379
|
+
deduped = dedupeSingleTag(deduped, key);
|
|
3380
|
+
}
|
|
3381
|
+
if (deduped !== ctx.rawSegment) {
|
|
3382
|
+
return { rawSegment: deduped, reparse: true };
|
|
3383
|
+
}
|
|
3384
|
+
return {};
|
|
3385
|
+
}
|
|
3386
|
+
};
|
|
3387
|
+
var repairAgainstSchemaHeuristic = {
|
|
3388
|
+
id: "repair-against-schema",
|
|
3389
|
+
phase: "post-parse",
|
|
3390
|
+
applies: (ctx) => ctx.parsed !== null && typeof ctx.parsed === "object",
|
|
3391
|
+
run: (ctx) => {
|
|
3392
|
+
const repaired = repairParsedAgainstSchema(ctx.parsed, ctx.schema);
|
|
3393
|
+
if (repaired !== ctx.parsed) {
|
|
3394
|
+
return { parsed: repaired };
|
|
3395
|
+
}
|
|
3396
|
+
return {};
|
|
3397
|
+
}
|
|
3398
|
+
};
|
|
3399
|
+
var defaultPipelineConfig = {
|
|
3400
|
+
preParse: [normalizeCloseTagsHeuristic, escapeInvalidLtHeuristic],
|
|
3401
|
+
fallbackReparse: [balanceTagsHeuristic, dedupeShellStringTagsHeuristic],
|
|
3402
|
+
postParse: [repairAgainstSchemaHeuristic]
|
|
3403
|
+
};
|
|
3404
|
+
var INDEX_TAG_RE = /^<(\d+)(?:>|\/?>)/;
|
|
3405
|
+
function isIndexTagAt(xml, pos) {
|
|
3406
|
+
const remaining = xml.slice(pos);
|
|
3407
|
+
return INDEX_TAG_RE.test(remaining);
|
|
3408
|
+
}
|
|
3409
|
+
function escapeInvalidLt(xml) {
|
|
3410
|
+
const len = xml.length;
|
|
3411
|
+
let out = "";
|
|
3412
|
+
for (let i = 0; i < len; i += 1) {
|
|
3413
|
+
const ch = xml[i];
|
|
3414
|
+
if (ch === "<") {
|
|
3415
|
+
const next = i + 1 < len ? xml[i + 1] : "";
|
|
3416
|
+
const isValidStart = NAME_START_CHAR_RE.test(next) || next === "/" || next === "!" || next === "?";
|
|
3417
|
+
const isIndexTag = !isValidStart && isIndexTagAt(xml, i);
|
|
3418
|
+
if (!(isValidStart || isIndexTag)) {
|
|
3419
|
+
out += "<";
|
|
3420
|
+
continue;
|
|
3421
|
+
}
|
|
3422
|
+
}
|
|
3423
|
+
out += ch;
|
|
3424
|
+
}
|
|
3425
|
+
return out;
|
|
3426
|
+
}
|
|
3427
|
+
function balanceTags(xml) {
|
|
3428
|
+
const src = xml.replace(MALFORMED_CLOSE_RE_G, "</$1>").replace(STATUS_TO_STEP_BOUNDARY_RE, "</status></step><step>");
|
|
3429
|
+
let i = 0;
|
|
3430
|
+
const len = src.length;
|
|
3431
|
+
const out = [];
|
|
3432
|
+
const stack = [];
|
|
3433
|
+
while (i < len) {
|
|
3434
|
+
const lt = src.indexOf("<", i);
|
|
3435
|
+
if (lt === -1) {
|
|
3436
|
+
out.push(src.slice(i));
|
|
3437
|
+
break;
|
|
3438
|
+
}
|
|
3439
|
+
out.push(src.slice(i, lt));
|
|
3440
|
+
if (lt + 1 >= len) {
|
|
3441
|
+
break;
|
|
3442
|
+
}
|
|
3443
|
+
const next = src[lt + 1];
|
|
3444
|
+
if (next === "!" || next === "?") {
|
|
3445
|
+
i = handleSpecialTagSegment(src, lt, out);
|
|
3446
|
+
continue;
|
|
3447
|
+
}
|
|
3448
|
+
if (next === "/") {
|
|
3449
|
+
i = handleClosingTagSegment(src, lt, out, stack);
|
|
3450
|
+
continue;
|
|
3451
|
+
}
|
|
3452
|
+
i = handleOpeningTagSegment(src, lt, out, stack);
|
|
3453
|
+
}
|
|
3454
|
+
for (let k = stack.length - 1; k >= 0; k -= 1) {
|
|
3455
|
+
out.push(`</${stack[k]}>`);
|
|
3456
|
+
}
|
|
3457
|
+
return out.join("");
|
|
3458
|
+
}
|
|
3459
|
+
function skipWs(s, p, len) {
|
|
3460
|
+
let idx = p;
|
|
3461
|
+
while (idx < len && WHITESPACE_REGEX3.test(s[idx])) {
|
|
3462
|
+
idx += 1;
|
|
3463
|
+
}
|
|
3464
|
+
return idx;
|
|
3465
|
+
}
|
|
3466
|
+
function parseTagNameAt(s, p, len) {
|
|
3467
|
+
let idx = p;
|
|
3468
|
+
const start = idx;
|
|
3469
|
+
while (idx < len && NAME_CHAR_RE.test(s[idx])) {
|
|
3470
|
+
idx += 1;
|
|
3471
|
+
}
|
|
3472
|
+
return { name: s.slice(start, idx), pos: idx };
|
|
3473
|
+
}
|
|
3474
|
+
function handleSpecialTagSegment(src, lt, out) {
|
|
3475
|
+
const gt = src.indexOf(">", lt + 1);
|
|
3476
|
+
if (gt === -1) {
|
|
3477
|
+
out.push(src.slice(lt));
|
|
3478
|
+
return src.length;
|
|
3479
|
+
}
|
|
3480
|
+
out.push(src.slice(lt, gt + 1));
|
|
3481
|
+
return gt + 1;
|
|
3482
|
+
}
|
|
3483
|
+
function handleClosingTagSegment(src, lt, out, stack) {
|
|
3484
|
+
const len = src.length;
|
|
3485
|
+
let p = skipWs(src, lt + 2, len);
|
|
3486
|
+
const { name, pos } = parseTagNameAt(src, p, len);
|
|
3487
|
+
p = pos;
|
|
3488
|
+
const gt = src.indexOf(">", p);
|
|
3489
|
+
const closingText = gt === -1 ? src.slice(lt) : src.slice(lt, gt + 1);
|
|
3490
|
+
const idx = stack.lastIndexOf(name);
|
|
3491
|
+
if (idx !== -1) {
|
|
3492
|
+
for (let k = stack.length - 1; k > idx; k -= 1) {
|
|
3493
|
+
out.push(`</${stack[k]}>`);
|
|
3494
|
+
stack.pop();
|
|
3495
|
+
}
|
|
3496
|
+
out.push(closingText);
|
|
3497
|
+
stack.pop();
|
|
3498
|
+
}
|
|
3499
|
+
return gt === -1 ? len : gt + 1;
|
|
3500
|
+
}
|
|
3501
|
+
function handleOpeningTagSegment(src, lt, out, stack) {
|
|
3502
|
+
const len = src.length;
|
|
3503
|
+
let p = skipWs(src, lt + 1, len);
|
|
3504
|
+
const nameStart = p;
|
|
3505
|
+
const parsed = parseTagNameAt(src, p, len);
|
|
3506
|
+
p = parsed.pos;
|
|
3507
|
+
const name = src.slice(nameStart, p);
|
|
3508
|
+
const q = src.indexOf(">", p);
|
|
3509
|
+
if (q === -1) {
|
|
3510
|
+
out.push(src.slice(lt));
|
|
3511
|
+
return len;
|
|
3512
|
+
}
|
|
3513
|
+
let r = q - 1;
|
|
3514
|
+
while (r >= nameStart && WHITESPACE_REGEX3.test(src[r])) {
|
|
3515
|
+
r -= 1;
|
|
3516
|
+
}
|
|
3517
|
+
const selfClosing = src[r] === "/";
|
|
3518
|
+
out.push(src.slice(lt, q + 1));
|
|
3519
|
+
if (!selfClosing && name) {
|
|
3520
|
+
stack.push(name);
|
|
3521
|
+
}
|
|
3522
|
+
return q + 1;
|
|
3523
|
+
}
|
|
3524
|
+
function extractSchemaProperties(schema) {
|
|
3525
|
+
const unwrapped = unwrapJsonSchema(schema);
|
|
3526
|
+
if (!unwrapped || typeof unwrapped !== "object") {
|
|
3527
|
+
return void 0;
|
|
330
3528
|
}
|
|
331
|
-
|
|
332
|
-
phase: "stream",
|
|
333
|
-
reason: "Incomplete streaming tool call segment emitted as text",
|
|
334
|
-
snippet: `${toolCallStart}${state.currentToolCallJson}`
|
|
335
|
-
});
|
|
336
|
-
const errorId = generateId();
|
|
337
|
-
const errorContent = `${toolCallStart}${state.currentToolCallJson}`;
|
|
338
|
-
controller.enqueue({
|
|
339
|
-
type: "text-start",
|
|
340
|
-
id: errorId
|
|
341
|
-
});
|
|
342
|
-
controller.enqueue({
|
|
343
|
-
type: "text-delta",
|
|
344
|
-
id: errorId,
|
|
345
|
-
delta: errorContent
|
|
346
|
-
});
|
|
347
|
-
controller.enqueue({
|
|
348
|
-
type: "text-end",
|
|
349
|
-
id: errorId
|
|
350
|
-
});
|
|
351
|
-
state.currentToolCallJson = "";
|
|
3529
|
+
return unwrapped.properties;
|
|
352
3530
|
}
|
|
353
|
-
function
|
|
354
|
-
|
|
355
|
-
|
|
3531
|
+
function shouldDeduplicateStringTags(schema) {
|
|
3532
|
+
const props = extractSchemaProperties(schema);
|
|
3533
|
+
if (!props) {
|
|
3534
|
+
return false;
|
|
356
3535
|
}
|
|
357
|
-
|
|
358
|
-
|
|
359
|
-
|
|
3536
|
+
const commandRaw = props.command;
|
|
3537
|
+
if (!commandRaw) {
|
|
3538
|
+
return false;
|
|
3539
|
+
}
|
|
3540
|
+
const command = unwrapJsonSchema(commandRaw);
|
|
3541
|
+
return (command == null ? void 0 : command.type) === "array";
|
|
360
3542
|
}
|
|
361
|
-
function
|
|
362
|
-
|
|
363
|
-
|
|
364
|
-
|
|
365
|
-
}
|
|
366
|
-
|
|
367
|
-
|
|
368
|
-
|
|
369
|
-
|
|
370
|
-
|
|
371
|
-
});
|
|
372
|
-
state.hasEmittedTextStart = true;
|
|
3543
|
+
function getStringPropertyNames(schema) {
|
|
3544
|
+
const props = extractSchemaProperties(schema);
|
|
3545
|
+
if (!props) {
|
|
3546
|
+
return [];
|
|
3547
|
+
}
|
|
3548
|
+
const names = [];
|
|
3549
|
+
for (const key of Object.keys(props)) {
|
|
3550
|
+
const prop = unwrapJsonSchema(props[key]);
|
|
3551
|
+
if ((prop == null ? void 0 : prop.type) === "string") {
|
|
3552
|
+
names.push(key);
|
|
373
3553
|
}
|
|
374
|
-
controller.enqueue({
|
|
375
|
-
type: "text-delta",
|
|
376
|
-
id: state.currentTextId,
|
|
377
|
-
delta: text
|
|
378
|
-
});
|
|
379
3554
|
}
|
|
3555
|
+
return names;
|
|
380
3556
|
}
|
|
381
|
-
function
|
|
3557
|
+
function escapeRegExp2(s) {
|
|
3558
|
+
return s.replace(/[.*+?^${}()|[\]\\]/g, "\\$&");
|
|
3559
|
+
}
|
|
3560
|
+
function dedupeSingleTag(xml, key) {
|
|
382
3561
|
var _a, _b;
|
|
383
|
-
const
|
|
384
|
-
|
|
385
|
-
|
|
386
|
-
|
|
387
|
-
|
|
388
|
-
|
|
389
|
-
|
|
390
|
-
|
|
391
|
-
|
|
392
|
-
|
|
393
|
-
|
|
394
|
-
|
|
395
|
-
|
|
396
|
-
|
|
397
|
-
|
|
398
|
-
|
|
399
|
-
});
|
|
400
|
-
const errorId = generateId();
|
|
401
|
-
const errorContent = `${toolCallStart}${state.currentToolCallJson}${toolCallEnd}`;
|
|
402
|
-
controller.enqueue({
|
|
403
|
-
type: "text-start",
|
|
404
|
-
id: errorId
|
|
405
|
-
});
|
|
406
|
-
controller.enqueue({
|
|
407
|
-
type: "text-delta",
|
|
408
|
-
id: errorId,
|
|
409
|
-
delta: errorContent
|
|
410
|
-
});
|
|
411
|
-
controller.enqueue({
|
|
412
|
-
type: "text-end",
|
|
413
|
-
id: errorId
|
|
414
|
-
});
|
|
415
|
-
(_b = options == null ? void 0 : options.onError) == null ? void 0 : _b.call(
|
|
416
|
-
options,
|
|
417
|
-
"Could not process streaming JSON tool call; emitting original text.",
|
|
418
|
-
{
|
|
419
|
-
toolCall: errorContent
|
|
420
|
-
}
|
|
421
|
-
);
|
|
3562
|
+
const escaped = escapeRegExp2(key);
|
|
3563
|
+
const re = new RegExp(`<${escaped}>([\\s\\S]*?)<\\/${escaped}>`, "g");
|
|
3564
|
+
const matches = Array.from(xml.matchAll(re));
|
|
3565
|
+
if (matches.length <= 1) {
|
|
3566
|
+
return xml;
|
|
3567
|
+
}
|
|
3568
|
+
const last = matches.at(-1);
|
|
3569
|
+
let result = "";
|
|
3570
|
+
let cursor = 0;
|
|
3571
|
+
for (const m of matches) {
|
|
3572
|
+
const idx = (_a = m.index) != null ? _a : 0;
|
|
3573
|
+
result += xml.slice(cursor, idx);
|
|
3574
|
+
if (last && idx === ((_b = last.index) != null ? _b : -1)) {
|
|
3575
|
+
result += m[0];
|
|
3576
|
+
}
|
|
3577
|
+
cursor = idx + m[0].length;
|
|
422
3578
|
}
|
|
3579
|
+
result += xml.slice(cursor);
|
|
3580
|
+
return result;
|
|
423
3581
|
}
|
|
424
|
-
function
|
|
425
|
-
|
|
426
|
-
|
|
427
|
-
|
|
428
|
-
|
|
429
|
-
|
|
430
|
-
|
|
431
|
-
state.currentToolCallJson = "";
|
|
432
|
-
state.isInsideToolCall = true;
|
|
3582
|
+
function repairParsedAgainstSchema(input, schema) {
|
|
3583
|
+
if (!input || typeof input !== "object") {
|
|
3584
|
+
return input;
|
|
3585
|
+
}
|
|
3586
|
+
const properties = extractSchemaProperties(schema);
|
|
3587
|
+
if (!properties) {
|
|
3588
|
+
return input;
|
|
433
3589
|
}
|
|
3590
|
+
applySchemaProps(input, properties);
|
|
3591
|
+
return input;
|
|
434
3592
|
}
|
|
435
|
-
function
|
|
436
|
-
const
|
|
437
|
-
|
|
438
|
-
|
|
439
|
-
|
|
440
|
-
|
|
441
|
-
|
|
442
|
-
|
|
443
|
-
|
|
444
|
-
|
|
3593
|
+
function applySchemaProps(obj, properties) {
|
|
3594
|
+
for (const key of Object.keys(obj)) {
|
|
3595
|
+
const propSchema = properties[key];
|
|
3596
|
+
if (!propSchema) {
|
|
3597
|
+
continue;
|
|
3598
|
+
}
|
|
3599
|
+
const prop = unwrapJsonSchema(propSchema);
|
|
3600
|
+
if ((prop == null ? void 0 : prop.type) === "array" && prop.items) {
|
|
3601
|
+
const itemSchema = unwrapJsonSchema(prop.items);
|
|
3602
|
+
obj[key] = coerceArrayItems(obj[key], itemSchema);
|
|
3603
|
+
continue;
|
|
3604
|
+
}
|
|
3605
|
+
if ((prop == null ? void 0 : prop.type) === "object") {
|
|
3606
|
+
const val = obj[key];
|
|
3607
|
+
if (val && typeof val === "object") {
|
|
3608
|
+
obj[key] = repairParsedAgainstSchema(val, prop);
|
|
3609
|
+
}
|
|
445
3610
|
}
|
|
446
|
-
publishText(state.buffer.slice(0, startIndex), state, controller);
|
|
447
|
-
state.buffer = state.buffer.slice(startIndex + tag.length);
|
|
448
|
-
processTagMatch(context);
|
|
449
|
-
startIndex = getPotentialStartIndex(
|
|
450
|
-
state.buffer,
|
|
451
|
-
state.isInsideToolCall ? toolCallEnd : toolCallStart
|
|
452
|
-
);
|
|
453
3611
|
}
|
|
454
3612
|
}
|
|
455
|
-
function
|
|
456
|
-
if (
|
|
457
|
-
return;
|
|
458
|
-
}
|
|
459
|
-
const potentialIndex = getPotentialStartIndex(state.buffer, toolCallStart);
|
|
460
|
-
if (potentialIndex != null && potentialIndex + toolCallStart.length > state.buffer.length) {
|
|
461
|
-
publishText(state.buffer.slice(0, potentialIndex), state, controller);
|
|
462
|
-
state.buffer = state.buffer.slice(potentialIndex);
|
|
463
|
-
} else {
|
|
464
|
-
publishText(state.buffer, state, controller);
|
|
465
|
-
state.buffer = "";
|
|
3613
|
+
function coerceArrayItems(val, itemSchema) {
|
|
3614
|
+
if (!Array.isArray(val)) {
|
|
3615
|
+
return val;
|
|
466
3616
|
}
|
|
3617
|
+
return val.map((v) => coerceArrayItem(v, itemSchema));
|
|
467
3618
|
}
|
|
468
|
-
|
|
469
|
-
|
|
470
|
-
|
|
471
|
-
|
|
472
|
-
|
|
473
|
-
|
|
474
|
-
toolSystemPromptTemplate
|
|
475
|
-
}) {
|
|
476
|
-
return toolSystemPromptTemplate(tools || []);
|
|
477
|
-
},
|
|
478
|
-
formatToolCall(toolCall) {
|
|
479
|
-
let args = {};
|
|
480
|
-
if (toolCall.input != null) {
|
|
481
|
-
try {
|
|
482
|
-
args = JSON.parse(toolCall.input);
|
|
483
|
-
} catch (e) {
|
|
484
|
-
args = toolCall.input;
|
|
485
|
-
}
|
|
3619
|
+
function coerceArrayItem(v, itemSchema) {
|
|
3620
|
+
const itemType = itemSchema == null ? void 0 : itemSchema.type;
|
|
3621
|
+
if (typeof v === "string" && itemType === "object") {
|
|
3622
|
+
const parsed = tryParseStringToSchemaObject(v, itemSchema);
|
|
3623
|
+
if (parsed !== null) {
|
|
3624
|
+
return parsed;
|
|
486
3625
|
}
|
|
487
|
-
|
|
488
|
-
|
|
489
|
-
arguments: args
|
|
490
|
-
})}${toolCallEnd}`;
|
|
491
|
-
},
|
|
492
|
-
parseGeneratedText({
|
|
493
|
-
text,
|
|
494
|
-
options
|
|
495
|
-
}) {
|
|
496
|
-
const startEsc = escapeRegExp(toolCallStart);
|
|
497
|
-
const endEsc = escapeRegExp(toolCallEnd);
|
|
498
|
-
const toolCallRegex = new RegExp(
|
|
499
|
-
`${startEsc}([\0-\uFFFF]*?)${endEsc}`,
|
|
500
|
-
"gs"
|
|
3626
|
+
const fallback = extractStepStatusFromString(
|
|
3627
|
+
v.replace(MALFORMED_CLOSE_RE_G, "</$1>")
|
|
501
3628
|
);
|
|
502
|
-
|
|
503
|
-
|
|
504
|
-
let match = toolCallRegex.exec(text);
|
|
505
|
-
while (match !== null) {
|
|
506
|
-
currentIndex = processMatchedToolCall({
|
|
507
|
-
match,
|
|
508
|
-
text,
|
|
509
|
-
currentIndex,
|
|
510
|
-
processedElements,
|
|
511
|
-
options
|
|
512
|
-
});
|
|
513
|
-
match = toolCallRegex.exec(text);
|
|
514
|
-
}
|
|
515
|
-
if (currentIndex < text.length) {
|
|
516
|
-
const remainingText = text.substring(currentIndex);
|
|
517
|
-
addTextSegment(remainingText, processedElements);
|
|
518
|
-
}
|
|
519
|
-
return processedElements;
|
|
520
|
-
},
|
|
521
|
-
createStreamParser({
|
|
522
|
-
options
|
|
523
|
-
}) {
|
|
524
|
-
const state = {
|
|
525
|
-
isInsideToolCall: false,
|
|
526
|
-
buffer: "",
|
|
527
|
-
currentToolCallJson: "",
|
|
528
|
-
currentTextId: null,
|
|
529
|
-
hasEmittedTextStart: false
|
|
530
|
-
};
|
|
531
|
-
return new TransformStream({
|
|
532
|
-
transform(chunk, controller) {
|
|
533
|
-
var _a;
|
|
534
|
-
if (chunk.type === "finish") {
|
|
535
|
-
handleFinishChunk(state, controller, toolCallStart, chunk);
|
|
536
|
-
return;
|
|
537
|
-
}
|
|
538
|
-
if (chunk.type !== "text-delta") {
|
|
539
|
-
controller.enqueue(chunk);
|
|
540
|
-
return;
|
|
541
|
-
}
|
|
542
|
-
const textContent = (_a = chunk.delta) != null ? _a : "";
|
|
543
|
-
state.buffer += textContent;
|
|
544
|
-
processBufferTags({
|
|
545
|
-
state,
|
|
546
|
-
controller,
|
|
547
|
-
toolCallStart,
|
|
548
|
-
toolCallEnd,
|
|
549
|
-
options
|
|
550
|
-
});
|
|
551
|
-
handlePartialTag(state, controller, toolCallStart);
|
|
552
|
-
}
|
|
553
|
-
});
|
|
554
|
-
},
|
|
555
|
-
extractToolCallSegments({ text }) {
|
|
556
|
-
const startEsc = escapeRegExp(toolCallStart);
|
|
557
|
-
const endEsc = escapeRegExp(toolCallEnd);
|
|
558
|
-
const regex = new RegExp(`${startEsc}([\0-\uFFFF]*?)${endEsc}`, "gs");
|
|
559
|
-
const segments = [];
|
|
560
|
-
let m = regex.exec(text);
|
|
561
|
-
while (m != null) {
|
|
562
|
-
segments.push(m[0]);
|
|
563
|
-
m = regex.exec(text);
|
|
3629
|
+
if (fallback) {
|
|
3630
|
+
return fallback;
|
|
564
3631
|
}
|
|
565
|
-
return
|
|
3632
|
+
return v;
|
|
566
3633
|
}
|
|
567
|
-
|
|
568
|
-
|
|
569
|
-
|
|
570
|
-
|
|
571
|
-
return typeof protocol === "function";
|
|
3634
|
+
if (v && typeof v === "object" && itemType === "object") {
|
|
3635
|
+
return repairParsedAgainstSchema(v, itemSchema);
|
|
3636
|
+
}
|
|
3637
|
+
return v;
|
|
572
3638
|
}
|
|
573
|
-
function
|
|
574
|
-
|
|
3639
|
+
function tryParseStringToSchemaObject(xml, itemSchema) {
|
|
3640
|
+
try {
|
|
3641
|
+
const normalized = xml.replace(MALFORMED_CLOSE_RE_G, "</$1>");
|
|
3642
|
+
const fixed = parse2(normalized, itemSchema, { noChildNodes: [] });
|
|
3643
|
+
return typeof fixed === "string" ? null : fixed;
|
|
3644
|
+
} catch (e) {
|
|
3645
|
+
return null;
|
|
3646
|
+
}
|
|
3647
|
+
}
|
|
3648
|
+
function extractStepStatusFromString(normXml) {
|
|
3649
|
+
const stepMatch = normXml.match(STEP_TAG_RE);
|
|
3650
|
+
const statusMatch = normXml.match(STATUS_TAG_RE);
|
|
3651
|
+
if (stepMatch && statusMatch) {
|
|
3652
|
+
return { step: stepMatch[1], status: statusMatch[1] };
|
|
3653
|
+
}
|
|
3654
|
+
return null;
|
|
3655
|
+
}
|
|
3656
|
+
|
|
3657
|
+
// src/rxml/parse.ts
|
|
3658
|
+
function parse3(xml, schema, options = {}) {
|
|
3659
|
+
if (!options.repair) {
|
|
3660
|
+
return parse2(xml, schema, options);
|
|
3661
|
+
}
|
|
3662
|
+
const baseOptions = {
|
|
3663
|
+
...options,
|
|
3664
|
+
repair: false
|
|
3665
|
+
};
|
|
3666
|
+
const ctx = createIntermediateCall("", xml, schema);
|
|
3667
|
+
const result = applyHeuristicPipeline(ctx, defaultPipelineConfig, {
|
|
3668
|
+
parse: (raw, s) => parse2(raw, s, baseOptions),
|
|
3669
|
+
onError: options.onError,
|
|
3670
|
+
maxReparses: options.maxReparses
|
|
3671
|
+
});
|
|
3672
|
+
if (result.parsed !== null) {
|
|
3673
|
+
return result.parsed;
|
|
3674
|
+
}
|
|
3675
|
+
const error = result.errors[0];
|
|
3676
|
+
throw new RXMLParseError("Failed to parse XML with repair heuristics", error);
|
|
575
3677
|
}
|
|
576
3678
|
|
|
577
3679
|
// src/core/protocols/xml-protocol.ts
|
|
578
|
-
var
|
|
579
|
-
var
|
|
580
|
-
var WHITESPACE_REGEX = /\s/;
|
|
3680
|
+
var NAME_CHAR_RE2 = /[A-Za-z0-9_:-]/;
|
|
3681
|
+
var WHITESPACE_REGEX4 = /\s/;
|
|
581
3682
|
function getToolSchema(tools, toolName) {
|
|
582
3683
|
var _a;
|
|
583
3684
|
return (_a = tools.find((t) => t.name === toolName)) == null ? void 0 : _a.inputSchema;
|
|
@@ -591,7 +3692,7 @@ function processToolCall(params) {
|
|
|
591
3692
|
onError: (_a = options == null ? void 0 : options.onError) != null ? _a : parseOptions == null ? void 0 : parseOptions.onError
|
|
592
3693
|
};
|
|
593
3694
|
try {
|
|
594
|
-
const parsed = (
|
|
3695
|
+
const parsed = parse3(toolCall.content, toolSchema, parseConfig);
|
|
595
3696
|
processedElements.push({
|
|
596
3697
|
type: "tool-call",
|
|
597
3698
|
toolCallId: generateId(),
|
|
@@ -629,7 +3730,7 @@ function handleStreamingToolCallEnd(params) {
|
|
|
629
3730
|
};
|
|
630
3731
|
flushText(ctrl);
|
|
631
3732
|
try {
|
|
632
|
-
const parsedResult = (
|
|
3733
|
+
const parsedResult = parse3(toolContent, toolSchema, parseConfig);
|
|
633
3734
|
ctrl.enqueue({
|
|
634
3735
|
type: "tool-call",
|
|
635
3736
|
toolCallId: generateId(),
|
|
@@ -679,11 +3780,11 @@ function consumeClosingTag(text, lt) {
|
|
|
679
3780
|
}
|
|
680
3781
|
function consumeOpenTag(text, lt) {
|
|
681
3782
|
let p = lt + 1;
|
|
682
|
-
while (p < text.length &&
|
|
3783
|
+
while (p < text.length && WHITESPACE_REGEX4.test(text[p])) {
|
|
683
3784
|
p += 1;
|
|
684
3785
|
}
|
|
685
3786
|
const nameStart = p;
|
|
686
|
-
while (p < text.length &&
|
|
3787
|
+
while (p < text.length && NAME_CHAR_RE2.test(text.charAt(p))) {
|
|
687
3788
|
p += 1;
|
|
688
3789
|
}
|
|
689
3790
|
const name = text.slice(nameStart, p);
|
|
@@ -692,7 +3793,7 @@ function consumeOpenTag(text, lt) {
|
|
|
692
3793
|
return null;
|
|
693
3794
|
}
|
|
694
3795
|
let r = q - 1;
|
|
695
|
-
while (r >= nameStart &&
|
|
3796
|
+
while (r >= nameStart && WHITESPACE_REGEX4.test(text[r])) {
|
|
696
3797
|
r -= 1;
|
|
697
3798
|
}
|
|
698
3799
|
const selfClosing = text[r] === "/";
|
|
@@ -721,11 +3822,11 @@ function nextTagToken(text, fromPos) {
|
|
|
721
3822
|
if (next === "/") {
|
|
722
3823
|
const closing = consumeClosingTag(text, lt);
|
|
723
3824
|
let p = lt + 2;
|
|
724
|
-
while (p < text.length &&
|
|
3825
|
+
while (p < text.length && WHITESPACE_REGEX4.test(text[p])) {
|
|
725
3826
|
p += 1;
|
|
726
3827
|
}
|
|
727
3828
|
const nameStart = p;
|
|
728
|
-
while (p < text.length &&
|
|
3829
|
+
while (p < text.length && NAME_CHAR_RE2.test(text.charAt(p))) {
|
|
729
3830
|
p += 1;
|
|
730
3831
|
}
|
|
731
3832
|
const name = text.slice(nameStart, p);
|
|
@@ -896,7 +3997,7 @@ function isOpenTagPrefix(suffix, toolName) {
|
|
|
896
3997
|
}
|
|
897
3998
|
function consumeWhitespace(text, index) {
|
|
898
3999
|
let i = index;
|
|
899
|
-
while (i < text.length &&
|
|
4000
|
+
while (i < text.length && WHITESPACE_REGEX4.test(text.charAt(i))) {
|
|
900
4001
|
i += 1;
|
|
901
4002
|
}
|
|
902
4003
|
return i;
|
|
@@ -1167,7 +4268,7 @@ var xmlProtocol = (protocolOptions) => {
|
|
|
1167
4268
|
args = toolCall.input;
|
|
1168
4269
|
}
|
|
1169
4270
|
}
|
|
1170
|
-
return (
|
|
4271
|
+
return stringify2(toolCall.toolName, args, {
|
|
1171
4272
|
suppressEmptyNode: false,
|
|
1172
4273
|
format: true,
|
|
1173
4274
|
minimalEscaping: true
|
|
@@ -1285,8 +4386,8 @@ var xmlProtocol = (protocolOptions) => {
|
|
|
1285
4386
|
|
|
1286
4387
|
// src/core/protocols/yaml-protocol.ts
|
|
1287
4388
|
var import_yaml = __toESM(require("yaml"), 1);
|
|
1288
|
-
var
|
|
1289
|
-
var
|
|
4389
|
+
var NAME_CHAR_RE3 = /[A-Za-z0-9_:-]/;
|
|
4390
|
+
var WHITESPACE_REGEX5 = /\s/;
|
|
1290
4391
|
var LEADING_WHITESPACE_RE = /^(\s*)/;
|
|
1291
4392
|
function findClosingTagEnd(text, contentStart, toolName) {
|
|
1292
4393
|
let pos = contentStart;
|
|
@@ -1303,11 +4404,11 @@ function findClosingTagEnd(text, contentStart, toolName) {
|
|
|
1303
4404
|
break;
|
|
1304
4405
|
}
|
|
1305
4406
|
let p = ltIdx + 2;
|
|
1306
|
-
while (p < gtIdx &&
|
|
4407
|
+
while (p < gtIdx && WHITESPACE_REGEX5.test(text[p])) {
|
|
1307
4408
|
p++;
|
|
1308
4409
|
}
|
|
1309
4410
|
const nameStart = p;
|
|
1310
|
-
while (p < gtIdx &&
|
|
4411
|
+
while (p < gtIdx && NAME_CHAR_RE3.test(text.charAt(p))) {
|
|
1311
4412
|
p++;
|
|
1312
4413
|
}
|
|
1313
4414
|
const name = text.slice(nameStart, p);
|
|
@@ -1323,11 +4424,11 @@ function findClosingTagEnd(text, contentStart, toolName) {
|
|
|
1323
4424
|
pos = gtIdx === -1 ? text.length : gtIdx + 1;
|
|
1324
4425
|
} else {
|
|
1325
4426
|
let p = ltIdx + 1;
|
|
1326
|
-
while (p < text.length &&
|
|
4427
|
+
while (p < text.length && WHITESPACE_REGEX5.test(text[p])) {
|
|
1327
4428
|
p++;
|
|
1328
4429
|
}
|
|
1329
4430
|
const nameStart = p;
|
|
1330
|
-
while (p < text.length &&
|
|
4431
|
+
while (p < text.length && NAME_CHAR_RE3.test(text.charAt(p))) {
|
|
1331
4432
|
p++;
|
|
1332
4433
|
}
|
|
1333
4434
|
const name = text.slice(nameStart, p);
|
|
@@ -1336,7 +4437,7 @@ function findClosingTagEnd(text, contentStart, toolName) {
|
|
|
1336
4437
|
break;
|
|
1337
4438
|
}
|
|
1338
4439
|
let r = gtIdx - 1;
|
|
1339
|
-
while (r >= nameStart &&
|
|
4440
|
+
while (r >= nameStart && WHITESPACE_REGEX5.test(text[r])) {
|
|
1340
4441
|
r--;
|
|
1341
4442
|
}
|
|
1342
4443
|
const selfClosing = text[r] === "/";
|
|
@@ -1841,7 +4942,6 @@ function hasInputProperty(obj) {
|
|
|
1841
4942
|
|
|
1842
4943
|
// src/generate-handler.ts
|
|
1843
4944
|
var import_provider_utils = require("@ai-sdk/provider-utils");
|
|
1844
|
-
var import_schema_coerce = require("@ai-sdk-tool/schema-coerce");
|
|
1845
4945
|
function parseToolChoiceJson(text, providerOptions) {
|
|
1846
4946
|
var _a;
|
|
1847
4947
|
try {
|
|
@@ -2003,7 +5103,7 @@ function fixToolCallWithSchema(part, tools) {
|
|
|
2003
5103
|
args = part.input;
|
|
2004
5104
|
}
|
|
2005
5105
|
const schema = (_a = tools.find((t) => t.name === part.toolName)) == null ? void 0 : _a.inputSchema;
|
|
2006
|
-
const coerced =
|
|
5106
|
+
const coerced = coerceBySchema(args, schema);
|
|
2007
5107
|
return {
|
|
2008
5108
|
...part,
|
|
2009
5109
|
input: JSON.stringify(coerced != null ? coerced : {})
|
|
@@ -2928,14 +6028,16 @@ var yamlToolMiddleware = createToolMiddleware({
|
|
|
2928
6028
|
logParsedSummary,
|
|
2929
6029
|
logRawChunk,
|
|
2930
6030
|
originalToolsSchema,
|
|
6031
|
+
parse,
|
|
6032
|
+
stringify,
|
|
2931
6033
|
toolChoiceStream,
|
|
6034
|
+
transform,
|
|
2932
6035
|
transformParams,
|
|
2933
6036
|
wrapGenerate,
|
|
2934
6037
|
wrapStream,
|
|
2935
6038
|
xmlProtocol,
|
|
2936
6039
|
xmlToolMiddleware,
|
|
2937
6040
|
yamlProtocol,
|
|
2938
|
-
yamlToolMiddleware
|
|
2939
|
-
...require("@ai-sdk-tool/rjson")
|
|
6041
|
+
yamlToolMiddleware
|
|
2940
6042
|
});
|
|
2941
6043
|
//# sourceMappingURL=index.cjs.map
|