glost-processor 0.7.0 → 1.0.2
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/LICENSE +21 -21
- package/README.md +262 -262
- package/package.json +4 -4
- package/src/__benchmarks__/stream-processor.bench.d.ts +8 -0
- package/src/__benchmarks__/stream-processor.bench.d.ts.map +1 -0
- package/src/__benchmarks__/stream-processor.bench.js +202 -0
- package/src/__benchmarks__/stream-processor.bench.js.map +1 -0
- package/src/index.d.ts +59 -0
- package/src/index.d.ts.map +1 -0
- package/src/index.js +57 -0
- package/src/index.js.map +1 -0
- package/src/processor.d.ts +256 -0
- package/src/processor.d.ts.map +1 -0
- package/src/processor.js +482 -0
- package/src/processor.js.map +1 -0
- package/src/stream-processor.d.ts +145 -0
- package/src/stream-processor.d.ts.map +1 -0
- package/src/stream-processor.js +245 -0
- package/src/stream-processor.js.map +1 -0
- package/src/types.d.ts +167 -0
- package/src/types.d.ts.map +1 -0
- package/src/types.js +9 -0
- package/src/types.js.map +1 -0
package/src/processor.js
ADDED
|
@@ -0,0 +1,482 @@
|
|
|
1
|
+
/**
|
|
2
|
+
* GLOST Processor
|
|
3
|
+
*
|
|
4
|
+
* Unified-style processor for GLOST documents with fluent API.
|
|
5
|
+
* Similar to unified/remark processors.
|
|
6
|
+
*
|
|
7
|
+
* @packageDocumentation
|
|
8
|
+
*/
|
|
9
|
+
import { processGLOSTWithExtensionsAsync, extensionRegistry } from "glost-extensions";
|
|
10
|
+
/**
|
|
11
|
+
* GLOST Processor
|
|
12
|
+
*
|
|
13
|
+
* Fluent API for processing GLOST documents through plugin pipelines.
|
|
14
|
+
*
|
|
15
|
+
* @example
|
|
16
|
+
* ```typescript
|
|
17
|
+
* import { GLOSTProcessor } from "glost-processor";
|
|
18
|
+
*
|
|
19
|
+
* const processor = new GLOSTProcessor()
|
|
20
|
+
* .use(transcription, { scheme: "ipa" })
|
|
21
|
+
* .use(translation, { target: "en" })
|
|
22
|
+
* .use(frequency);
|
|
23
|
+
*
|
|
24
|
+
* const result = await processor.process(document);
|
|
25
|
+
* ```
|
|
26
|
+
*/
|
|
27
|
+
export class GLOSTProcessor {
|
|
28
|
+
plugins = [];
|
|
29
|
+
hooks = {
|
|
30
|
+
before: new Map(),
|
|
31
|
+
after: new Map(),
|
|
32
|
+
onError: [],
|
|
33
|
+
onSkip: [],
|
|
34
|
+
onProgress: [],
|
|
35
|
+
};
|
|
36
|
+
dataStore = new Map();
|
|
37
|
+
options = {};
|
|
38
|
+
frozen = false;
|
|
39
|
+
/**
|
|
40
|
+
* Create a new processor instance
|
|
41
|
+
*
|
|
42
|
+
* @param options - Initial processor options
|
|
43
|
+
*/
|
|
44
|
+
constructor(options = {}) {
|
|
45
|
+
this.options = { ...options };
|
|
46
|
+
if (options.data) {
|
|
47
|
+
this.dataStore = new Map(options.data);
|
|
48
|
+
}
|
|
49
|
+
}
|
|
50
|
+
/**
|
|
51
|
+
* Use a plugin, preset, or extension
|
|
52
|
+
*
|
|
53
|
+
* @param spec - Plugin function, extension object, preset, or plugin ID
|
|
54
|
+
* @param options - Plugin options
|
|
55
|
+
* @returns This processor for chaining
|
|
56
|
+
*
|
|
57
|
+
* @example
|
|
58
|
+
* ```typescript
|
|
59
|
+
* processor
|
|
60
|
+
* .use(transcription, { scheme: "ipa" })
|
|
61
|
+
* .use(translation)
|
|
62
|
+
* .use("frequency");
|
|
63
|
+
* ```
|
|
64
|
+
*/
|
|
65
|
+
use(spec, options) {
|
|
66
|
+
this.assertNotFrozen();
|
|
67
|
+
// Handle presets
|
|
68
|
+
if (this.isPreset(spec)) {
|
|
69
|
+
return this.usePreset(spec);
|
|
70
|
+
}
|
|
71
|
+
// Add plugin to pipeline
|
|
72
|
+
this.plugins.push({ spec, options });
|
|
73
|
+
return this;
|
|
74
|
+
}
|
|
75
|
+
/**
|
|
76
|
+
* Apply a preset (collection of plugins)
|
|
77
|
+
*
|
|
78
|
+
* @param preset - Preset to apply
|
|
79
|
+
* @returns This processor for chaining
|
|
80
|
+
*/
|
|
81
|
+
usePreset(preset) {
|
|
82
|
+
for (const pluginEntry of preset.plugins) {
|
|
83
|
+
if (Array.isArray(pluginEntry)) {
|
|
84
|
+
const [plugin, opts] = pluginEntry;
|
|
85
|
+
this.use(plugin, opts);
|
|
86
|
+
}
|
|
87
|
+
else {
|
|
88
|
+
this.use(pluginEntry);
|
|
89
|
+
}
|
|
90
|
+
}
|
|
91
|
+
return this;
|
|
92
|
+
}
|
|
93
|
+
/**
|
|
94
|
+
* Register a hook to run before a plugin
|
|
95
|
+
*
|
|
96
|
+
* @param pluginId - Plugin ID to hook into
|
|
97
|
+
* @param hook - Hook function to run
|
|
98
|
+
* @returns This processor for chaining
|
|
99
|
+
*
|
|
100
|
+
* @example
|
|
101
|
+
* ```typescript
|
|
102
|
+
* processor.before("translation", (doc) => {
|
|
103
|
+
* console.log("About to translate");
|
|
104
|
+
* });
|
|
105
|
+
* ```
|
|
106
|
+
*/
|
|
107
|
+
before(pluginId, hook) {
|
|
108
|
+
this.assertNotFrozen();
|
|
109
|
+
if (!this.hooks.before.has(pluginId)) {
|
|
110
|
+
this.hooks.before.set(pluginId, []);
|
|
111
|
+
}
|
|
112
|
+
this.hooks.before.get(pluginId).push(hook);
|
|
113
|
+
return this;
|
|
114
|
+
}
|
|
115
|
+
/**
|
|
116
|
+
* Register a hook to run after a plugin
|
|
117
|
+
*
|
|
118
|
+
* @param pluginId - Plugin ID to hook into
|
|
119
|
+
* @param hook - Hook function to run
|
|
120
|
+
* @returns This processor for chaining
|
|
121
|
+
*
|
|
122
|
+
* @example
|
|
123
|
+
* ```typescript
|
|
124
|
+
* processor.after("translation", (doc) => {
|
|
125
|
+
* console.log("Translation complete");
|
|
126
|
+
* });
|
|
127
|
+
* ```
|
|
128
|
+
*/
|
|
129
|
+
after(pluginId, hook) {
|
|
130
|
+
this.assertNotFrozen();
|
|
131
|
+
if (!this.hooks.after.has(pluginId)) {
|
|
132
|
+
this.hooks.after.set(pluginId, []);
|
|
133
|
+
}
|
|
134
|
+
this.hooks.after.get(pluginId).push(hook);
|
|
135
|
+
return this;
|
|
136
|
+
}
|
|
137
|
+
/**
|
|
138
|
+
* Register an error handler
|
|
139
|
+
*
|
|
140
|
+
* @param hook - Error handler function
|
|
141
|
+
* @returns This processor for chaining
|
|
142
|
+
*
|
|
143
|
+
* @example
|
|
144
|
+
* ```typescript
|
|
145
|
+
* processor.onError((error, plugin) => {
|
|
146
|
+
* console.error(`Plugin ${plugin} failed:`, error);
|
|
147
|
+
* });
|
|
148
|
+
* ```
|
|
149
|
+
*/
|
|
150
|
+
onError(hook) {
|
|
151
|
+
this.assertNotFrozen();
|
|
152
|
+
this.hooks.onError.push(hook);
|
|
153
|
+
return this;
|
|
154
|
+
}
|
|
155
|
+
/**
|
|
156
|
+
* Register a skip handler
|
|
157
|
+
*
|
|
158
|
+
* @param hook - Skip handler function
|
|
159
|
+
* @returns This processor for chaining
|
|
160
|
+
*
|
|
161
|
+
* @example
|
|
162
|
+
* ```typescript
|
|
163
|
+
* processor.onSkip((plugin, reason) => {
|
|
164
|
+
* console.log(`Plugin ${plugin} skipped: ${reason}`);
|
|
165
|
+
* });
|
|
166
|
+
* ```
|
|
167
|
+
*/
|
|
168
|
+
onSkip(hook) {
|
|
169
|
+
this.assertNotFrozen();
|
|
170
|
+
this.hooks.onSkip.push(hook);
|
|
171
|
+
return this;
|
|
172
|
+
}
|
|
173
|
+
/**
|
|
174
|
+
* Register a progress handler
|
|
175
|
+
*
|
|
176
|
+
* @param hook - Progress handler function
|
|
177
|
+
* @returns This processor for chaining
|
|
178
|
+
*
|
|
179
|
+
* @example
|
|
180
|
+
* ```typescript
|
|
181
|
+
* processor.onProgress((stats) => {
|
|
182
|
+
* console.log(`Progress: ${stats.completed}/${stats.total}`);
|
|
183
|
+
* });
|
|
184
|
+
* ```
|
|
185
|
+
*/
|
|
186
|
+
onProgress(hook) {
|
|
187
|
+
this.assertNotFrozen();
|
|
188
|
+
this.hooks.onProgress.push(hook);
|
|
189
|
+
return this;
|
|
190
|
+
}
|
|
191
|
+
data(key, value) {
|
|
192
|
+
if (arguments.length === 1) {
|
|
193
|
+
return this.dataStore.get(key);
|
|
194
|
+
}
|
|
195
|
+
this.assertNotFrozen();
|
|
196
|
+
this.dataStore.set(key, value);
|
|
197
|
+
return this;
|
|
198
|
+
}
|
|
199
|
+
/**
|
|
200
|
+
* Freeze the processor
|
|
201
|
+
*
|
|
202
|
+
* Returns a frozen processor that cannot be modified.
|
|
203
|
+
* Useful for reusing the same configuration across multiple documents.
|
|
204
|
+
*
|
|
205
|
+
* @returns A frozen copy of this processor
|
|
206
|
+
*
|
|
207
|
+
* @example
|
|
208
|
+
* ```typescript
|
|
209
|
+
* const frozen = processor
|
|
210
|
+
* .use(transcription)
|
|
211
|
+
* .use(translation)
|
|
212
|
+
* .freeze();
|
|
213
|
+
*
|
|
214
|
+
* // Can process multiple documents with the same pipeline
|
|
215
|
+
* const result1 = await frozen.process(doc1);
|
|
216
|
+
* const result2 = await frozen.process(doc2);
|
|
217
|
+
* ```
|
|
218
|
+
*/
|
|
219
|
+
freeze() {
|
|
220
|
+
const frozen = new GLOSTProcessor(this.options);
|
|
221
|
+
frozen.plugins = [...this.plugins];
|
|
222
|
+
frozen.hooks = {
|
|
223
|
+
before: new Map(this.hooks.before),
|
|
224
|
+
after: new Map(this.hooks.after),
|
|
225
|
+
onError: [...this.hooks.onError],
|
|
226
|
+
onSkip: [...this.hooks.onSkip],
|
|
227
|
+
onProgress: [...this.hooks.onProgress],
|
|
228
|
+
};
|
|
229
|
+
frozen.dataStore = new Map(this.dataStore);
|
|
230
|
+
frozen.frozen = true;
|
|
231
|
+
return frozen;
|
|
232
|
+
}
|
|
233
|
+
/**
|
|
234
|
+
* Process a document through the pipeline
|
|
235
|
+
*
|
|
236
|
+
* @param document - GLOST document to process
|
|
237
|
+
* @returns Promise resolving to the processed document
|
|
238
|
+
*
|
|
239
|
+
* @example
|
|
240
|
+
* ```typescript
|
|
241
|
+
* const result = await processor.process(document);
|
|
242
|
+
* console.log(result);
|
|
243
|
+
* ```
|
|
244
|
+
*/
|
|
245
|
+
async process(document) {
|
|
246
|
+
const result = await this.processWithMeta(document);
|
|
247
|
+
return result.document;
|
|
248
|
+
}
|
|
249
|
+
/**
|
|
250
|
+
* Process a document and return detailed metadata
|
|
251
|
+
*
|
|
252
|
+
* @param document - GLOST document to process
|
|
253
|
+
* @returns Promise resolving to processing result with metadata
|
|
254
|
+
*
|
|
255
|
+
* @example
|
|
256
|
+
* ```typescript
|
|
257
|
+
* const result = await processor.processWithMeta(document);
|
|
258
|
+
* console.log(result.metadata.appliedPlugins);
|
|
259
|
+
* console.log(result.metadata.stats.totalTime);
|
|
260
|
+
* ```
|
|
261
|
+
*/
|
|
262
|
+
async processWithMeta(document) {
|
|
263
|
+
const startTime = Date.now();
|
|
264
|
+
const extensions = await this.resolveExtensions();
|
|
265
|
+
const timing = new Map();
|
|
266
|
+
const errors = [];
|
|
267
|
+
const warnings = [];
|
|
268
|
+
const appliedPlugins = [];
|
|
269
|
+
const skippedPlugins = [];
|
|
270
|
+
// Emit initial progress
|
|
271
|
+
this.emitProgress({
|
|
272
|
+
total: extensions.length,
|
|
273
|
+
completed: 0,
|
|
274
|
+
startTime,
|
|
275
|
+
elapsed: 0,
|
|
276
|
+
});
|
|
277
|
+
// Process extensions with hooks
|
|
278
|
+
let processedDoc = document;
|
|
279
|
+
for (let i = 0; i < extensions.length; i++) {
|
|
280
|
+
const extension = extensions[i];
|
|
281
|
+
const pluginStart = Date.now();
|
|
282
|
+
try {
|
|
283
|
+
// Run before hooks
|
|
284
|
+
await this.runBeforeHooks(processedDoc, extension.id);
|
|
285
|
+
// Process with this extension
|
|
286
|
+
const { data: _, ...extensionOptions } = this.options;
|
|
287
|
+
const result = await processGLOSTWithExtensionsAsync(processedDoc, [extension], extensionOptions);
|
|
288
|
+
processedDoc = result.document;
|
|
289
|
+
// Check for errors/skips
|
|
290
|
+
if (result.metadata.errors.length > 0) {
|
|
291
|
+
for (const err of result.metadata.errors) {
|
|
292
|
+
errors.push({
|
|
293
|
+
plugin: extension.id,
|
|
294
|
+
phase: "transform",
|
|
295
|
+
message: err.error.message,
|
|
296
|
+
stack: err.error.stack,
|
|
297
|
+
recoverable: true,
|
|
298
|
+
error: err.error,
|
|
299
|
+
});
|
|
300
|
+
}
|
|
301
|
+
}
|
|
302
|
+
if (result.metadata.skippedExtensions.includes(extension.id)) {
|
|
303
|
+
skippedPlugins.push(extension.id);
|
|
304
|
+
this.emitSkip(extension.id, "Skipped by processor");
|
|
305
|
+
}
|
|
306
|
+
else {
|
|
307
|
+
appliedPlugins.push(extension.id);
|
|
308
|
+
}
|
|
309
|
+
// Run after hooks
|
|
310
|
+
await this.runAfterHooks(processedDoc, extension.id);
|
|
311
|
+
// Record timing
|
|
312
|
+
timing.set(extension.id, Date.now() - pluginStart);
|
|
313
|
+
// Emit progress
|
|
314
|
+
this.emitProgress({
|
|
315
|
+
total: extensions.length,
|
|
316
|
+
completed: i + 1,
|
|
317
|
+
current: extension.id,
|
|
318
|
+
startTime,
|
|
319
|
+
elapsed: Date.now() - startTime,
|
|
320
|
+
});
|
|
321
|
+
}
|
|
322
|
+
catch (error) {
|
|
323
|
+
const err = error instanceof Error ? error : new Error(String(error));
|
|
324
|
+
errors.push({
|
|
325
|
+
plugin: extension.id,
|
|
326
|
+
phase: "transform",
|
|
327
|
+
message: err.message,
|
|
328
|
+
stack: err.stack,
|
|
329
|
+
recoverable: false,
|
|
330
|
+
error: err,
|
|
331
|
+
});
|
|
332
|
+
skippedPlugins.push(extension.id);
|
|
333
|
+
this.emitError(err, extension.id);
|
|
334
|
+
this.emitSkip(extension.id, err.message);
|
|
335
|
+
// Re-throw in strict mode
|
|
336
|
+
if (!this.options.lenient) {
|
|
337
|
+
throw err;
|
|
338
|
+
}
|
|
339
|
+
}
|
|
340
|
+
}
|
|
341
|
+
const endTime = Date.now();
|
|
342
|
+
return {
|
|
343
|
+
document: processedDoc,
|
|
344
|
+
metadata: {
|
|
345
|
+
appliedPlugins,
|
|
346
|
+
skippedPlugins,
|
|
347
|
+
errors,
|
|
348
|
+
warnings,
|
|
349
|
+
stats: {
|
|
350
|
+
totalTime: endTime - startTime,
|
|
351
|
+
timing,
|
|
352
|
+
nodesProcessed: 0, // Could be calculated by visiting the tree
|
|
353
|
+
startTime,
|
|
354
|
+
endTime,
|
|
355
|
+
},
|
|
356
|
+
},
|
|
357
|
+
};
|
|
358
|
+
}
|
|
359
|
+
/**
|
|
360
|
+
* Process a document synchronously (only if all plugins are sync)
|
|
361
|
+
*
|
|
362
|
+
* @param document - GLOST document to process
|
|
363
|
+
* @returns The processed document
|
|
364
|
+
* @throws {Error} If any plugin is async
|
|
365
|
+
*
|
|
366
|
+
* @example
|
|
367
|
+
* ```typescript
|
|
368
|
+
* const result = processor.processSync(document);
|
|
369
|
+
* ```
|
|
370
|
+
*/
|
|
371
|
+
processSync(document) {
|
|
372
|
+
throw new Error("Synchronous processing not yet implemented. Use process() instead.");
|
|
373
|
+
}
|
|
374
|
+
/**
|
|
375
|
+
* Resolve all plugins to extensions
|
|
376
|
+
*/
|
|
377
|
+
async resolveExtensions() {
|
|
378
|
+
const extensions = [];
|
|
379
|
+
for (const { spec, options } of this.plugins) {
|
|
380
|
+
const extension = await this.resolvePlugin(spec, options);
|
|
381
|
+
if (extension) {
|
|
382
|
+
extensions.push(extension);
|
|
383
|
+
}
|
|
384
|
+
}
|
|
385
|
+
return extensions;
|
|
386
|
+
}
|
|
387
|
+
/**
|
|
388
|
+
* Resolve a single plugin to an extension
|
|
389
|
+
*/
|
|
390
|
+
async resolvePlugin(spec, options) {
|
|
391
|
+
// String ID - lookup in registry
|
|
392
|
+
if (typeof spec === "string") {
|
|
393
|
+
const ext = extensionRegistry.get(spec);
|
|
394
|
+
if (!ext) {
|
|
395
|
+
throw new Error(`Plugin "${spec}" not found in registry`);
|
|
396
|
+
}
|
|
397
|
+
return ext;
|
|
398
|
+
}
|
|
399
|
+
// Function - call to get extension
|
|
400
|
+
if (typeof spec === "function") {
|
|
401
|
+
const result = spec(options);
|
|
402
|
+
return result || null;
|
|
403
|
+
}
|
|
404
|
+
// Extension object - use directly
|
|
405
|
+
return spec;
|
|
406
|
+
}
|
|
407
|
+
/**
|
|
408
|
+
* Check if a spec is a preset
|
|
409
|
+
*/
|
|
410
|
+
isPreset(spec) {
|
|
411
|
+
return (spec &&
|
|
412
|
+
typeof spec === "object" &&
|
|
413
|
+
"plugins" in spec &&
|
|
414
|
+
Array.isArray(spec.plugins));
|
|
415
|
+
}
|
|
416
|
+
/**
|
|
417
|
+
* Run before hooks for a plugin
|
|
418
|
+
*/
|
|
419
|
+
async runBeforeHooks(document, pluginId) {
|
|
420
|
+
const hooks = this.hooks.before.get(pluginId) || [];
|
|
421
|
+
for (const hook of hooks) {
|
|
422
|
+
await hook(document, pluginId);
|
|
423
|
+
}
|
|
424
|
+
}
|
|
425
|
+
/**
|
|
426
|
+
* Run after hooks for a plugin
|
|
427
|
+
*/
|
|
428
|
+
async runAfterHooks(document, pluginId) {
|
|
429
|
+
const hooks = this.hooks.after.get(pluginId) || [];
|
|
430
|
+
for (const hook of hooks) {
|
|
431
|
+
await hook(document, pluginId);
|
|
432
|
+
}
|
|
433
|
+
}
|
|
434
|
+
/**
|
|
435
|
+
* Emit error to error handlers
|
|
436
|
+
*/
|
|
437
|
+
emitError(error, pluginId) {
|
|
438
|
+
for (const hook of this.hooks.onError) {
|
|
439
|
+
try {
|
|
440
|
+
hook(error, pluginId);
|
|
441
|
+
}
|
|
442
|
+
catch (err) {
|
|
443
|
+
console.error("Error in error hook:", err);
|
|
444
|
+
}
|
|
445
|
+
}
|
|
446
|
+
}
|
|
447
|
+
/**
|
|
448
|
+
* Emit skip to skip handlers
|
|
449
|
+
*/
|
|
450
|
+
emitSkip(pluginId, reason) {
|
|
451
|
+
for (const hook of this.hooks.onSkip) {
|
|
452
|
+
try {
|
|
453
|
+
hook(pluginId, reason);
|
|
454
|
+
}
|
|
455
|
+
catch (err) {
|
|
456
|
+
console.error("Error in skip hook:", err);
|
|
457
|
+
}
|
|
458
|
+
}
|
|
459
|
+
}
|
|
460
|
+
/**
|
|
461
|
+
* Emit progress to progress handlers
|
|
462
|
+
*/
|
|
463
|
+
emitProgress(stats) {
|
|
464
|
+
for (const hook of this.hooks.onProgress) {
|
|
465
|
+
try {
|
|
466
|
+
hook(stats);
|
|
467
|
+
}
|
|
468
|
+
catch (err) {
|
|
469
|
+
console.error("Error in progress hook:", err);
|
|
470
|
+
}
|
|
471
|
+
}
|
|
472
|
+
}
|
|
473
|
+
/**
|
|
474
|
+
* Assert that the processor is not frozen
|
|
475
|
+
*/
|
|
476
|
+
assertNotFrozen() {
|
|
477
|
+
if (this.frozen) {
|
|
478
|
+
throw new Error("Cannot modify frozen processor");
|
|
479
|
+
}
|
|
480
|
+
}
|
|
481
|
+
}
|
|
482
|
+
//# sourceMappingURL=processor.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"processor.js","sourceRoot":"","sources":["processor.ts"],"names":[],"mappings":"AAAA;;;;;;;GAOG;AAIH,OAAO,EAAE,+BAA+B,EAAE,iBAAiB,EAAE,MAAM,kBAAkB,CAAC;AAiBtF;;;;;;;;;;;;;;;;GAgBG;AACH,MAAM,OAAO,cAAc;IACjB,OAAO,GAA+C,EAAE,CAAC;IACzD,KAAK,GAAmB;QAC9B,MAAM,EAAE,IAAI,GAAG,EAAE;QACjB,KAAK,EAAE,IAAI,GAAG,EAAE;QAChB,OAAO,EAAE,EAAE;QACX,MAAM,EAAE,EAAE;QACV,UAAU,EAAE,EAAE;KACf,CAAC;IACM,SAAS,GAAqB,IAAI,GAAG,EAAE,CAAC;IACxC,OAAO,GAAqB,EAAE,CAAC;IAC/B,MAAM,GAAG,KAAK,CAAC;IAEvB;;;;OAIG;IACH,YAAY,UAA4B,EAAE;QACxC,IAAI,CAAC,OAAO,GAAG,EAAE,GAAG,OAAO,EAAE,CAAC;QAC9B,IAAI,OAAO,CAAC,IAAI,EAAE,CAAC;YACjB,IAAI,CAAC,SAAS,GAAG,IAAI,GAAG,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC;QACzC,CAAC;IACH,CAAC;IAED;;;;;;;;;;;;;;OAcG;IACH,GAAG,CAAC,IAAyB,EAAE,OAAa;QAC1C,IAAI,CAAC,eAAe,EAAE,CAAC;QAEvB,iBAAiB;QACjB,IAAI,IAAI,CAAC,QAAQ,CAAC,IAAI,CAAC,EAAE,CAAC;YACxB,OAAO,IAAI,CAAC,SAAS,CAAC,IAAI,CAAC,CAAC;QAC9B,CAAC;QAED,yBAAyB;QACzB,IAAI,CAAC,OAAO,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,OAAO,EAAE,CAAC,CAAC;QACrC,OAAO,IAAI,CAAC;IACd,CAAC;IAED;;;;;OAKG;IACK,SAAS,CAAC,MAAc;QAC9B,KAAK,MAAM,WAAW,IAAI,MAAM,CAAC,OAAO,EAAE,CAAC;YACzC,IAAI,KAAK,CAAC,OAAO,CAAC,WAAW,CAAC,EAAE,CAAC;gBAC/B,MAAM,CAAC,MAAM,EAAE,IAAI,CAAC,GAAG,WAAW,CAAC;gBACnC,IAAI,CAAC,GAAG,CAAC,MAAM,EAAE,IAAI,CAAC,CAAC;YACzB,CAAC;iBAAM,CAAC;gBACN,IAAI,CAAC,GAAG,CAAC,WAAW,CAAC,CAAC;YACxB,CAAC;QACH,CAAC;QACD,OAAO,IAAI,CAAC;IACd,CAAC;IAED;;;;;;;;;;;;;OAaG;IACH,MAAM,CAAC,QAAgB,EAAE,IAAgB;QACvC,IAAI,CAAC,eAAe,EAAE,CAAC;QACvB,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,MAAM,CAAC,GAAG,CAAC,QAAQ,CAAC,EAAE,CAAC;YACrC,IAAI,CAAC,KAAK,CAAC,MAAM,CAAC,GAAG,CAAC,QAAQ,EAAE,EAAE,CAAC,CAAC;QACtC,CAAC;QACD,IAAI,CAAC,KAAK,CAAC,MAAM,CAAC,GAAG,CAAC,QAAQ,CAAE,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QAC5C,OAAO,IAAI,CAAC;IACd,CAAC;IAED;;;;;;;;;;;;;OAaG;IACH,KAAK,CAAC,QAAgB,EAAE,IAAe;QACrC,IAAI,CAAC,eAAe,EAAE,CAAC;QACvB,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,KAAK,CAAC,GAAG,CAAC,QAAQ,CAAC,EAAE,CAAC;YACpC,IAAI,CAAC,KAAK,CAAC,KAAK,CAAC,GAAG,CAAC,QAAQ,EAAE,EAAE,CAAC,CAAC;QACrC,CAAC;QACD,IAAI,CAAC,KAAK,CAAC,KAAK,CAAC,GAAG,CAAC,QAAQ,CAAE,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QAC3C,OAAO,IAAI,CAAC;IACd,CAAC;IAED;;;;;;;;;;;;OAYG;IACH,OAAO,CAAC,IAAe;QACrB,IAAI,CAAC,eAAe,EAAE,CAAC;QACvB,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QAC9B,OAAO,IAAI,CAAC;IACd,CAAC;IAED;;;;;;;;;;;;OAYG;IACH,MAAM,CAAC,IAAc;QACnB,IAAI,CAAC,eAAe,EAAE,CAAC;QACvB,IAAI,CAAC,KAAK,CAAC,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QAC7B,OAAO,IAAI,CAAC;IACd,CAAC;IAED;;;;;;;;;;;;OAYG;IACH,UAAU,CAAC,IAAkB;QAC3B,IAAI,CAAC,eAAe,EAAE,CAAC;QACvB,IAAI,CAAC,KAAK,CAAC,UAAU,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QACjC,OAAO,IAAI,CAAC;IACd,CAAC;IAiBD,IAAI,CAAC,GAAW,EAAE,KAAW;QAC3B,IAAI,SAAS,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;YAC3B,OAAO,IAAI,CAAC,SAAS,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;QACjC,CAAC;QACD,IAAI,CAAC,eAAe,EAAE,CAAC;QACvB,IAAI,CAAC,SAAS,CAAC,GAAG,CAAC,GAAG,EAAE,KAAK,CAAC,CAAC;QAC/B,OAAO,IAAI,CAAC;IACd,CAAC;IAED;;;;;;;;;;;;;;;;;;;OAmBG;IACH,MAAM;QACJ,MAAM,MAAM,GAAG,IAAI,cAAc,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC;QAChD,MAAM,CAAC,OAAO,GAAG,CAAC,GAAG,IAAI,CAAC,OAAO,CAAC,CAAC;QACnC,MAAM,CAAC,KAAK,GAAG;YACb,MAAM,EAAE,IAAI,GAAG,CAAC,IAAI,CAAC,KAAK,CAAC,MAAM,CAAC;YAClC,KAAK,EAAE,IAAI,GAAG,CAAC,IAAI,CAAC,KAAK,CAAC,KAAK,CAAC;YAChC,OAAO,EAAE,CAAC,GAAG,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC;YAChC,MAAM,EAAE,CAAC,GAAG,IAAI,CAAC,KAAK,CAAC,MAAM,CAAC;YAC9B,UAAU,EAAE,CAAC,GAAG,IAAI,CAAC,KAAK,CAAC,UAAU,CAAC;SACvC,CAAC;QACF,MAAM,CAAC,SAAS,GAAG,IAAI,GAAG,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC;QAC1C,MAAc,CAAC,MAAM,GAAG,IAAI,CAAC;QAC9B,OAAO,MAAoC,CAAC;IAC9C,CAAC;IAED;;;;;;;;;;;OAWG;IACH,KAAK,CAAC,OAAO,CAAC,QAAmB;QAC/B,MAAM,MAAM,GAAG,MAAM,IAAI,CAAC,eAAe,CAAC,QAAQ,CAAC,CAAC;QACpD,OAAO,MAAM,CAAC,QAAQ,CAAC;IACzB,CAAC;IAED;;;;;;;;;;;;OAYG;IACH,KAAK,CAAC,eAAe,CAAC,QAAmB;QACvC,MAAM,SAAS,GAAG,IAAI,CAAC,GAAG,EAAE,CAAC;QAC7B,MAAM,UAAU,GAAG,MAAM,IAAI,CAAC,iBAAiB,EAAE,CAAC;QAClD,MAAM,MAAM,GAAG,IAAI,GAAG,EAAkB,CAAC;QACzC,MAAM,MAAM,GAAsB,EAAE,CAAC;QACrC,MAAM,QAAQ,GAAwB,EAAE,CAAC;QACzC,MAAM,cAAc,GAAa,EAAE,CAAC;QACpC,MAAM,cAAc,GAAa,EAAE,CAAC;QAEpC,wBAAwB;QACxB,IAAI,CAAC,YAAY,CAAC;YAChB,KAAK,EAAE,UAAU,CAAC,MAAM;YACxB,SAAS,EAAE,CAAC;YACZ,SAAS;YACT,OAAO,EAAE,CAAC;SACX,CAAC,CAAC;QAEH,gCAAgC;QAChC,IAAI,YAAY,GAAG,QAAQ,CAAC;QAE5B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,UAAU,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;YAC3C,MAAM,SAAS,GAAG,UAAU,CAAC,CAAC,CAAE,CAAC;YACjC,MAAM,WAAW,GAAG,IAAI,CAAC,GAAG,EAAE,CAAC;YAE/B,IAAI,CAAC;gBACH,mBAAmB;gBACnB,MAAM,IAAI,CAAC,cAAc,CAAC,YAAY,EAAE,SAAS,CAAC,EAAE,CAAC,CAAC;gBAEtD,8BAA8B;gBAC9B,MAAM,EAAE,IAAI,EAAE,CAAC,EAAE,GAAG,gBAAgB,EAAE,GAAG,IAAI,CAAC,OAAc,CAAC;gBAC7D,MAAM,MAAM,GAAG,MAAM,+BAA+B,CAClD,YAAY,EACZ,CAAC,SAAS,CAAC,EACX,gBAAgB,CACjB,CAAC;gBAEF,YAAY,GAAG,MAAM,CAAC,QAAQ,CAAC;gBAE/B,yBAAyB;gBACzB,IAAI,MAAM,CAAC,QAAQ,CAAC,MAAM,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;oBACtC,KAAK,MAAM,GAAG,IAAI,MAAM,CAAC,QAAQ,CAAC,MAAM,EAAE,CAAC;wBACzC,MAAM,CAAC,IAAI,CAAC;4BACV,MAAM,EAAE,SAAS,CAAC,EAAE;4BACpB,KAAK,EAAE,WAAW;4BAClB,OAAO,EAAE,GAAG,CAAC,KAAK,CAAC,OAAO;4BAC1B,KAAK,EAAE,GAAG,CAAC,KAAK,CAAC,KAAK;4BACtB,WAAW,EAAE,IAAI;4BACjB,KAAK,EAAE,GAAG,CAAC,KAAK;yBACjB,CAAC,CAAC;oBACL,CAAC;gBACH,CAAC;gBAED,IAAI,MAAM,CAAC,QAAQ,CAAC,iBAAiB,CAAC,QAAQ,CAAC,SAAS,CAAC,EAAE,CAAC,EAAE,CAAC;oBAC7D,cAAc,CAAC,IAAI,CAAC,SAAS,CAAC,EAAE,CAAC,CAAC;oBAClC,IAAI,CAAC,QAAQ,CAAC,SAAS,CAAC,EAAE,EAAE,sBAAsB,CAAC,CAAC;gBACtD,CAAC;qBAAM,CAAC;oBACN,cAAc,CAAC,IAAI,CAAC,SAAS,CAAC,EAAE,CAAC,CAAC;gBACpC,CAAC;gBAED,kBAAkB;gBAClB,MAAM,IAAI,CAAC,aAAa,CAAC,YAAY,EAAE,SAAS,CAAC,EAAE,CAAC,CAAC;gBAErD,gBAAgB;gBAChB,MAAM,CAAC,GAAG,CAAC,SAAS,CAAC,EAAE,EAAE,IAAI,CAAC,GAAG,EAAE,GAAG,WAAW,CAAC,CAAC;gBAEnD,gBAAgB;gBAChB,IAAI,CAAC,YAAY,CAAC;oBAChB,KAAK,EAAE,UAAU,CAAC,MAAM;oBACxB,SAAS,EAAE,CAAC,GAAG,CAAC;oBAChB,OAAO,EAAE,SAAS,CAAC,EAAE;oBACrB,SAAS;oBACT,OAAO,EAAE,IAAI,CAAC,GAAG,EAAE,GAAG,SAAS;iBAChC,CAAC,CAAC;YACL,CAAC;YAAC,OAAO,KAAK,EAAE,CAAC;gBACf,MAAM,GAAG,GAAG,KAAK,YAAY,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,IAAI,KAAK,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC;gBACtE,MAAM,CAAC,IAAI,CAAC;oBACV,MAAM,EAAE,SAAS,CAAC,EAAE;oBACpB,KAAK,EAAE,WAAW;oBAClB,OAAO,EAAE,GAAG,CAAC,OAAO;oBACpB,KAAK,EAAE,GAAG,CAAC,KAAK;oBAChB,WAAW,EAAE,KAAK;oBAClB,KAAK,EAAE,GAAG;iBACX,CAAC,CAAC;gBACH,cAAc,CAAC,IAAI,CAAC,SAAS,CAAC,EAAE,CAAC,CAAC;gBAClC,IAAI,CAAC,SAAS,CAAC,GAAG,EAAE,SAAS,CAAC,EAAE,CAAC,CAAC;gBAClC,IAAI,CAAC,QAAQ,CAAC,SAAS,CAAC,EAAE,EAAE,GAAG,CAAC,OAAO,CAAC,CAAC;gBAEzC,0BAA0B;gBAC1B,IAAI,CAAC,IAAI,CAAC,OAAO,CAAC,OAAO,EAAE,CAAC;oBAC1B,MAAM,GAAG,CAAC;gBACZ,CAAC;YACH,CAAC;QACH,CAAC;QAED,MAAM,OAAO,GAAG,IAAI,CAAC,GAAG,EAAE,CAAC;QAE3B,OAAO;YACL,QAAQ,EAAE,YAAY;YACtB,QAAQ,EAAE;gBACR,cAAc;gBACd,cAAc;gBACd,MAAM;gBACN,QAAQ;gBACR,KAAK,EAAE;oBACL,SAAS,EAAE,OAAO,GAAG,SAAS;oBAC9B,MAAM;oBACN,cAAc,EAAE,CAAC,EAAE,2CAA2C;oBAC9D,SAAS;oBACT,OAAO;iBACR;aACF;SACF,CAAC;IACJ,CAAC;IAED;;;;;;;;;;;OAWG;IACH,WAAW,CAAC,QAAmB;QAC7B,MAAM,IAAI,KAAK,CACb,oEAAoE,CACrE,CAAC;IACJ,CAAC;IAED;;OAEG;IACK,KAAK,CAAC,iBAAiB;QAC7B,MAAM,UAAU,GAAqB,EAAE,CAAC;QAExC,KAAK,MAAM,EAAE,IAAI,EAAE,OAAO,EAAE,IAAI,IAAI,CAAC,OAAO,EAAE,CAAC;YAC7C,MAAM,SAAS,GAAG,MAAM,IAAI,CAAC,aAAa,CAAC,IAAI,EAAE,OAAO,CAAC,CAAC;YAC1D,IAAI,SAAS,EAAE,CAAC;gBACd,UAAU,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC;YAC7B,CAAC;QACH,CAAC;QAED,OAAO,UAAU,CAAC;IACpB,CAAC;IAED;;OAEG;IACK,KAAK,CAAC,aAAa,CACzB,IAAgB,EAChB,OAAa;QAEb,iCAAiC;QACjC,IAAI,OAAO,IAAI,KAAK,QAAQ,EAAE,CAAC;YAC7B,MAAM,GAAG,GAAG,iBAAiB,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC;YACxC,IAAI,CAAC,GAAG,EAAE,CAAC;gBACT,MAAM,IAAI,KAAK,CAAC,WAAW,IAAI,yBAAyB,CAAC,CAAC;YAC5D,CAAC;YACD,OAAO,GAAG,CAAC;QACb,CAAC;QAED,mCAAmC;QACnC,IAAI,OAAO,IAAI,KAAK,UAAU,EAAE,CAAC;YAC/B,MAAM,MAAM,GAAG,IAAI,CAAC,OAAO,CAAC,CAAC;YAC7B,OAAO,MAAM,IAAI,IAAI,CAAC;QACxB,CAAC;QAED,kCAAkC;QAClC,OAAO,IAAI,CAAC;IACd,CAAC;IAED;;OAEG;IACK,QAAQ,CAAC,IAAS;QACxB,OAAO,CACL,IAAI;YACJ,OAAO,IAAI,KAAK,QAAQ;YACxB,SAAS,IAAI,IAAI;YACjB,KAAK,CAAC,OAAO,CAAC,IAAI,CAAC,OAAO,CAAC,CAC5B,CAAC;IACJ,CAAC;IAED;;OAEG;IACK,KAAK,CAAC,cAAc,CAAC,QAAmB,EAAE,QAAgB;QAChE,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,MAAM,CAAC,GAAG,CAAC,QAAQ,CAAC,IAAI,EAAE,CAAC;QACpD,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;YACzB,MAAM,IAAI,CAAC,QAAQ,EAAE,QAAQ,CAAC,CAAC;QACjC,CAAC;IACH,CAAC;IAED;;OAEG;IACK,KAAK,CAAC,aAAa,CAAC,QAAmB,EAAE,QAAgB;QAC/D,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,KAAK,CAAC,GAAG,CAAC,QAAQ,CAAC,IAAI,EAAE,CAAC;QACnD,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;YACzB,MAAM,IAAI,CAAC,QAAQ,EAAE,QAAQ,CAAC,CAAC;QACjC,CAAC;IACH,CAAC;IAED;;OAEG;IACK,SAAS,CAAC,KAAY,EAAE,QAAgB;QAC9C,KAAK,MAAM,IAAI,IAAI,IAAI,CAAC,KAAK,CAAC,OAAO,EAAE,CAAC;YACtC,IAAI,CAAC;gBACH,IAAI,CAAC,KAAK,EAAE,QAAQ,CAAC,CAAC;YACxB,CAAC;YAAC,OAAO,GAAG,EAAE,CAAC;gBACb,OAAO,CAAC,KAAK,CAAC,sBAAsB,EAAE,GAAG,CAAC,CAAC;YAC7C,CAAC;QACH,CAAC;IACH,CAAC;IAED;;OAEG;IACK,QAAQ,CAAC,QAAgB,EAAE,MAAc;QAC/C,KAAK,MAAM,IAAI,IAAI,IAAI,CAAC,KAAK,CAAC,MAAM,EAAE,CAAC;YACrC,IAAI,CAAC;gBACH,IAAI,CAAC,QAAQ,EAAE,MAAM,CAAC,CAAC;YACzB,CAAC;YAAC,OAAO,GAAG,EAAE,CAAC;gBACb,OAAO,CAAC,KAAK,CAAC,qBAAqB,EAAE,GAAG,CAAC,CAAC;YAC5C,CAAC;QACH,CAAC;IACH,CAAC;IAED;;OAEG;IACK,YAAY,CAAC,KAAoB;QACvC,KAAK,MAAM,IAAI,IAAI,IAAI,CAAC,KAAK,CAAC,UAAU,EAAE,CAAC;YACzC,IAAI,CAAC;gBACH,IAAI,CAAC,KAAK,CAAC,CAAC;YACd,CAAC;YAAC,OAAO,GAAG,EAAE,CAAC;gBACb,OAAO,CAAC,KAAK,CAAC,yBAAyB,EAAE,GAAG,CAAC,CAAC;YAChD,CAAC;QACH,CAAC;IACH,CAAC;IAED;;OAEG;IACK,eAAe;QACrB,IAAI,IAAI,CAAC,MAAM,EAAE,CAAC;YAChB,MAAM,IAAI,KAAK,CAAC,gCAAgC,CAAC,CAAC;QACpD,CAAC;IACH,CAAC;CACF"}
|
|
@@ -0,0 +1,145 @@
|
|
|
1
|
+
/**
|
|
2
|
+
* GLOST Stream Processor
|
|
3
|
+
*
|
|
4
|
+
* Streaming variant of GLOSTProcessor that yields processed sentence
|
|
5
|
+
* batches progressively using AsyncGenerator. Keeps the full document
|
|
6
|
+
* out of memory between chunks.
|
|
7
|
+
*
|
|
8
|
+
* Document-level transforms (extensions with streamingSupport !== 'chunk')
|
|
9
|
+
* run once on the full document before streaming begins. Chunk-compatible
|
|
10
|
+
* extensions (streamingSupport === 'chunk') then run on each batch.
|
|
11
|
+
*
|
|
12
|
+
* @packageDocumentation
|
|
13
|
+
*
|
|
14
|
+
* @example
|
|
15
|
+
* ```typescript
|
|
16
|
+
* import { GLOSTStreamProcessor } from "glost-processor";
|
|
17
|
+
*
|
|
18
|
+
* const processor = new GLOSTStreamProcessor()
|
|
19
|
+
* .use(transcription)
|
|
20
|
+
* .use(translation);
|
|
21
|
+
*
|
|
22
|
+
* for await (const chunk of processor.stream(document)) {
|
|
23
|
+
* console.log(chunk.sentences, chunk.isLast);
|
|
24
|
+
* }
|
|
25
|
+
* ```
|
|
26
|
+
*
|
|
27
|
+
* @since 0.7.0
|
|
28
|
+
*/
|
|
29
|
+
import type { GLOSTRoot, GLOSTSentence } from "glost-core";
|
|
30
|
+
import type { PluginSpec, Preset, ProcessorOptions } from "./types.js";
|
|
31
|
+
/**
|
|
32
|
+
* Options for stream() method
|
|
33
|
+
*
|
|
34
|
+
* @since 0.7.0
|
|
35
|
+
*/
|
|
36
|
+
export interface StreamOptions {
|
|
37
|
+
/**
|
|
38
|
+
* Number of sentences per chunk.
|
|
39
|
+
*
|
|
40
|
+
* A smaller value reduces latency to first yielded chunk; a larger
|
|
41
|
+
* value amortises per-chunk overhead. Default: 50.
|
|
42
|
+
*/
|
|
43
|
+
batchSize?: number;
|
|
44
|
+
}
|
|
45
|
+
/**
|
|
46
|
+
* A single yielded chunk from the stream
|
|
47
|
+
*
|
|
48
|
+
* @since 0.7.0
|
|
49
|
+
*/
|
|
50
|
+
export interface ProcessedChunk {
|
|
51
|
+
/** Processed sentences in this batch */
|
|
52
|
+
sentences: GLOSTSentence[];
|
|
53
|
+
/**
|
|
54
|
+
* Index of the source paragraph this chunk came from.
|
|
55
|
+
*
|
|
56
|
+
* When the document has multiple paragraphs each paragraph is
|
|
57
|
+
* chunked independently, so multiple consecutive chunks may share
|
|
58
|
+
* the same paragraphIndex.
|
|
59
|
+
*/
|
|
60
|
+
paragraphIndex: number;
|
|
61
|
+
/**
|
|
62
|
+
* Index of this chunk within its paragraph (0-based).
|
|
63
|
+
*/
|
|
64
|
+
chunkIndex: number;
|
|
65
|
+
/** True for the final chunk across the whole document */
|
|
66
|
+
isLast: boolean;
|
|
67
|
+
}
|
|
68
|
+
/**
|
|
69
|
+
* Streaming processor for GLOST documents
|
|
70
|
+
*
|
|
71
|
+
* Mirrors the `GLOSTProcessor` API (`.use()`, `.freeze()`) but adds a
|
|
72
|
+
* `.stream()` method that returns an `AsyncGenerator<ProcessedChunk>`.
|
|
73
|
+
*
|
|
74
|
+
* @since 0.7.0
|
|
75
|
+
*/
|
|
76
|
+
export declare class GLOSTStreamProcessor {
|
|
77
|
+
private plugins;
|
|
78
|
+
private options;
|
|
79
|
+
private frozen;
|
|
80
|
+
/**
|
|
81
|
+
* Create a new stream processor instance
|
|
82
|
+
*
|
|
83
|
+
* @param options - Initial processor options
|
|
84
|
+
*/
|
|
85
|
+
constructor(options?: ProcessorOptions);
|
|
86
|
+
/**
|
|
87
|
+
* Add a plugin, preset, or extension to the pipeline
|
|
88
|
+
*
|
|
89
|
+
* @param spec - Plugin function, extension object, preset, or ID
|
|
90
|
+
* @param options - Plugin options
|
|
91
|
+
* @returns This processor for chaining
|
|
92
|
+
*/
|
|
93
|
+
use(spec: PluginSpec | Preset, options?: unknown): this;
|
|
94
|
+
/**
|
|
95
|
+
* Freeze the processor
|
|
96
|
+
*
|
|
97
|
+
* Returns a frozen processor that cannot be modified. Useful for
|
|
98
|
+
* reusing the same pipeline configuration across multiple documents.
|
|
99
|
+
*
|
|
100
|
+
* @returns A frozen copy of this processor
|
|
101
|
+
*/
|
|
102
|
+
freeze(): FrozenStreamProcessor;
|
|
103
|
+
/**
|
|
104
|
+
* Stream a document as progressive sentence batches
|
|
105
|
+
*
|
|
106
|
+
* Processing phases:
|
|
107
|
+
* 1. All extensions with `streamingSupport !== 'chunk'` (i.e. `'none'`
|
|
108
|
+
* or `'full'`, or unset) run their `transform`, `visit`, and
|
|
109
|
+
* `enhanceMetadata` hooks on the **full** document.
|
|
110
|
+
* 2. The resulting document is split into sentence batches.
|
|
111
|
+
* 3. For each batch, extensions with `streamingSupport === 'chunk'`
|
|
112
|
+
* run their `visit` and `enhanceMetadata` hooks.
|
|
113
|
+
* 4. A `ProcessedChunk` is yielded.
|
|
114
|
+
*
|
|
115
|
+
* Cancellation: break out of the `for await` loop at any time. The
|
|
116
|
+
* generator will stop without processing remaining chunks.
|
|
117
|
+
*
|
|
118
|
+
* @param document - GLOST document to stream
|
|
119
|
+
* @param streamOptions - Streaming options (batchSize etc.)
|
|
120
|
+
* @yields `ProcessedChunk` objects in document order
|
|
121
|
+
*
|
|
122
|
+
* @example
|
|
123
|
+
* ```typescript
|
|
124
|
+
* for await (const chunk of processor.stream(doc, { batchSize: 20 })) {
|
|
125
|
+
* console.log(`para ${chunk.paragraphIndex} chunk ${chunk.chunkIndex}`);
|
|
126
|
+
* if (chunk.isLast) console.log("done");
|
|
127
|
+
* }
|
|
128
|
+
* ```
|
|
129
|
+
*/
|
|
130
|
+
stream(document: GLOSTRoot, streamOptions?: StreamOptions): AsyncGenerator<ProcessedChunk>;
|
|
131
|
+
private usePreset;
|
|
132
|
+
private resolveExtensions;
|
|
133
|
+
private resolvePlugin;
|
|
134
|
+
private isPreset;
|
|
135
|
+
private assertNotFrozen;
|
|
136
|
+
}
|
|
137
|
+
/**
|
|
138
|
+
* A frozen `GLOSTStreamProcessor` that cannot be modified.
|
|
139
|
+
*
|
|
140
|
+
* @since 0.7.0
|
|
141
|
+
*/
|
|
142
|
+
export type FrozenStreamProcessor = Omit<GLOSTStreamProcessor, "use" | "freeze"> & {
|
|
143
|
+
readonly frozen: true;
|
|
144
|
+
};
|
|
145
|
+
//# sourceMappingURL=stream-processor.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"stream-processor.d.ts","sourceRoot":"","sources":["stream-processor.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;GA2BG;AAEH,OAAO,KAAK,EAAE,SAAS,EAAE,aAAa,EAAE,MAAM,YAAY,CAAC;AAO3D,OAAO,KAAK,EAAE,UAAU,EAAE,MAAM,EAAE,gBAAgB,EAAE,MAAM,YAAY,CAAC;AAMvE;;;;GAIG;AACH,MAAM,WAAW,aAAa;IAC5B;;;;;OAKG;IACH,SAAS,CAAC,EAAE,MAAM,CAAC;CACpB;AAED;;;;GAIG;AACH,MAAM,WAAW,cAAc;IAC7B,wCAAwC;IACxC,SAAS,EAAE,aAAa,EAAE,CAAC;IAE3B;;;;;;OAMG;IACH,cAAc,EAAE,MAAM,CAAC;IAEvB;;OAEG;IACH,UAAU,EAAE,MAAM,CAAC;IAEnB,yDAAyD;IACzD,MAAM,EAAE,OAAO,CAAC;CACjB;AAMD;;;;;;;GAOG;AACH,qBAAa,oBAAoB;IAC/B,OAAO,CAAC,OAAO,CAAsD;IACrE,OAAO,CAAC,OAAO,CAAwB;IACvC,OAAO,CAAC,MAAM,CAAS;IAEvB;;;;OAIG;gBACS,OAAO,GAAE,gBAAqB;IAI1C;;;;;;OAMG;IACH,GAAG,CAAC,IAAI,EAAE,UAAU,GAAG,MAAM,EAAE,OAAO,CAAC,EAAE,OAAO,GAAG,IAAI;IAWvD;;;;;;;OAOG;IACH,MAAM,IAAI,qBAAqB;IAO/B;;;;;;;;;;;;;;;;;;;;;;;;;;OA0BG;IACI,MAAM,CACX,QAAQ,EAAE,SAAS,EACnB,aAAa,CAAC,EAAE,aAAa,GAC5B,cAAc,CAAC,cAAc,CAAC;IAiGjC,OAAO,CAAC,SAAS;YAYH,iBAAiB;YAWjB,aAAa;IAsB3B,OAAO,CAAC,QAAQ;IAShB,OAAO,CAAC,eAAe;CAKxB;AAMD;;;;GAIG;AACH,MAAM,MAAM,qBAAqB,GAAG,IAAI,CACtC,oBAAoB,EACpB,KAAK,GAAG,QAAQ,CACjB,GAAG;IAAE,QAAQ,CAAC,MAAM,EAAE,IAAI,CAAA;CAAE,CAAC"}
|