@absolutejs/rag 0.10.0 → 0.12.0

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
@@ -0,0 +1,341 @@
1
+ // @bun
2
+ var __create = Object.create;
3
+ var __getProtoOf = Object.getPrototypeOf;
4
+ var __defProp = Object.defineProperty;
5
+ var __getOwnPropNames = Object.getOwnPropertyNames;
6
+ var __hasOwnProp = Object.prototype.hasOwnProperty;
7
+ function __accessProp(key) {
8
+ return this[key];
9
+ }
10
+ var __toESMCache_node;
11
+ var __toESMCache_esm;
12
+ var __toESM = (mod, isNodeMode, target) => {
13
+ var canCache = mod != null && typeof mod === "object";
14
+ if (canCache) {
15
+ var cache = isNodeMode ? __toESMCache_node ??= new WeakMap : __toESMCache_esm ??= new WeakMap;
16
+ var cached = cache.get(mod);
17
+ if (cached)
18
+ return cached;
19
+ }
20
+ target = mod != null ? __create(__getProtoOf(mod)) : {};
21
+ const to = isNodeMode || !mod || !mod.__esModule ? __defProp(target, "default", { value: mod, enumerable: true }) : target;
22
+ if (mod && typeof mod === "object" || typeof mod === "function") {
23
+ for (let key of __getOwnPropNames(mod))
24
+ if (!__hasOwnProp.call(to, key))
25
+ __defProp(to, key, {
26
+ get: __accessProp.bind(mod, key),
27
+ enumerable: true
28
+ });
29
+ }
30
+ if (canCache)
31
+ cache.set(mod, to);
32
+ return to;
33
+ };
34
+ var __commonJS = (cb, mod) => () => (mod || cb((mod = { exports: {} }).exports, mod), mod.exports);
35
+ var __returnValue = (v) => v;
36
+ function __exportSetter(name, newValue) {
37
+ this[name] = __returnValue.bind(null, newValue);
38
+ }
39
+ var __export = (target, all) => {
40
+ for (var name in all)
41
+ __defProp(target, name, {
42
+ get: all[name],
43
+ enumerable: true,
44
+ configurable: true,
45
+ set: __exportSetter.bind(all, name)
46
+ });
47
+ };
48
+ var __require = import.meta.require;
49
+
50
+ // src/web/playwright.ts
51
+ import { chromium } from "playwright-core";
52
+
53
+ // src/web/transport.ts
54
+ import { lookup } from "dns/promises";
55
+ import { isIP } from "net";
56
+ import { request as httpRequest } from "http";
57
+ import { request as httpsRequest } from "https";
58
+ import { brotliDecompressSync, gunzipSync, inflateSync } from "zlib";
59
+
60
+ class WebReadError extends Error {
61
+ code;
62
+ constructor(code, message) {
63
+ super(message);
64
+ this.code = code;
65
+ this.name = "WebReadError";
66
+ }
67
+ }
68
+ var isPublicWebAddress = (address) => {
69
+ if (isIP(address) === 4) {
70
+ const [a = 0, b = 0, c = 0] = address.split(".").map(Number);
71
+ return !(a === 0 || a === 10 || a === 127 || a >= 224 || a === 100 && b >= 64 && b <= 127 || a === 169 && b === 254 || a === 172 && b >= 16 && b <= 31 || a === 192 && (b === 168 || b === 0 || b === 88 && c === 99) || a === 198 && (b === 18 || b === 19 || b === 51 && c === 100) || a === 203 && b === 0 && c === 113);
72
+ }
73
+ const value = address.toLowerCase();
74
+ return isIP(value) === 6 && /^[23]/u.test(value) && !/^2001:(?:db8|0):/u.test(value) && !value.startsWith("2002:");
75
+ };
76
+ var validatePublicWebUrl = (raw) => {
77
+ let url;
78
+ try {
79
+ url = new URL(raw);
80
+ } catch {
81
+ throw new WebReadError("invalid_url", "Provide a valid public HTTP or HTTPS URL.");
82
+ }
83
+ if (!["http:", "https:"].includes(url.protocol) || url.username || url.password || url.port && !["80", "443"].includes(url.port))
84
+ throw new WebReadError("blocked_url", "Only public HTTP/HTTPS URLs on standard ports are supported.");
85
+ const host = url.hostname.replace(/^\[|\]$/gu, "").toLowerCase();
86
+ if (host === "localhost" || /\.(?:localhost|local|internal)$/u.test(host) || isIP(host) && !isPublicWebAddress(host))
87
+ throw new WebReadError("blocked_url", "Private and local network destinations are not allowed.");
88
+ return url;
89
+ };
90
+ var fetchPublicWebResource = async (raw, options = {}) => {
91
+ const signal = options.signal ?? AbortSignal.timeout(15000);
92
+ const maxBytes = options.maxBytes ?? 5000000;
93
+ let url = validatePublicWebUrl(raw);
94
+ const redirects = [];
95
+ let method = options.method ?? "GET";
96
+ let body = options.body;
97
+ for (let hop = 0;hop <= 5; hop++) {
98
+ signal.throwIfAborted();
99
+ const host = url.hostname.replace(/^\[|\]$/gu, "");
100
+ const addresses = await lookup(host, { all: true });
101
+ if (!addresses.length || addresses.some(({ address }) => !isPublicWebAddress(address)))
102
+ throw new WebReadError("blocked_url", "The destination resolved to a private or reserved network.");
103
+ signal.throwIfAborted();
104
+ const selected = addresses.find(({ family }) => family === 4) ?? addresses[0];
105
+ const response = await new Promise((resolve, reject) => {
106
+ const request = url.protocol === "https:" ? httpsRequest : httpRequest;
107
+ const req = request({
108
+ protocol: url.protocol,
109
+ hostname: selected.address,
110
+ family: selected.family,
111
+ servername: isIP(host) ? undefined : host,
112
+ port: url.port || undefined,
113
+ path: url.pathname + url.search,
114
+ method,
115
+ signal,
116
+ agent: false,
117
+ headers: {
118
+ "user-agent": "onSparkReader/1.0 (+https://onspark.com)",
119
+ accept: "*/*",
120
+ "accept-encoding": "identity",
121
+ ...options.headers,
122
+ host: url.host
123
+ }
124
+ }, (res) => {
125
+ const chunks = [];
126
+ let bytes = 0;
127
+ res.on("error", reject);
128
+ res.on("data", (chunk) => {
129
+ bytes += chunk.length;
130
+ if (bytes > maxBytes) {
131
+ const error = new WebReadError("too_large", "The response exceeded the website read limit.");
132
+ res.destroy(error);
133
+ reject(error);
134
+ } else
135
+ chunks.push(chunk);
136
+ });
137
+ res.on("end", () => {
138
+ try {
139
+ let data = Buffer.concat(chunks);
140
+ const encoding = res.headers["content-encoding"];
141
+ const limits = { maxOutputLength: maxBytes };
142
+ if (encoding === "gzip")
143
+ data = gunzipSync(data, limits);
144
+ else if (encoding === "br")
145
+ data = brotliDecompressSync(data, limits);
146
+ else if (encoding === "deflate")
147
+ data = inflateSync(data, limits);
148
+ const headers = Object.fromEntries(Object.entries(res.headers).filter(([key]) => ![
149
+ "content-encoding",
150
+ "content-length",
151
+ "transfer-encoding",
152
+ "set-cookie"
153
+ ].includes(key)).map(([key, value]) => [
154
+ key,
155
+ Array.isArray(value) ? value.join(", ") : value ?? ""
156
+ ]));
157
+ resolve({
158
+ url: url.href,
159
+ status: res.statusCode ?? 502,
160
+ headers,
161
+ body: data
162
+ });
163
+ } catch (error) {
164
+ reject(error);
165
+ }
166
+ });
167
+ });
168
+ req.on("error", reject);
169
+ if (body)
170
+ req.write(body);
171
+ req.end();
172
+ });
173
+ if (options.redirect === "manual")
174
+ return response;
175
+ if ([301, 302, 303, 307, 308].includes(response.status) && response.headers.location) {
176
+ const next = validatePublicWebUrl(new URL(response.headers.location, url).href);
177
+ redirects.push({
178
+ from: url.href,
179
+ to: next.href,
180
+ kind: "http",
181
+ status: response.status
182
+ });
183
+ url = next;
184
+ if (response.status === 303 || [301, 302].includes(response.status) && method === "POST") {
185
+ method = "GET";
186
+ body = undefined;
187
+ }
188
+ continue;
189
+ }
190
+ return { ...response, redirects };
191
+ }
192
+ throw new WebReadError("redirect_limit", "The site exceeded the redirect limit.");
193
+ };
194
+
195
+ // src/web/playwright.ts
196
+ var createPlaywrightWebRenderer = (options = {}) => {
197
+ let browser;
198
+ let launching;
199
+ const launch = async () => {
200
+ if (browser?.isConnected())
201
+ return browser;
202
+ launching ??= chromium.launch({
203
+ headless: true,
204
+ executablePath: options.executablePath,
205
+ args: ["--disable-dev-shm-usage"]
206
+ });
207
+ try {
208
+ browser = await launching;
209
+ return browser;
210
+ } finally {
211
+ launching = undefined;
212
+ }
213
+ };
214
+ const renderOnce = async (url, { signal }) => {
215
+ validatePublicWebUrl(url);
216
+ signal.throwIfAborted();
217
+ const instance = await launch();
218
+ const context = await instance.newContext({
219
+ serviceWorkers: "block",
220
+ acceptDownloads: false
221
+ });
222
+ const abort = () => {
223
+ context.close().catch(() => {});
224
+ };
225
+ signal.addEventListener("abort", abort, { once: true });
226
+ const redirects = [];
227
+ let count = 0, bytes = 0;
228
+ try {
229
+ await context.routeWebSocket("**/*", (socket) => socket.close());
230
+ await context.route("**/*", async (route) => {
231
+ const request = route.request();
232
+ try {
233
+ if (++count > 80 || bytes > 20000000 || ["image", "media", "font"].includes(request.resourceType()))
234
+ return await route.abort();
235
+ const body = request.postData() ?? undefined;
236
+ if (body && body.length > 1e5)
237
+ return await route.abort();
238
+ const original = request.headers();
239
+ const headers = {};
240
+ for (const key of ["accept", "content-type", "origin", "user-agent"])
241
+ if (original[key])
242
+ headers[key] = original[key];
243
+ const response2 = await fetchPublicWebResource(request.url(), {
244
+ signal,
245
+ maxBytes: 5000000,
246
+ method: request.method(),
247
+ body,
248
+ headers,
249
+ redirect: "manual"
250
+ });
251
+ if (request.isNavigationRequest() && request.frame() === context.pages()[0]?.mainFrame() && [301, 302, 303, 307, 308].includes(response2.status) && response2.headers.location) {
252
+ redirects.push({
253
+ from: request.url(),
254
+ to: new URL(response2.headers.location, request.url()).href,
255
+ kind: "http",
256
+ status: response2.status
257
+ });
258
+ }
259
+ bytes += response2.body.byteLength;
260
+ await route.fulfill({
261
+ status: response2.status,
262
+ headers: response2.headers,
263
+ body: Buffer.from(response2.body)
264
+ });
265
+ } catch {
266
+ await route.abort().catch(() => {});
267
+ }
268
+ });
269
+ const page = await context.newPage();
270
+ let previousUrl = url;
271
+ page.on("framenavigated", (frame) => {
272
+ if (frame !== page.mainFrame() || !/^https?:/u.test(frame.url()))
273
+ return;
274
+ const next = frame.url();
275
+ if (next !== previousUrl && !redirects.some((hop) => hop.to === next))
276
+ redirects.push({ from: previousUrl, to: next, kind: "client" });
277
+ previousUrl = next;
278
+ });
279
+ const response = await page.goto(url, {
280
+ waitUntil: "domcontentloaded",
281
+ timeout: 25000
282
+ });
283
+ const settled = await page.waitForFunction(() => {
284
+ const root = document.querySelector("#root,#app,#__next,app-root,[ng-version]");
285
+ const rootText = root?.textContent?.trim() ?? "";
286
+ return document.body.innerText.trim().length > 100 && (!root || rootText.length > 20 && !/^(loading|please wait)/i.test(rootText));
287
+ }, undefined, { timeout: 8000 }).then(() => true, () => false);
288
+ await page.waitForTimeout(1200);
289
+ for (let step = 0;step < 3; step++) {
290
+ await page.evaluate(() => window.scrollBy(0, window.innerHeight));
291
+ await page.waitForTimeout(350);
292
+ }
293
+ await page.evaluate(() => {
294
+ document.querySelectorAll("details").forEach((element) => {
295
+ element.open = true;
296
+ });
297
+ window.scrollTo(0, 0);
298
+ });
299
+ signal.throwIfAborted();
300
+ const html = await page.content();
301
+ if (html.length > 5000000)
302
+ throw new WebReadError("too_large", "Rendered page exceeded the content limit.");
303
+ return {
304
+ html,
305
+ redirects,
306
+ settled,
307
+ text: await page.locator("body").innerText(),
308
+ url: page.url(),
309
+ status: response?.status() ?? 200
310
+ };
311
+ } finally {
312
+ signal.removeEventListener("abort", abort);
313
+ await context.close().catch(() => {});
314
+ }
315
+ };
316
+ const render = async (url, options2) => {
317
+ try {
318
+ return await renderOnce(url, options2);
319
+ } catch (error) {
320
+ if (options2.signal.aborted || browser?.isConnected())
321
+ throw error;
322
+ return renderOnce(url, options2);
323
+ }
324
+ };
325
+ return {
326
+ render,
327
+ ready: async () => {
328
+ await launch();
329
+ },
330
+ close: async () => {
331
+ await browser?.close();
332
+ browser = undefined;
333
+ }
334
+ };
335
+ };
336
+ export {
337
+ createPlaywrightWebRenderer
338
+ };
339
+
340
+ //# debugId=7A89F48E5D741A4264756E2164756E21
341
+ //# sourceMappingURL=playwright.js.map
@@ -0,0 +1,11 @@
1
+ {
2
+ "version": 3,
3
+ "sources": ["../src/web/playwright.ts", "../src/web/transport.ts"],
4
+ "sourcesContent": [
5
+ "import { chromium, type Browser } from \"playwright-core\";\nimport {\n fetchPublicWebResource,\n validatePublicWebUrl,\n WebReadError,\n} from \"./transport\";\nimport type { WebRedirect } from \"./transport\";\nimport type { WebRenderer } from \"./index\";\n/** Fresh contexts share a browser, never cookies; all HTTP traffic uses pinned public DNS. */\nexport const createPlaywrightWebRenderer = (\n options: { executablePath?: string } = {},\n) => {\n let browser: Browser | undefined;\n let launching: Promise<Browser> | undefined;\n const launch = async () => {\n if (browser?.isConnected()) return browser;\n launching ??= chromium.launch({\n headless: true,\n executablePath: options.executablePath,\n args: [\"--disable-dev-shm-usage\"],\n });\n try {\n browser = await launching;\n return browser;\n } finally {\n launching = undefined;\n }\n };\n const renderOnce: WebRenderer = async (url, { signal }) => {\n validatePublicWebUrl(url);\n signal.throwIfAborted();\n const instance = await launch();\n const context = await instance.newContext({\n serviceWorkers: \"block\",\n acceptDownloads: false,\n });\n const abort = () => {\n void context.close().catch(() => {});\n };\n signal.addEventListener(\"abort\", abort, { once: true });\n const redirects: WebRedirect[] = [];\n let count = 0,\n bytes = 0;\n try {\n await context.routeWebSocket(\"**/*\", (socket) => socket.close());\n await context.route(\"**/*\", async (route) => {\n const request = route.request();\n try {\n if (\n ++count > 80 ||\n bytes > 20_000_000 ||\n [\"image\", \"media\", \"font\"].includes(request.resourceType())\n )\n return await route.abort();\n const body = request.postData() ?? undefined;\n if (body && body.length > 100_000) return await route.abort();\n const original = request.headers();\n const headers: Record<string, string> = {};\n for (const key of [\"accept\", \"content-type\", \"origin\", \"user-agent\"])\n if (original[key]) headers[key] = original[key];\n const response = await fetchPublicWebResource(request.url(), {\n signal,\n maxBytes: 5_000_000,\n method: request.method(),\n body,\n headers,\n redirect: \"manual\",\n });\n if (\n request.isNavigationRequest() &&\n request.frame() === context.pages()[0]?.mainFrame() &&\n [301, 302, 303, 307, 308].includes(response.status) &&\n response.headers.location\n ) {\n redirects.push({\n from: request.url(),\n to: new URL(response.headers.location, request.url()).href,\n kind: \"http\",\n status: response.status,\n });\n }\n bytes += response.body.byteLength;\n await route.fulfill({\n status: response.status,\n headers: response.headers,\n body: Buffer.from(response.body),\n });\n } catch {\n await route.abort().catch(() => {});\n }\n });\n const page = await context.newPage();\n let previousUrl = url;\n page.on(\"framenavigated\", (frame) => {\n if (frame !== page.mainFrame() || !/^https?:/u.test(frame.url()))\n return;\n const next = frame.url();\n if (next !== previousUrl && !redirects.some((hop) => hop.to === next))\n redirects.push({ from: previousUrl, to: next, kind: \"client\" });\n previousUrl = next;\n });\n const response = await page.goto(url, {\n waitUntil: \"domcontentloaded\",\n timeout: 25000,\n });\n const settled = await page\n .waitForFunction(\n () => {\n const root = document.querySelector(\n \"#root,#app,#__next,app-root,[ng-version]\",\n );\n const rootText = root?.textContent?.trim() ?? \"\";\n return (\n document.body.innerText.trim().length > 100 &&\n (!root ||\n (rootText.length > 20 &&\n !/^(loading|please wait)/i.test(rootText)))\n );\n },\n undefined,\n { timeout: 8000 },\n )\n .then(\n () => true,\n () => false,\n );\n // Allow hydration and async content to settle after the first readable frame.\n await page.waitForTimeout(1200);\n // Bounded scrolling exposes common lazy-loaded sections without clicking actions.\n for (let step = 0; step < 3; step++) {\n await page.evaluate(() => window.scrollBy(0, window.innerHeight));\n await page.waitForTimeout(350);\n }\n await page.evaluate(() => {\n document.querySelectorAll(\"details\").forEach((element) => {\n element.open = true;\n });\n window.scrollTo(0, 0);\n });\n signal.throwIfAborted();\n const html = await page.content();\n if (html.length > 5_000_000)\n throw new WebReadError(\n \"too_large\",\n \"Rendered page exceeded the content limit.\",\n );\n return {\n html,\n redirects,\n settled,\n text: await page.locator(\"body\").innerText(),\n url: page.url(),\n status: response?.status() ?? 200,\n };\n } finally {\n signal.removeEventListener(\"abort\", abort);\n await context.close().catch(() => {});\n }\n };\n const render: WebRenderer = async (url, options) => {\n try {\n return await renderOnce(url, options);\n } catch (error) {\n if (options.signal.aborted || browser?.isConnected()) throw error;\n // A crashed worker may be replaced once; the same total deadline still applies.\n return renderOnce(url, options);\n }\n };\n return {\n render,\n ready: async () => {\n await launch();\n },\n close: async () => {\n await browser?.close();\n browser = undefined;\n },\n };\n};\n",
6
+ "import { lookup } from \"node:dns/promises\";\nimport { isIP } from \"node:net\";\nimport { request as httpRequest } from \"node:http\";\nimport { request as httpsRequest } from \"node:https\";\nimport { brotliDecompressSync, gunzipSync, inflateSync } from \"node:zlib\";\n\nexport class WebReadError extends Error {\n constructor(\n public code: string,\n message: string,\n ) {\n super(message);\n this.name = \"WebReadError\";\n }\n}\nexport const isPublicWebAddress = (address: string) => {\n if (isIP(address) === 4) {\n const [a = 0, b = 0, c = 0] = address.split(\".\").map(Number);\n return !(\n a === 0 ||\n a === 10 ||\n a === 127 ||\n a >= 224 ||\n (a === 100 && b >= 64 && b <= 127) ||\n (a === 169 && b === 254) ||\n (a === 172 && b >= 16 && b <= 31) ||\n (a === 192 && (b === 168 || b === 0 || (b === 88 && c === 99))) ||\n (a === 198 && (b === 18 || b === 19 || (b === 51 && c === 100))) ||\n (a === 203 && b === 0 && c === 113)\n );\n }\n const value = address.toLowerCase();\n return (\n isIP(value) === 6 &&\n /^[23]/u.test(value) &&\n !/^2001:(?:db8|0):/u.test(value) &&\n !value.startsWith(\"2002:\")\n );\n};\nexport type WebRedirect = {\n from: string;\n to: string;\n kind: \"http\" | \"client\";\n status?: number;\n};\nexport type WebFetchResult = {\n redirects?: WebRedirect[];\n url: string;\n status: number;\n headers: Record<string, string>;\n body: Uint8Array;\n};\nexport type WebFetchOptions = {\n signal?: AbortSignal;\n redirect?: \"follow\" | \"manual\";\n maxBytes?: number;\n method?: string;\n body?: string;\n headers?: Record<string, string>;\n};\nexport const validatePublicWebUrl = (raw: string) => {\n let url: URL;\n try {\n url = new URL(raw);\n } catch {\n throw new WebReadError(\n \"invalid_url\",\n \"Provide a valid public HTTP or HTTPS URL.\",\n );\n }\n if (\n ![\"http:\", \"https:\"].includes(url.protocol) ||\n url.username ||\n url.password ||\n (url.port && ![\"80\", \"443\"].includes(url.port))\n )\n throw new WebReadError(\n \"blocked_url\",\n \"Only public HTTP/HTTPS URLs on standard ports are supported.\",\n );\n const host = url.hostname.replace(/^\\[|\\]$/gu, \"\").toLowerCase();\n if (\n host === \"localhost\" ||\n /\\.(?:localhost|local|internal)$/u.test(host) ||\n (isIP(host) && !isPublicWebAddress(host))\n )\n throw new WebReadError(\n \"blocked_url\",\n \"Private and local network destinations are not allowed.\",\n );\n return url;\n};\n/** Pin the validated DNS answer into the connection, including every redirect. */\nexport const fetchPublicWebResource = async (\n raw: string,\n options: WebFetchOptions = {},\n): Promise<WebFetchResult> => {\n const signal = options.signal ?? AbortSignal.timeout(15000);\n const maxBytes = options.maxBytes ?? 5_000_000;\n let url = validatePublicWebUrl(raw);\n const redirects: WebRedirect[] = [];\n let method = options.method ?? \"GET\";\n let body = options.body;\n for (let hop = 0; hop <= 5; hop++) {\n signal.throwIfAborted();\n const host = url.hostname.replace(/^\\[|\\]$/gu, \"\");\n const addresses = await lookup(host, { all: true });\n if (\n !addresses.length ||\n addresses.some(({ address }) => !isPublicWebAddress(address))\n )\n throw new WebReadError(\n \"blocked_url\",\n \"The destination resolved to a private or reserved network.\",\n );\n signal.throwIfAborted();\n const selected =\n addresses.find(({ family }) => family === 4) ?? addresses[0]!;\n const response = await new Promise<WebFetchResult>((resolve, reject) => {\n const request = url.protocol === \"https:\" ? httpsRequest : httpRequest;\n const req = request(\n {\n protocol: url.protocol,\n hostname: selected.address,\n family: selected.family,\n servername: isIP(host) ? undefined : host,\n port: url.port || undefined,\n path: url.pathname + url.search,\n method,\n signal,\n agent: false,\n headers: {\n \"user-agent\": \"onSparkReader/1.0 (+https://onspark.com)\",\n accept: \"*/*\",\n \"accept-encoding\": \"identity\",\n ...options.headers,\n host: url.host,\n },\n },\n (res) => {\n const chunks: Buffer[] = [];\n let bytes = 0;\n res.on(\"error\", reject);\n res.on(\"data\", (chunk: Buffer) => {\n bytes += chunk.length;\n if (bytes > maxBytes) {\n const error = new WebReadError(\n \"too_large\",\n \"The response exceeded the website read limit.\",\n );\n res.destroy(error);\n reject(error);\n } else chunks.push(chunk);\n });\n res.on(\"end\", () => {\n try {\n let data: Uint8Array = Buffer.concat(chunks);\n const encoding = res.headers[\"content-encoding\"];\n const limits = { maxOutputLength: maxBytes };\n if (encoding === \"gzip\") data = gunzipSync(data, limits);\n else if (encoding === \"br\")\n data = brotliDecompressSync(data, limits);\n else if (encoding === \"deflate\") data = inflateSync(data, limits);\n const headers = Object.fromEntries(\n Object.entries(res.headers)\n .filter(\n ([key]) =>\n ![\n \"content-encoding\",\n \"content-length\",\n \"transfer-encoding\",\n \"set-cookie\",\n ].includes(key),\n )\n .map(([key, value]) => [\n key,\n Array.isArray(value) ? value.join(\", \") : (value ?? \"\"),\n ]),\n );\n resolve({\n url: url.href,\n status: res.statusCode ?? 502,\n headers,\n body: data,\n });\n } catch (error) {\n reject(error);\n }\n });\n },\n );\n req.on(\"error\", reject);\n if (body) req.write(body);\n req.end();\n });\n if (options.redirect === \"manual\") return response;\n if (\n [301, 302, 303, 307, 308].includes(response.status) &&\n response.headers.location\n ) {\n const next = validatePublicWebUrl(\n new URL(response.headers.location, url).href,\n );\n redirects.push({\n from: url.href,\n to: next.href,\n kind: \"http\",\n status: response.status,\n });\n url = next;\n if (\n response.status === 303 ||\n ([301, 302].includes(response.status) && method === \"POST\")\n ) {\n method = \"GET\";\n body = undefined;\n }\n continue;\n }\n return { ...response, redirects };\n }\n throw new WebReadError(\n \"redirect_limit\",\n \"The site exceeded the redirect limit.\",\n );\n};\n"
7
+ ],
8
+ "mappings": ";;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;AAAA;;;ACAA;AACA;AACA,oBAAS;AACT,oBAAS;AACT;AAAA;AAEO,MAAM,qBAAqB,MAAM;AAAA,EAE7B;AAAA,EADT,WAAW,CACF,MACP,SACA;AAAA,IACA,MAAM,OAAO;AAAA,IAHN;AAAA,IAIP,KAAK,OAAO;AAAA;AAEhB;AACO,IAAM,qBAAqB,CAAC,YAAoB;AAAA,EACrD,IAAI,KAAK,OAAO,MAAM,GAAG;AAAA,IACvB,OAAO,IAAI,GAAG,IAAI,GAAG,IAAI,KAAK,QAAQ,MAAM,GAAG,EAAE,IAAI,MAAM;AAAA,IAC3D,OAAO,EACL,MAAM,KACN,MAAM,MACN,MAAM,OACN,KAAK,OACJ,MAAM,OAAO,KAAK,MAAM,KAAK,OAC7B,MAAM,OAAO,MAAM,OACnB,MAAM,OAAO,KAAK,MAAM,KAAK,MAC7B,MAAM,QAAQ,MAAM,OAAO,MAAM,KAAM,MAAM,MAAM,MAAM,OACzD,MAAM,QAAQ,MAAM,MAAM,MAAM,MAAO,MAAM,MAAM,MAAM,QACzD,MAAM,OAAO,MAAM,KAAK,MAAM;AAAA,EAEnC;AAAA,EACA,MAAM,QAAQ,QAAQ,YAAY;AAAA,EAClC,OACE,KAAK,KAAK,MAAM,KAChB,SAAS,KAAK,KAAK,KACnB,CAAC,oBAAoB,KAAK,KAAK,KAC/B,CAAC,MAAM,WAAW,OAAO;AAAA;AAwBtB,IAAM,uBAAuB,CAAC,QAAgB;AAAA,EACnD,IAAI;AAAA,EACJ,IAAI;AAAA,IACF,MAAM,IAAI,IAAI,GAAG;AAAA,IACjB,MAAM;AAAA,IACN,MAAM,IAAI,aACR,eACA,2CACF;AAAA;AAAA,EAEF,IACE,CAAC,CAAC,SAAS,QAAQ,EAAE,SAAS,IAAI,QAAQ,KAC1C,IAAI,YACJ,IAAI,YACH,IAAI,QAAQ,CAAC,CAAC,MAAM,KAAK,EAAE,SAAS,IAAI,IAAI;AAAA,IAE7C,MAAM,IAAI,aACR,eACA,8DACF;AAAA,EACF,MAAM,OAAO,IAAI,SAAS,QAAQ,aAAa,EAAE,EAAE,YAAY;AAAA,EAC/D,IACE,SAAS,eACT,mCAAmC,KAAK,IAAI,KAC3C,KAAK,IAAI,KAAK,CAAC,mBAAmB,IAAI;AAAA,IAEvC,MAAM,IAAI,aACR,eACA,yDACF;AAAA,EACF,OAAO;AAAA;AAGF,IAAM,yBAAyB,OACpC,KACA,UAA2B,CAAC,MACA;AAAA,EAC5B,MAAM,SAAS,QAAQ,UAAU,YAAY,QAAQ,KAAK;AAAA,EAC1D,MAAM,WAAW,QAAQ,YAAY;AAAA,EACrC,IAAI,MAAM,qBAAqB,GAAG;AAAA,EAClC,MAAM,YAA2B,CAAC;AAAA,EAClC,IAAI,SAAS,QAAQ,UAAU;AAAA,EAC/B,IAAI,OAAO,QAAQ;AAAA,EACnB,SAAS,MAAM,EAAG,OAAO,GAAG,OAAO;AAAA,IACjC,OAAO,eAAe;AAAA,IACtB,MAAM,OAAO,IAAI,SAAS,QAAQ,aAAa,EAAE;AAAA,IACjD,MAAM,YAAY,MAAM,OAAO,MAAM,EAAE,KAAK,KAAK,CAAC;AAAA,IAClD,IACE,CAAC,UAAU,UACX,UAAU,KAAK,GAAG,cAAc,CAAC,mBAAmB,OAAO,CAAC;AAAA,MAE5D,MAAM,IAAI,aACR,eACA,4DACF;AAAA,IACF,OAAO,eAAe;AAAA,IACtB,MAAM,WACJ,UAAU,KAAK,GAAG,aAAa,WAAW,CAAC,KAAK,UAAU;AAAA,IAC5D,MAAM,WAAW,MAAM,IAAI,QAAwB,CAAC,SAAS,WAAW;AAAA,MACtE,MAAM,UAAU,IAAI,aAAa,WAAW,eAAe;AAAA,MAC3D,MAAM,MAAM,QACV;AAAA,QACE,UAAU,IAAI;AAAA,QACd,UAAU,SAAS;AAAA,QACnB,QAAQ,SAAS;AAAA,QACjB,YAAY,KAAK,IAAI,IAAI,YAAY;AAAA,QACrC,MAAM,IAAI,QAAQ;AAAA,QAClB,MAAM,IAAI,WAAW,IAAI;AAAA,QACzB;AAAA,QACA;AAAA,QACA,OAAO;AAAA,QACP,SAAS;AAAA,UACP,cAAc;AAAA,UACd,QAAQ;AAAA,UACR,mBAAmB;AAAA,aAChB,QAAQ;AAAA,UACX,MAAM,IAAI;AAAA,QACZ;AAAA,MACF,GACA,CAAC,QAAQ;AAAA,QACP,MAAM,SAAmB,CAAC;AAAA,QAC1B,IAAI,QAAQ;AAAA,QACZ,IAAI,GAAG,SAAS,MAAM;AAAA,QACtB,IAAI,GAAG,QAAQ,CAAC,UAAkB;AAAA,UAChC,SAAS,MAAM;AAAA,UACf,IAAI,QAAQ,UAAU;AAAA,YACpB,MAAM,QAAQ,IAAI,aAChB,aACA,+CACF;AAAA,YACA,IAAI,QAAQ,KAAK;AAAA,YACjB,OAAO,KAAK;AAAA,UACd,EAAO;AAAA,mBAAO,KAAK,KAAK;AAAA,SACzB;AAAA,QACD,IAAI,GAAG,OAAO,MAAM;AAAA,UAClB,IAAI;AAAA,YACF,IAAI,OAAmB,OAAO,OAAO,MAAM;AAAA,YAC3C,MAAM,WAAW,IAAI,QAAQ;AAAA,YAC7B,MAAM,SAAS,EAAE,iBAAiB,SAAS;AAAA,YAC3C,IAAI,aAAa;AAAA,cAAQ,OAAO,WAAW,MAAM,MAAM;AAAA,YAClD,SAAI,aAAa;AAAA,cACpB,OAAO,qBAAqB,MAAM,MAAM;AAAA,YACrC,SAAI,aAAa;AAAA,cAAW,OAAO,YAAY,MAAM,MAAM;AAAA,YAChE,MAAM,UAAU,OAAO,YACrB,OAAO,QAAQ,IAAI,OAAO,EACvB,OACC,EAAE,SACA,CAAC;AAAA,cACC;AAAA,cACA;AAAA,cACA;AAAA,cACA;AAAA,YACF,EAAE,SAAS,GAAG,CAClB,EACC,IAAI,EAAE,KAAK,WAAW;AAAA,cACrB;AAAA,cACA,MAAM,QAAQ,KAAK,IAAI,MAAM,KAAK,IAAI,IAAK,SAAS;AAAA,YACtD,CAAC,CACL;AAAA,YACA,QAAQ;AAAA,cACN,KAAK,IAAI;AAAA,cACT,QAAQ,IAAI,cAAc;AAAA,cAC1B;AAAA,cACA,MAAM;AAAA,YACR,CAAC;AAAA,YACD,OAAO,OAAO;AAAA,YACd,OAAO,KAAK;AAAA;AAAA,SAEf;AAAA,OAEL;AAAA,MACA,IAAI,GAAG,SAAS,MAAM;AAAA,MACtB,IAAI;AAAA,QAAM,IAAI,MAAM,IAAI;AAAA,MACxB,IAAI,IAAI;AAAA,KACT;AAAA,IACD,IAAI,QAAQ,aAAa;AAAA,MAAU,OAAO;AAAA,IAC1C,IACE,CAAC,KAAK,KAAK,KAAK,KAAK,GAAG,EAAE,SAAS,SAAS,MAAM,KAClD,SAAS,QAAQ,UACjB;AAAA,MACA,MAAM,OAAO,qBACX,IAAI,IAAI,SAAS,QAAQ,UAAU,GAAG,EAAE,IAC1C;AAAA,MACA,UAAU,KAAK;AAAA,QACb,MAAM,IAAI;AAAA,QACV,IAAI,KAAK;AAAA,QACT,MAAM;AAAA,QACN,QAAQ,SAAS;AAAA,MACnB,CAAC;AAAA,MACD,MAAM;AAAA,MACN,IACE,SAAS,WAAW,OACnB,CAAC,KAAK,GAAG,EAAE,SAAS,SAAS,MAAM,KAAK,WAAW,QACpD;AAAA,QACA,SAAS;AAAA,QACT,OAAO;AAAA,MACT;AAAA,MACA;AAAA,IACF;AAAA,IACA,OAAO,KAAK,UAAU,UAAU;AAAA,EAClC;AAAA,EACA,MAAM,IAAI,aACR,kBACA,uCACF;AAAA;;;ADvNK,IAAM,8BAA8B,CACzC,UAAuC,CAAC,MACrC;AAAA,EACH,IAAI;AAAA,EACJ,IAAI;AAAA,EACJ,MAAM,SAAS,YAAY;AAAA,IACzB,IAAI,SAAS,YAAY;AAAA,MAAG,OAAO;AAAA,IACnC,cAAc,SAAS,OAAO;AAAA,MAC5B,UAAU;AAAA,MACV,gBAAgB,QAAQ;AAAA,MACxB,MAAM,CAAC,yBAAyB;AAAA,IAClC,CAAC;AAAA,IACD,IAAI;AAAA,MACF,UAAU,MAAM;AAAA,MAChB,OAAO;AAAA,cACP;AAAA,MACA,YAAY;AAAA;AAAA;AAAA,EAGhB,MAAM,aAA0B,OAAO,OAAO,aAAa;AAAA,IACzD,qBAAqB,GAAG;AAAA,IACxB,OAAO,eAAe;AAAA,IACtB,MAAM,WAAW,MAAM,OAAO;AAAA,IAC9B,MAAM,UAAU,MAAM,SAAS,WAAW;AAAA,MACxC,gBAAgB;AAAA,MAChB,iBAAiB;AAAA,IACnB,CAAC;AAAA,IACD,MAAM,QAAQ,MAAM;AAAA,MACb,QAAQ,MAAM,EAAE,MAAM,MAAM,EAAE;AAAA;AAAA,IAErC,OAAO,iBAAiB,SAAS,OAAO,EAAE,MAAM,KAAK,CAAC;AAAA,IACtD,MAAM,YAA2B,CAAC;AAAA,IAClC,IAAI,QAAQ,GACV,QAAQ;AAAA,IACV,IAAI;AAAA,MACF,MAAM,QAAQ,eAAe,QAAQ,CAAC,WAAW,OAAO,MAAM,CAAC;AAAA,MAC/D,MAAM,QAAQ,MAAM,QAAQ,OAAO,UAAU;AAAA,QAC3C,MAAM,UAAU,MAAM,QAAQ;AAAA,QAC9B,IAAI;AAAA,UACF,IACE,EAAE,QAAQ,MACV,QAAQ,YACR,CAAC,SAAS,SAAS,MAAM,EAAE,SAAS,QAAQ,aAAa,CAAC;AAAA,YAE1D,OAAO,MAAM,MAAM,MAAM;AAAA,UAC3B,MAAM,OAAO,QAAQ,SAAS,KAAK;AAAA,UACnC,IAAI,QAAQ,KAAK,SAAS;AAAA,YAAS,OAAO,MAAM,MAAM,MAAM;AAAA,UAC5D,MAAM,WAAW,QAAQ,QAAQ;AAAA,UACjC,MAAM,UAAkC,CAAC;AAAA,UACzC,WAAW,OAAO,CAAC,UAAU,gBAAgB,UAAU,YAAY;AAAA,YACjE,IAAI,SAAS;AAAA,cAAM,QAAQ,OAAO,SAAS;AAAA,UAC7C,MAAM,YAAW,MAAM,uBAAuB,QAAQ,IAAI,GAAG;AAAA,YAC3D;AAAA,YACA,UAAU;AAAA,YACV,QAAQ,QAAQ,OAAO;AAAA,YACvB;AAAA,YACA;AAAA,YACA,UAAU;AAAA,UACZ,CAAC;AAAA,UACD,IACE,QAAQ,oBAAoB,KAC5B,QAAQ,MAAM,MAAM,QAAQ,MAAM,EAAE,IAAI,UAAU,KAClD,CAAC,KAAK,KAAK,KAAK,KAAK,GAAG,EAAE,SAAS,UAAS,MAAM,KAClD,UAAS,QAAQ,UACjB;AAAA,YACA,UAAU,KAAK;AAAA,cACb,MAAM,QAAQ,IAAI;AAAA,cAClB,IAAI,IAAI,IAAI,UAAS,QAAQ,UAAU,QAAQ,IAAI,CAAC,EAAE;AAAA,cACtD,MAAM;AAAA,cACN,QAAQ,UAAS;AAAA,YACnB,CAAC;AAAA,UACH;AAAA,UACA,SAAS,UAAS,KAAK;AAAA,UACvB,MAAM,MAAM,QAAQ;AAAA,YAClB,QAAQ,UAAS;AAAA,YACjB,SAAS,UAAS;AAAA,YAClB,MAAM,OAAO,KAAK,UAAS,IAAI;AAAA,UACjC,CAAC;AAAA,UACD,MAAM;AAAA,UACN,MAAM,MAAM,MAAM,EAAE,MAAM,MAAM,EAAE;AAAA;AAAA,OAErC;AAAA,MACD,MAAM,OAAO,MAAM,QAAQ,QAAQ;AAAA,MACnC,IAAI,cAAc;AAAA,MAClB,KAAK,GAAG,kBAAkB,CAAC,UAAU;AAAA,QACnC,IAAI,UAAU,KAAK,UAAU,KAAK,CAAC,YAAY,KAAK,MAAM,IAAI,CAAC;AAAA,UAC7D;AAAA,QACF,MAAM,OAAO,MAAM,IAAI;AAAA,QACvB,IAAI,SAAS,eAAe,CAAC,UAAU,KAAK,CAAC,QAAQ,IAAI,OAAO,IAAI;AAAA,UAClE,UAAU,KAAK,EAAE,MAAM,aAAa,IAAI,MAAM,MAAM,SAAS,CAAC;AAAA,QAChE,cAAc;AAAA,OACf;AAAA,MACD,MAAM,WAAW,MAAM,KAAK,KAAK,KAAK;AAAA,QACpC,WAAW;AAAA,QACX,SAAS;AAAA,MACX,CAAC;AAAA,MACD,MAAM,UAAU,MAAM,KACnB,gBACC,MAAM;AAAA,QACJ,MAAM,OAAO,SAAS,cACpB,0CACF;AAAA,QACA,MAAM,WAAW,MAAM,aAAa,KAAK,KAAK;AAAA,QAC9C,OACE,SAAS,KAAK,UAAU,KAAK,EAAE,SAAS,QACvC,CAAC,QACC,SAAS,SAAS,MACjB,CAAC,0BAA0B,KAAK,QAAQ;AAAA,SAGhD,WACA,EAAE,SAAS,KAAK,CAClB,EACC,KACC,MAAM,MACN,MAAM,KACR;AAAA,MAEF,MAAM,KAAK,eAAe,IAAI;AAAA,MAE9B,SAAS,OAAO,EAAG,OAAO,GAAG,QAAQ;AAAA,QACnC,MAAM,KAAK,SAAS,MAAM,OAAO,SAAS,GAAG,OAAO,WAAW,CAAC;AAAA,QAChE,MAAM,KAAK,eAAe,GAAG;AAAA,MAC/B;AAAA,MACA,MAAM,KAAK,SAAS,MAAM;AAAA,QACxB,SAAS,iBAAiB,SAAS,EAAE,QAAQ,CAAC,YAAY;AAAA,UACxD,QAAQ,OAAO;AAAA,SAChB;AAAA,QACD,OAAO,SAAS,GAAG,CAAC;AAAA,OACrB;AAAA,MACD,OAAO,eAAe;AAAA,MACtB,MAAM,OAAO,MAAM,KAAK,QAAQ;AAAA,MAChC,IAAI,KAAK,SAAS;AAAA,QAChB,MAAM,IAAI,aACR,aACA,2CACF;AAAA,MACF,OAAO;AAAA,QACL;AAAA,QACA;AAAA,QACA;AAAA,QACA,MAAM,MAAM,KAAK,QAAQ,MAAM,EAAE,UAAU;AAAA,QAC3C,KAAK,KAAK,IAAI;AAAA,QACd,QAAQ,UAAU,OAAO,KAAK;AAAA,MAChC;AAAA,cACA;AAAA,MACA,OAAO,oBAAoB,SAAS,KAAK;AAAA,MACzC,MAAM,QAAQ,MAAM,EAAE,MAAM,MAAM,EAAE;AAAA;AAAA;AAAA,EAGxC,MAAM,SAAsB,OAAO,KAAK,aAAY;AAAA,IAClD,IAAI;AAAA,MACF,OAAO,MAAM,WAAW,KAAK,QAAO;AAAA,MACpC,OAAO,OAAO;AAAA,MACd,IAAI,SAAQ,OAAO,WAAW,SAAS,YAAY;AAAA,QAAG,MAAM;AAAA,MAE5D,OAAO,WAAW,KAAK,QAAO;AAAA;AAAA;AAAA,EAGlC,OAAO;AAAA,IACL;AAAA,IACA,OAAO,YAAY;AAAA,MACjB,MAAM,OAAO;AAAA;AAAA,IAEf,OAAO,YAAY;AAAA,MACjB,MAAM,SAAS,MAAM;AAAA,MACrB,UAAU;AAAA;AAAA,EAEd;AAAA;",
9
+ "debugId": "7A89F48E5D741A4264756E2164756E21",
10
+ "names": []
11
+ }
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "@absolutejs/rag",
3
- "version": "0.10.0",
3
+ "version": "0.12.0",
4
4
  "homepage": "https://github.com/absolutejs/rag",
5
5
  "bugs": {
6
6
  "url": "https://github.com/absolutejs/rag/issues"
@@ -61,14 +61,23 @@
61
61
  "import": "./dist/manifest.js",
62
62
  "types": "./dist/src/manifest.d.ts"
63
63
  },
64
- "./manifest.json": "./dist/manifest.json"
64
+ "./manifest.json": "./dist/manifest.json",
65
+ "./web": {
66
+ "import": "./dist/web/index.js",
67
+ "types": "./dist/src/web/index.d.ts"
68
+ },
69
+ "./web/playwright": {
70
+ "import": "./dist/web/playwright.js",
71
+ "types": "./dist/src/web/playwright.d.ts"
72
+ }
65
73
  },
66
74
  "absolutejs": {
67
75
  "manifestContract": 2
68
76
  },
69
77
  "dependencies": {
70
78
  "@absolutejs/manifest": "^0.9.0",
71
- "@sinclair/typebox": "^0.34.0"
79
+ "@sinclair/typebox": "^0.34.0",
80
+ "linkedom": "0.18.13"
72
81
  },
73
82
  "peerDependencies": {
74
83
  "@angular/core": "^21.0.0",
@@ -77,7 +86,8 @@
77
86
  "svelte": "^5.35.2",
78
87
  "vue": "^3.5.27",
79
88
  "@absolutejs/sync": ">=2.14.0 <3",
80
- "@absolutejs/ai": ">=0.0.54 <0.3"
89
+ "@absolutejs/ai": ">=0.0.54 <0.3",
90
+ "playwright-core": "^1.60.0"
81
91
  },
82
92
  "peerDependenciesMeta": {
83
93
  "@angular/core": {
@@ -91,6 +101,9 @@
91
101
  },
92
102
  "vue": {
93
103
  "optional": true
104
+ },
105
+ "playwright-core": {
106
+ "optional": true
94
107
  }
95
108
  },
96
109
  "devDependencies": {
@@ -115,7 +128,8 @@
115
128
  "svelte": "5.55.0",
116
129
  "typescript": "5.9.3",
117
130
  "typescript-eslint": "^8.56.1",
118
- "vue": "3.5.27"
131
+ "vue": "3.5.27",
132
+ "playwright-core": "1.60.0"
119
133
  },
120
134
  "scripts": {
121
135
  "build": "bun run scripts/build.ts && absolute-manifest emit",