webforai 1.5.1 → 1.6.0
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/dist/bin.js +277 -334
- package/dist/cjs/cli/bin.js +3 -4
- package/dist/cjs/cli/commands/{handlers → webforai}/index.js +33 -31
- package/dist/cjs/cli/commands/webforai/loadHtml.js +93 -0
- package/dist/cjs/cli/utils.js +18 -51
- package/dist/cjs/extract-hast/index.js +4 -3
- package/dist/cjs/extract-hast/readability.js +28 -61
- package/dist/cjs/extract-hast/strongly.js +2 -1
- package/dist/cjs/extract-hast/utils.js +0 -15
- package/dist/cjs/html-to-mdast.js +6 -6
- package/dist/cjs/link-replacer.js +10 -5
- package/dist/cjs/loaders/playwright.js +36 -7
- package/dist/cjs/mdast-handlers/custom-br-handler.js +31 -0
- package/dist/cjs/mdast-to-markdown.js +1 -1
- package/dist/cjs/utils/hast-utils.js +71 -0
- package/dist/cli/bin.js +3 -4
- package/dist/cli/commands/webforai/index.js +53 -0
- package/dist/cli/commands/webforai/loadHtml.js +60 -0
- package/dist/cli/utils.js +18 -51
- package/dist/extract-hast/index.js +4 -3
- package/dist/extract-hast/readability.js +29 -62
- package/dist/extract-hast/strongly.js +2 -1
- package/dist/extract-hast/utils.js +0 -14
- package/dist/html-to-mdast.js +6 -6
- package/dist/link-replacer.js +10 -5
- package/dist/loaders/playwright.js +36 -7
- package/dist/mdast-handlers/custom-br-handler.js +8 -0
- package/dist/mdast-to-markdown.js +1 -1
- package/dist/types/cli/bin.js.map +1 -1
- package/dist/types/cli/commands/webforai/index.d.ts +6 -0
- package/dist/types/cli/commands/webforai/index.js.map +1 -0
- package/dist/types/cli/commands/webforai/loadHtml.d.ts +3 -0
- package/dist/types/cli/commands/webforai/loadHtml.js.map +1 -0
- package/dist/types/cli/utils.d.ts +27 -0
- package/dist/types/cli/utils.js.map +1 -1
- package/dist/types/extract-hast/index.d.ts +9 -4
- package/dist/types/extract-hast/index.js.map +1 -1
- package/dist/types/extract-hast/readability.d.ts +2 -1
- package/dist/types/extract-hast/readability.js.map +1 -1
- package/dist/types/extract-hast/strongly.d.ts +2 -1
- package/dist/types/extract-hast/strongly.js.map +1 -1
- package/dist/types/extract-hast/utils.d.ts +0 -1
- package/dist/types/extract-hast/utils.js.map +1 -1
- package/dist/types/html-to-mdast.d.ts +4 -0
- package/dist/types/html-to-mdast.js.map +1 -1
- package/dist/types/link-replacer.d.ts +1 -1
- package/dist/types/link-replacer.js.map +1 -1
- package/dist/types/loaders/playwright.d.ts +8 -2
- package/dist/types/loaders/playwright.js.map +1 -1
- package/dist/types/mdast-handlers/custom-br-handler.d.ts +2 -0
- package/dist/types/mdast-handlers/custom-br-handler.js.map +1 -0
- package/dist/types/mdast-to-markdown.js.map +1 -1
- package/dist/types/utils/hast-utils.d.ts +4 -0
- package/dist/types/utils/hast-utils.js.map +1 -0
- package/dist/utils/hast-utils.js +46 -0
- package/package.json +21 -10
- package/dist/cjs/cli/commands/handlers/loadHtml.js +0 -127
- package/dist/cjs/html-to-markdown.test.js +0 -118
- package/dist/cjs/link-replacer.test.js +0 -47
- package/dist/cjs/loaders/fetch.test.js +0 -9
- package/dist/cjs/loaders/playwright.test.js +0 -17
- package/dist/cjs/loaders/puppeteer.test.js +0 -17
- package/dist/cli/commands/handlers/index.js +0 -51
- package/dist/cli/commands/handlers/loadHtml.js +0 -94
- package/dist/html-to-markdown.test.js +0 -118
- package/dist/link-replacer.test.js +0 -47
- package/dist/loaders/fetch.test.js +0 -9
- package/dist/loaders/playwright.test.js +0 -17
- package/dist/loaders/puppeteer.test.js +0 -17
- package/dist/types/cli/commands/handlers/index.d.ts +0 -13
- package/dist/types/cli/commands/handlers/index.js.map +0 -1
- package/dist/types/cli/commands/handlers/loadHtml.d.ts +0 -2
- package/dist/types/cli/commands/handlers/loadHtml.js.map +0 -1
|
@@ -1,13 +1,42 @@
|
|
|
1
1
|
// src/loaders/playwright.ts
|
|
2
|
-
import { chromium } from "playwright";
|
|
3
|
-
var
|
|
4
|
-
|
|
5
|
-
|
|
6
|
-
|
|
2
|
+
import { chromium, devices } from "playwright-core";
|
|
3
|
+
var SUPER_BYPASS_DEVICE = {
|
|
4
|
+
userAgent: "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/91.0.4472.124 Safari/537.36",
|
|
5
|
+
viewport: { width: 1920, height: 1080 },
|
|
6
|
+
deviceScaleFactor: 1,
|
|
7
|
+
hasTouch: false,
|
|
8
|
+
isMobile: false,
|
|
9
|
+
javaScriptEnabled: true,
|
|
10
|
+
locale: "en-US",
|
|
11
|
+
timezoneId: "America/New_York"
|
|
12
|
+
};
|
|
13
|
+
var loadHtml = async (url, options) => {
|
|
14
|
+
const { browser, waitUntil, timeout, superBypassMode } = options ?? {};
|
|
15
|
+
const _browser = browser ?? await chromium.launch({ headless: true });
|
|
16
|
+
const context = await _browser.newContext(superBypassMode ? SUPER_BYPASS_DEVICE : devices["Desktop Chrome"]);
|
|
17
|
+
if (superBypassMode) {
|
|
18
|
+
await context.addInitScript(() => {
|
|
19
|
+
Object.defineProperty(navigator, "webdriver", { get: () => void 0 });
|
|
20
|
+
Object.defineProperty(navigator, "languages", { get: () => ["en-US", "en"] });
|
|
21
|
+
Object.defineProperty(navigator, "plugins", { get: () => [1, 2, 3, 4, 5] });
|
|
22
|
+
});
|
|
23
|
+
}
|
|
24
|
+
const page = await context.newPage();
|
|
25
|
+
if (superBypassMode) {
|
|
26
|
+
await page.route("**/*.js", (route) => {
|
|
27
|
+
if (route.request().url().includes("captcha-delivery")) {
|
|
28
|
+
return route.abort();
|
|
29
|
+
}
|
|
30
|
+
return route.continue();
|
|
31
|
+
});
|
|
32
|
+
}
|
|
33
|
+
await page.goto(url, { waitUntil: waitUntil ?? "load", timeout });
|
|
7
34
|
const html = await page.content();
|
|
8
35
|
await page.close();
|
|
9
|
-
if (
|
|
10
|
-
await
|
|
36
|
+
if (browser) {
|
|
37
|
+
await context.close();
|
|
38
|
+
} else {
|
|
39
|
+
await _browser.close();
|
|
11
40
|
}
|
|
12
41
|
return html;
|
|
13
42
|
};
|
|
@@ -10,7 +10,7 @@ var DEFAULT_MDAST_TO_MARKDOWN_OPTIONS = {
|
|
|
10
10
|
};
|
|
11
11
|
var mdastToMarkdown = (mdast, options) => {
|
|
12
12
|
const { baseUrl, ...toMarkdownOptions } = { ...DEFAULT_MDAST_TO_MARKDOWN_OPTIONS, ...options };
|
|
13
|
-
let markdown = toMarkdown(warpRoot(mdast), toMarkdownOptions);
|
|
13
|
+
let markdown = toMarkdown(warpRoot(mdast), toMarkdownOptions).replace(/\*\*\*\*/g, "");
|
|
14
14
|
if (baseUrl) {
|
|
15
15
|
markdown = linkReplacer(markdown, baseUrl);
|
|
16
16
|
}
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"bin.js","sourceRoot":"","sources":["../../../src/cli/bin.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,KAAK,EAAE,MAAM,gBAAgB,CAAC;AACvC,OAAO,EAAE,OAAO,EAAE,MAAM,WAAW,CAAC;AACpC,OAAO,EAAE,MAAM,YAAY,CAAC;AAC5B,OAAO,WAAW,MAAM,oBAAoB,CAAC;AAC7C,OAAO,EAAE,eAAe,EAAE,MAAM,qBAAqB,CAAC;AACtD,OAAO,EAAE,OAAO,EAAE,KAAK,EAAE,MAAM,aAAa,CAAC;AAE7C,OAAO;KACL,IAAI,CAAC,UAAU,CAAC;KAChB,WAAW,CAAC,wDAAwD,CAAC;KACrE,OAAO,CAAC,WAAW,CAAC,OAAO,EAAE,eAAe,EAAE,4BAA4B,CAAC;
|
|
1
|
+
{"version":3,"file":"bin.js","sourceRoot":"","sources":["../../../src/cli/bin.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,KAAK,EAAE,MAAM,gBAAgB,CAAC;AACvC,OAAO,EAAE,OAAO,EAAE,MAAM,WAAW,CAAC;AACpC,OAAO,EAAE,MAAM,YAAY,CAAC;AAC5B,OAAO,WAAW,MAAM,oBAAoB,CAAC;AAC7C,OAAO,EAAE,eAAe,EAAE,MAAM,qBAAqB,CAAC;AACtD,OAAO,EAAE,OAAO,EAAE,KAAK,EAAE,MAAM,aAAa,CAAC;AAE7C,OAAO;KACL,IAAI,CAAC,UAAU,CAAC;KAChB,WAAW,CAAC,wDAAwD,CAAC;KACrE,OAAO,CAAC,WAAW,CAAC,OAAO,EAAE,eAAe,EAAE,4BAA4B,CAAC,CAAC;AAE9E,OAAO;KACL,QAAQ,CAAC,UAAU,EAAE,wBAAwB,CAAC;KAC9C,MAAM,CAAC,uBAAuB,EAAE,kCAAkC,CAAC;KACnE,MAAM,CAAC,mBAAmB,EAAE,oBAAoB,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,GAAG,CAAC;KACpE,MAAM,CAAC,uBAAuB,EAAE,kBAAkB,OAAO,CAAC,IAAI,CAAC,IAAI,CAAC,GAAG,CAAC;KACxE,MAAM,CAAC,aAAa,EAAE,oCAAoC,CAAC;KAC3D,MAAM,CAAC,eAAe,CAAC,CAAC;AAE1B,KAAK,CAAC,EAAE,CAAC,OAAO,CAAC,wBAAwB,WAAW,CAAC,OAAO,EAAE,CAAC,CAAC,CAAC;AAEjE,OAAO,CAAC,KAAK,EAAE,CAAC"}
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"index.js","sourceRoot":"","sources":["../../../../../src/cli/commands/webforai/index.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,MAAM,kBAAkB,CAAC;AAClC,OAAO,IAAI,MAAM,WAAW,CAAC;AAC7B,OAAO,EAAE,KAAK,EAAE,KAAK,EAAE,OAAO,EAAE,MAAM,gBAAgB,CAAC;AACvD,OAAO,EAAE,MAAM,YAAY,CAAC;AAC5B,OAAO,WAAW,MAAM,0BAA0B,CAAC;AACnD,OAAO,EAAE,cAAc,EAAE,MAAM,2BAA2B,CAAC;AAC3D,OAAO,EAAE,eAAe,EAAE,MAAM,+BAA+B,CAAC;AAChE,OAAO,EAAE,eAAe,EAAE,MAAM,+BAA+B,CAAC;AAChE,OAAO,EAAE,iBAAiB,EAAE,MAAM,iCAAiC,CAAC;AACpE,OAAO,EAAE,YAAY,EAAE,MAAM,4BAA4B,CAAC;AAC1D,OAAO,EAAE,KAAK,EAAE,MAAM,aAAa,CAAC;AACpC,OAAO,EAAE,QAAQ,EAAE,MAAM,YAAY,CAAC;AAEtC,MAAM,aAAa,GAAG,EAAE,UAAU,EAAE,IAAI,EAAE,WAAW,EAAE,IAAI,EAAE,SAAS,EAAE,IAAI,EAAE,CAAC;AAC/E,MAAM,sBAAsB,GAAG,EAAE,UAAU,EAAE,KAAK,EAAE,WAAW,EAAE,KAAK,EAAE,SAAS,EAAE,KAAK,EAAE,CAAC;AAE3F,MAAM,CAAC,MAAM,eAAe,GAAG,KAAK,EACnC,WAAmB,EACnB,OAA6E,EAC5E,EAAE;IACH,KAAK,CAAC,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC,KAAK,CAAC,wBAAwB,WAAW,CAAC,OAAO,EAAE,CAAC,CAAC,CAAC,CAAC;IAExE,MAAM,UAAU,GAAG,WAAW,IAAI,CAAC,MAAM,eAAe,EAAE,CAAC,CAAC;IAC5D,OAAO,CAAC,KAAK,IAAI,OAAO,CAAC,KAAK,CAAC,eAAe,UAAU,EAAE,CAAC,CAAC;IAE5D,MAAM,MAAM,GAAG,KAAK,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,MAAM,IAAI,CAAC,MAAM,YAAY,EAAE,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC;IACtF,OAAO,CAAC,KAAK,IAAI,OAAO,CAAC,KAAK,CAAC,WAAW,MAAM,EAAE,CAAC,CAAC;IAEpD,MAAM,UAAU,GAAG,OAAO,CAAC,MAAM,IAAI,CAAC,MAAM,eAAe,CAAC,UAAU,CAAC,CAAC,CAAC;IACzE,OAAO,CAAC,KAAK,IAAI,OAAO,CAAC,KAAK,CAAC,eAAe,UAAU,EAAE,CAAC,CAAC;IAE5D,MAAM,IAAI,GAAG,OAAO,CAAC,IAAI,IAAI,CAAC,MAAM,iBAAiB,EAAE,CAAC,CAAC;IACzD,OAAO,CAAC,KAAK,IAAI,OAAO,CAAC,KAAK,CAAC,SAAS,IAAI,EAAE,CAAC,CAAC;IAEhD,IAAI,IAAY,CAAC;IACjB,MAAM,CAAC,GAAG,OAAO,EAAE,CAAC;IACpB,IAAI,CAAC;QACJ,CAAC,CAAC,KAAK,CAAC,oBAAoB,CAAC,CAAC;QAC9B,IAAI,GAAG,MAAM,QAAQ,CAAC,UAAU,EAAE,MAAM,EAAE,EAAE,KAAK,EAAE,OAAO,CAAC,KAAK,EAAE,CAAC,CAAC;QACpE,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC,KAAK,CAAC,iBAAiB,CAAC,CAAC,CAAC;IACrC,CAAC;IAAC,OAAO,KAAK,EAAE,CAAC;QAChB,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC,GAAG,CAAC,yBAAyB,CAAC,CAAC,CAAC;QAC1C,OAAO,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC;QACrB,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;IACjB,CAAC;IACD,OAAO,CAAC,KAAK,IAAI,OAAO,CAAC,KAAK,CAAC,SAAS,IAAI,EAAE,CAAC,CAAC;IAEhD,MAAM,QAAQ,GAAG,cAAc,CAAC,IAAI,EAAE;QACrC,OAAO,EAAE,KAAK,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC,UAAU,CAAC,CAAC,CAAC,SAAS;QACnD,GAAG,CAAC,IAAI,KAAK,IAAI,CAAC,CAAC,CAAC,aAAa,CAAC,CAAC,CAAC,sBAAsB,CAAC;KAC3D,CAAC,CAAC;IACH,OAAO,CAAC,KAAK,IAAI,OAAO,CAAC,KAAK,CAAC,aAAa,QAAQ,EAAE,CAAC,CAAC;IAExD,MAAM,SAAS,GAAG,IAAI,CAAC,OAAO,CAAC,UAAU,CAAC,CAAC;IAC3C,MAAM,iBAAiB,GAAG,MAAM,EAAE,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,WAAW,EAAE,CAAC,CAAC;IAEtF,IAAI,CAAC,iBAAiB,EAAE,CAAC;QACxB,MAAM,EAAE,CAAC,KAAK,CAAC,SAAS,EAAE,EAAE,SAAS,EAAE,IAAI,EAAE,CAAC,CAAC;IAChD,CAAC;IACD,MAAM,EAAE,CAAC,SAAS,CAAC,UAAU,EAAE,QAAQ,CAAC,CAAC;IAEzC,KAAK,CAAC,EAAE,CAAC,KAAK,CAAC,GAAG,EAAE,CAAC,IAAI,CAAC,OAAO,CAAC,sBAAsB,UAAU,EAAE,CAAC,CAAC,CAAC;AACxE,CAAC,CAAC"}
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"loadHtml.js","sourceRoot":"","sources":["../../../../../src/cli/commands/webforai/loadHtml.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,MAAM,kBAAkB,CAAC;AAClC,OAAO,EAAE,MAAM,YAAY,CAAC;AAC5B,OAAO,EAAE,QAAQ,EAAE,MAAM,iBAAiB,CAAC;AAC3C,OAAO,qBAAqB,MAAM,8BAA8B,CAAC;AACjE,OAAO,EAAE,QAAQ,IAAI,eAAe,EAAE,MAAM,wBAAwB,CAAC;AACrE,OAAO,EAAE,QAAQ,IAAI,oBAAoB,EAAE,MAAM,6BAA6B,CAAC;AAE/E,MAAM,wBAAwB,GAAG,KAAK,IAAI,EAAE;IAC3C,MAAM,IAAI,GAAG,QAAQ,CAAC,cAAc,EAAE,CAAC;IACvC,MAAM,WAAW,GAAG,MAAM,EAAE,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC,KAAK,CAAC,GAAG,EAAE,CAAC,KAAK,CAAC,CAAC;IAC7D,OAAO,CAAC,CAAC,WAAW,CAAC;AACtB,CAAC,CAAC;AAEF,MAAM,CAAC,MAAM,QAAQ,GAAG,KAAK,EAAE,UAAkB,EAAE,MAAc,EAAE,OAA4B,EAAE,EAAE;IAClG,IAAI,MAAM,KAAK,OAAO,EAAE,CAAC;QACxB,OAAO,CAAC,KAAK,IAAI,OAAO,CAAC,KAAK,CAAC,iCAAiC,UAAU,EAAE,CAAC,CAAC;QAC9E,MAAM,OAAO,GAAG,MAAM,EAAE,CAAC,QAAQ,CAAC,UAAU,EAAE,OAAO,CAAC,CAAC;QACvD,OAAO,CAAC,KAAK,IAAI,OAAO,CAAC,KAAK,CAAC,gBAAgB,OAAO,CAAC,KAAK,CAAC,CAAC,EAAE,GAAG,CAAC,EAAE,CAAC,CAAC;QACxE,OAAO,OAAO,CAAC;IAChB,CAAC;IAED,IAAI,MAAM,KAAK,OAAO,EAAE,CAAC;QACxB,OAAO,CAAC,KAAK,IAAI,OAAO,CAAC,KAAK,CAAC,0BAA0B,UAAU,EAAE,CAAC,CAAC;QACvE,MAAM,OAAO,GAAG,MAAM,eAAe,CAAC,UAAU,CAAC,CAAC;QAClD,OAAO,CAAC,KAAK,IAAI,OAAO,CAAC,KAAK,CAAC,gBAAgB,OAAO,CAAC,KAAK,CAAC,CAAC,EAAE,GAAG,CAAC,EAAE,CAAC,CAAC;QACxE,OAAO,OAAO,CAAC;IAChB,CAAC;IAED,IAAI,MAAM,KAAK,YAAY,EAAE,CAAC;QAC7B,OAAO,CAAC,KAAK,IAAI,OAAO,CAAC,KAAK,CAAC,iCAAiC,UAAU,EAAE,CAAC,CAAC;QAC9E,MAAM,qBAAqB,GAAG,MAAM,wBAAwB,EAAE,CAAC;QAC/D,OAAO,CAAC,KAAK,IAAI,OAAO,CAAC,KAAK,CAAC,yBAAyB,qBAAqB,EAAE,CAAC,CAAC;QACjF,IAAI,CAAC,qBAAqB,EAAE,CAAC;YAC5B,MAAM,OAAO,GAAG;gBACf,2DAA2D;gBAC3D,gCAAgC;gBAChC,EAAE;gBACF,IAAI,EAAE,CAAC,IAAI,CAAC,kBAAkB,qBAAqB,CAAC,OAAO,mBAAmB,CAAC,GAAG;gBAClF,EAAE;gBACF,sCAAsC;gBACtC,wHAAwH;gBACxH,GAAG,EAAE,CAAC,IAAI,CAAC,sBAAsB,CAAC,EAAE;gBACpC,EAAE;gBACF,sCAAsC;gBACtC,sDAAsD;gBACtD,6CAA6C;gBAC7C,EAAE;gBACF,IAAI,EAAE,CAAC,IAAI,CAAC,uBAAuB,qBAAqB,CAAC,OAAO,eAAe,CAAC,GAAG;gBACnF,EAAE;gBACF,EAAE,CAAC,IAAI,CAAC,4BAA4B,CAAC;aACrC,CAAC;YACF,OAAO,CAAC,IAAI,CAAC,OAAO,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC;YACjC,MAAM,IAAI,KAAK,CAAC,6BAA6B,CAAC,CAAC;QAChD,CAAC;QACD,MAAM,OAAO,GAAG,MAAM,oBAAoB,CAAC,UAAU,CAAC,CAAC;QACvD,OAAO,CAAC,KAAK,IAAI,OAAO,CAAC,KAAK,CAAC,gBAAgB,OAAO,CAAC,KAAK,CAAC,CAAC,EAAE,GAAG,CAAC,EAAE,CAAC,CAAC;QACxE,OAAO,OAAO,CAAC;IAChB,CAAC;IAED,MAAM,IAAI,KAAK,CAAC,uBAAuB,MAAM,EAAE,CAAC,CAAC;AAClD,CAAC,CAAC"}
|
|
@@ -1,5 +1,32 @@
|
|
|
1
|
+
/**
|
|
2
|
+
* Determines whether a string is a valid URL.
|
|
3
|
+
* @param maybeUrl - The string to check.
|
|
4
|
+
* @returns `true` if the string is a valid URL, otherwise `false`.
|
|
5
|
+
*/
|
|
1
6
|
export declare const isUrl: (maybeUrl: string) => boolean;
|
|
7
|
+
/**
|
|
8
|
+
* Changes the file extension of a given file path.
|
|
9
|
+
* @param filePath - The original file path.
|
|
10
|
+
* @param newExtension - The new extension (with or without the leading dot).
|
|
11
|
+
* @returns The file path with the updated extension.
|
|
12
|
+
*/
|
|
2
13
|
export declare function changeFileExtension(filePath: string, newExtension: string): string;
|
|
14
|
+
/**
|
|
15
|
+
* Converts a URL into a safe filename.
|
|
16
|
+
* @param url - The URL to convert.
|
|
17
|
+
* @returns The generated filename, or "output" if conversion fails.
|
|
18
|
+
*/
|
|
3
19
|
export declare function urlToFilename(url: string): string;
|
|
20
|
+
/**
|
|
21
|
+
* Converts a source path to an output path. If the source is a URL, it generates a filename and changes the extension to .md.
|
|
22
|
+
* If the source is a file path, it simply changes the extension to .md.
|
|
23
|
+
* @param sourcePath - The source path or URL.
|
|
24
|
+
* @returns The corresponding output path.
|
|
25
|
+
*/
|
|
4
26
|
export declare const sourcePathToOutputPath: (sourcePath: string) => string;
|
|
27
|
+
/**
|
|
28
|
+
* Retrieves the next available file path by appending a counter if the file already exists.
|
|
29
|
+
* @param filePath - The original file path.
|
|
30
|
+
* @returns A file path that does not currently exist.
|
|
31
|
+
*/
|
|
5
32
|
export declare function getNextAvailableFilePath(filePath: string): string;
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"utils.js","sourceRoot":"","sources":["../../../src/cli/utils.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,MAAM,SAAS,CAAC;AACzB,OAAO,IAAI,MAAM,WAAW,CAAC;AAE7B,MAAM,CAAC,MAAM,KAAK,GAAG,CAAC,QAAgB,
|
|
1
|
+
{"version":3,"file":"utils.js","sourceRoot":"","sources":["../../../src/cli/utils.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,MAAM,SAAS,CAAC;AACzB,OAAO,IAAI,MAAM,WAAW,CAAC;AAE7B;;;;GAIG;AACH,MAAM,CAAC,MAAM,KAAK,GAAG,CAAC,QAAgB,EAAW,EAAE;IAClD,IAAI,CAAC;QACJ,IAAI,GAAG,CAAC,QAAQ,CAAC,CAAC;QAClB,OAAO,IAAI,CAAC;IACb,CAAC;IAAC,MAAM,CAAC;QACR,OAAO,KAAK,CAAC;IACd,CAAC;AACF,CAAC,CAAC;AAEF;;;;;GAKG;AACH,MAAM,UAAU,mBAAmB,CAAC,QAAgB,EAAE,YAAoB;IACzE,MAAM,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,QAAQ,CAAC,CAAC;IACpC,MAAM,GAAG,GAAG,YAAY,CAAC,UAAU,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,YAAY,CAAC,CAAC,CAAC,IAAI,YAAY,EAAE,CAAC;IAC7E,OAAO,IAAI,CAAC,MAAM,CAAC,EAAE,GAAG,MAAM,EAAE,GAAG,EAAE,CAAC,CAAC;AACxC,CAAC;AAED;;;;GAIG;AACH,MAAM,UAAU,aAAa,CAAC,GAAW;IACxC,IAAI,CAAC;QACJ,MAAM,MAAM,GAAG,IAAI,GAAG,CAAC,GAAG,CAAC,CAAC;QAE5B,oEAAoE;QACpE,MAAM,QAAQ,GAAG,MAAM,CAAC,QAAQ,CAAC,UAAU,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,QAAQ,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,QAAQ,CAAC;QACjG,MAAM,WAAW,GAAG,QAAQ,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;QAE5D,qDAAqD;QACrD,MAAM,SAAS,GAAG,MAAM,CAAC,QAAQ,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC;QAC7D,MAAM,iBAAiB,GAAG,SAAS,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,kBAAkB,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;QAEhF,gCAAgC;QAChC,IAAI,QAAQ,GAAG,CAAC,WAAW,EAAE,iBAAiB,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;QAE1E,wBAAwB;QACxB,QAAQ,GAAG,QAAQ;aACjB,WAAW,EAAE;YACd,2DAA2D;aAC1D,OAAO,CAAC,wBAAwB,EAAE,EAAE,CAAC,CAAC,4BAA4B;aAClE,OAAO,CAAC,SAAS,EAAE,GAAG,CAAC,CAAC,uCAAuC;aAC/D,OAAO,CAAC,UAAU,EAAE,EAAE,CAAC,CAAC,CAAC,sCAAsC;QAEjE,OAAO,QAAQ,IAAI,QAAQ,CAAC;IAC7B,CAAC;IAAC,MAAM,CAAC;QACR,OAAO,QAAQ,CAAC;IACjB,CAAC;AACF,CAAC;AAED;;;;;GAKG;AACH,MAAM,CAAC,MAAM,sBAAsB,GAAG,CAAC,UAAkB,EAAU,EAAE,CACpE,KAAK,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC,GAAG,aAAa,CAAC,UAAU,CAAC,KAAK,CAAC,CAAC,CAAC,mBAAmB,CAAC,UAAU,EAAE,IAAI,CAAC,CAAC;AAE/F;;;;GAIG;AACH,MAAM,UAAU,wBAAwB,CAAC,QAAgB;IACxD,MAAM,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,QAAQ,CAAC,CAAC;IACpC,MAAM,GAAG,GAAG,MAAM,CAAC,GAAG,CAAC;IACvB,MAAM,GAAG,GAAG,MAAM,CAAC,GAAG,CAAC;IACvB,MAAM,QAAQ,GAAG,MAAM,CAAC,IAAI,CAAC,OAAO,CAAC,OAAO,EAAE,EAAE,CAAC,CAAC;IAClD,MAAM,KAAK,GAAG,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,SAAS,CAAC,CAAC;IAC3C,IAAI,OAAO,GAAG,KAAK,CAAC,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;IAE/C,IAAI,OAAO,GAAG,IAAI,CAAC,IAAI,CAAC,GAAG,EAAE,GAAG,QAAQ,IAAI,OAAO,GAAG,GAAG,EAAE,CAAC,CAAC;IAE7D,OAAO,EAAE,CAAC,UAAU,CAAC,OAAO,CAAC,EAAE,CAAC;QAC/B,OAAO,EAAE,CAAC;QACV,OAAO,GAAG,IAAI,CAAC,IAAI,CAAC,GAAG,EAAE,GAAG,QAAQ,IAAI,OAAO,GAAG,GAAG,EAAE,CAAC,CAAC;IAC1D,CAAC;IAED,OAAO,OAAO,CAAC;AAChB,CAAC"}
|
|
@@ -1,10 +1,15 @@
|
|
|
1
1
|
import type { Nodes as Hast } from "hast";
|
|
2
2
|
export type ExtractPresets = "strongly" | "readability";
|
|
3
|
-
export type
|
|
3
|
+
export type ExtractParams = {
|
|
4
|
+
hast: Hast;
|
|
5
|
+
lang?: string;
|
|
6
|
+
url?: string;
|
|
7
|
+
};
|
|
8
|
+
export type Extractor = ((param: ExtractParams) => Hast) | false | ExtractPresets;
|
|
4
9
|
export type Extracotrs = Extractor | Extractor[];
|
|
5
10
|
export declare const PRESET_EXTRACT_HAST: {
|
|
6
|
-
strongly: (
|
|
7
|
-
readability: (
|
|
11
|
+
strongly: (params: ExtractParams) => Hast;
|
|
12
|
+
readability: (params: ExtractParams) => Hast;
|
|
8
13
|
};
|
|
9
14
|
export declare const DEFAULT_EXTRACT_HAST: Extracotrs;
|
|
10
|
-
export declare const extractHast: (
|
|
15
|
+
export declare const extractHast: (params: ExtractParams, extractors?: Extracotrs) => Hast;
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"index.js","sourceRoot":"","sources":["../../../src/extract-hast/index.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,sBAAsB,EAAE,MAAM,eAAe,CAAC;AACvD,OAAO,EAAE,mBAAmB,EAAE,MAAM,YAAY,CAAC;
|
|
1
|
+
{"version":3,"file":"index.js","sourceRoot":"","sources":["../../../src/extract-hast/index.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,sBAAsB,EAAE,MAAM,eAAe,CAAC;AACvD,OAAO,EAAE,mBAAmB,EAAE,MAAM,YAAY,CAAC;AAOjD,MAAM,CAAC,MAAM,mBAAmB,GAAG;IAClC,QAAQ,EAAE,mBAAmB;IAC7B,WAAW,EAAE,sBAAsB;CACnC,CAAC;AAEF,MAAM,CAAC,MAAM,oBAAoB,GAAe,CAAC,aAAa,CAAC,CAAC;AAEhE,MAAM,CAAC,MAAM,WAAW,GAAG,CAAC,MAAqB,EAAE,aAAyB,oBAAoB,EAAQ,EAAE;IACzG,MAAM,EAAE,IAAI,EAAE,IAAI,EAAE,GAAG,MAAM,CAAC;IAC9B,MAAM,WAAW,GAAG,KAAK,CAAC,OAAO,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC,UAAU,CAAC,CAAC;IAE1E,MAAM,SAAS,GACd,WAAW,CAAC,MAAM,CAAO,CAAC,GAAG,EAAE,SAAS,EAAE,EAAE;QAC3C,IAAI,SAAS,KAAK,KAAK,EAAE,CAAC;YACzB,OAAO,GAAG,CAAC;QACZ,CAAC;QACD,IAAI,OAAO,SAAS,KAAK,QAAQ,IAAI,SAAS,IAAI,mBAAmB,EAAE,CAAC;YACvE,OAAO,mBAAmB,CAAC,SAAS,CAAC,CAAC,EAAE,IAAI,EAAE,GAAG,EAAE,IAAI,EAAE,CAAC,CAAC;QAC5D,CAAC;QACD,IAAI,OAAO,SAAS,KAAK,UAAU,EAAE,CAAC;YACrC,OAAO,SAAS,CAAC,EAAE,IAAI,EAAE,GAAG,EAAE,IAAI,EAAE,CAAC,CAAC;QACvC,CAAC;QACD,MAAM,IAAI,KAAK,CAAC,sBAAsB,SAAS,EAAE,CAAC,CAAC;IACpD,CAAC,EAAE,IAAI,CAAC,IAAI,IAAI,CAAC;IAElB,OAAO,SAAS,CAAC;AAClB,CAAC,CAAC"}
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"readability.js","sourceRoot":"","sources":["../../../src/extract-hast/readability.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,MAAM,EAAE,SAAS,EAAE,MAAM,kBAAkB,CAAC;AACrD,OAAO,EAAE,QAAQ,IAAI,YAAY,EAAE,MAAM,qBAAqB,CAAC;AAC/D,OAAO,EAAE,MAAM,EAAE,MAAM,mBAAmB,CAAC;
|
|
1
|
+
{"version":3,"file":"readability.js","sourceRoot":"","sources":["../../../src/extract-hast/readability.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,MAAM,EAAE,SAAS,EAAE,MAAM,kBAAkB,CAAC;AACrD,OAAO,EAAE,QAAQ,IAAI,YAAY,EAAE,MAAM,qBAAqB,CAAC;AAC/D,OAAO,EAAE,MAAM,EAAE,MAAM,mBAAmB,CAAC;AAE3C,OAAO,EAAE,UAAU,EAAE,YAAY,EAAE,WAAW,EAAE,MAAM,SAAS,CAAC;AAEhE,MAAM,cAAc,GAAG,CAAC,MAAM,EAAE,SAAS,EAAE,eAAe,EAAE,YAAY,EAAE,OAAO,EAAE,aAAa,EAAE,QAAQ,CAAC,CAAC;AAE5G,MAAM,OAAO,GAAG;IACf,MAAM,EAAE,kCAAkC;IAC1C,MAAM,EAAE,4CAA4C;IACpD,yBAAyB,EAAE,8BAA8B;IACzD,kBAAkB,EACjB,wRAAwR;IACzR,oBAAoB,EAAE,mDAAmD;CACzE,CAAC;AAEF,MAAM,cAAc,GAAG,CAAC,SAAS,EAAE,UAAU,EAAE,eAAe,EAAE,eAAe,EAAE,UAAU,EAAE,QAAQ,CAAC,CAAC;AAEvG,MAAM,cAAc,GAAG,CAAC,GAAG,EAAE,GAAG,EAAE,KAAK,EAAE,SAAS,EAAE,SAAS,EAAE,MAAM,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,CAAC,CAAC;AAEzF,MAAM,mBAAmB,GAAG;IAC3B,EAAE,EAAE,GAAG;IACP,EAAE,EAAE,GAAG;CACP,CAAC;AAEF,yBAAyB;AACzB,UAAU;AACV,WAAW;AACX,KAAK;AAEL,MAAM,cAAc,GAAG,CAAC,IAAU,EAAE,EAAE;IACrC,OAAO,CAAC,CACP,CAAC,SAAS,EAAE,SAAS,CAAC,CAAC,QAAQ,CAAC,IAAI,CAAC,IAAI,CAAC;QAC1C,CAAC,IAAI,CAAC,IAAI,KAAK,SAAS,IAAI,CAAC,QAAQ,EAAE,OAAO,EAAE,MAAM,EAAE,MAAM,EAAE,UAAU,EAAE,KAAK,EAAE,OAAO,CAAC,CAAC,QAAQ,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,CACnH,CAAC;AACH,CAAC,CAAC;AAEF,MAAM,sBAAsB,GAAG,CAAC,IAAU,EAAE,EAAE;IAC7C,IAAI,IAAI,CAAC,IAAI,KAAK,SAAS,EAAE,CAAC;QAC7B,OAAO,IAAI,CAAC;IACb,CAAC;IACD,MAAM,OAAO,GAAG,IAAe,CAAC;IAEhC,IAAI,CAAC,OAAO,EAAE,KAAK,CAAC,CAAC,QAAQ,CAAC,OAAO,CAAC,OAAO,CAAC,EAAE,CAAC;QAChD,OAAO,KAAK,CAAC;IACd,CAAC;IAED,yCAAyC;IACzC,IAAI,CAAC,QAAQ,EAAE,aAAa,CAAC,CAAC,IAAI,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,OAAO,CAAC,UAAU,CAAC,GAAG,CAAC,CAAC,EAAE,CAAC;QACtE,OAAO,KAAK,CAAC;IACd,CAAC;IACD,IAAI,UAAU,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,CAAC,SAAS,EAAE,EAAE,CAAC,OAAO,CAAC,MAAM,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC,EAAE,CAAC;QAC7E,OAAO,KAAK,CAAC;IACd,CAAC;IAED,yBAAyB;IACzB,IAAI,OAAO,CAAC,OAAO,KAAK,QAAQ,EAAE,CAAC;QAClC,OAAO,KAAK,CAAC;IACd,CAAC;IACD,IAAI,OAAO,CAAC,UAAU,CAAC,IAAI,KAAK,QAAQ,IAAI,OAAO,CAAC,UAAU,CAAC,YAAY,CAAC,EAAE,CAAC;QAC9E,OAAO,KAAK,CAAC;IACd,CAAC;IAED,yBAAyB;IACzB,IAAI,OAAO,CAAC,UAAU,CAAC,GAAG,KAAK,QAAQ,IAAI,YAAY,CAAC,OAAO,CAAC,UAAU,CAAC,QAAQ,EAAE,QAAQ,CAAC,EAAE,CAAC;QAChG,OAAO,KAAK,CAAC;IACd,CAAC;IACD,IAAI,OAAO,CAAC,MAAM,CAAC,IAAI,CAAC,WAAW,CAAC,OAAO,CAAC,CAAC,EAAE,CAAC;QAC/C,OAAO,KAAK,CAAC;IACd,CAAC;IAED,wBAAwB;IACxB,IAAI,OAAO,CAAC,UAAU,CAAC,IAAI,IAAI,cAAc,CAAC,QAAQ,CAAC,OAAO,CAAC,UAAU,CAAC,IAAc,CAAC,EAAE,CAAC;QAC3F,OAAO,KAAK,CAAC;IACd,CAAC;IAED,OAAO,IAAI,CAAC;AACb,CAAC,CAAC;AAEF,MAAM,qBAAqB,GAAG,CAAC,IAAU,EAAE,EAAE;IAC5C,IAAI,IAAI,CAAC,IAAI,KAAK,SAAS,EAAE,CAAC;QAC7B,OAAO,IAAI,CAAC;IACb,CAAC;IACD,MAAM,OAAO,GAAG,IAAe,CAAC;IAEhC,6BAA6B;IAC7B,IAAI,CAAC,MAAM,EAAE,SAAS,EAAE,MAAM,EAAE,SAAS,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,GAAG,CAAC,CAAC,QAAQ,CAAC,OAAO,CAAC,OAAO,CAAC,EAAE,CAAC;QAC/G,OAAO,IAAI,CAAC;IACb,CAAC;IACD,MAAM,KAAK,GAAG,WAAW,CAAC,OAAO,CAAC,CAAC;IAEnC,IAAI,OAAO,CAAC,yBAAyB,CAAC,IAAI,CAAC,KAAK,CAAC,EAAE,CAAC;QACnD,OAAO,KAAK,CAAC;IACd,CAAC;IAED,6BAA6B;IAC7B,IAAI,OAAO,CAAC,kBAAkB,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,OAAO,CAAC,oBAAoB,CAAC,IAAI,CAAC,KAAK,CAAC,EAAE,CAAC;QACzF,OAAO,KAAK,CAAC;IACd,CAAC;IAED,OAAO,IAAI,CAAC;AACb,CAAC,CAAC;AAEF,MAAM,iBAAiB,GAAG,CAAC,IAAU,EAAE,KAAa,EAAE,EAAE;IACvD,IAAI,IAAI,CAAC,IAAI,KAAK,SAAS,EAAE,CAAC;QAC7B,OAAO,IAAI,CAAC;IACb,CAAC;IACD,MAAM,OAAO,GAAG,IAAe,CAAC;IAEhC,IAAI,cAAc,CAAC,QAAQ,CAAC,OAAO,CAAC,OAAO,CAAC,EAAE,CAAC;QAC9C,OAAO,IAAI,CAAC;IACb,CAAC;IAED,IAAI,OAAO,CAAC,OAAO,KAAK,KAAK,IAAI,CAAC,OAAO,CAAC,UAAU,CAAC,GAAG,EAAE,CAAC;QAC1D,OAAO,KAAK,CAAC;IACd,CAAC;IAED,OAAO,IAAI,CAAC;AACb,CAAC,CAAC;AAEF,MAAM,CAAC,MAAM,sBAAsB,GAAG,CAAC,MAAqB,EAAQ,EAAE;IACrE,MAAM,EAAE,IAAI,EAAE,IAAI,GAAG,IAAI,EAAE,GAAG,MAAM,CAAC;IACrC,MAAM,IAAI,GAAG,MAAM,CAAC,MAAM,EAAE,IAAI,CAAC,IAAI,IAAI,CAAC;IAE1C,MAAM,oBAAoB,GAAG,MAAM,CAAC,IAAI,EAAE,CAAC,IAAI,EAAE,EAAE,CAAC,cAAc,CAAC,IAAY,CAAC,CAAC,CAAC;IAClF,MAAM,wBAAwB,GAAG,oBAAoB,IAAI,YAAY,CAAC,oBAAoB,CAAC,CAAC;IAC5F,IAAI,CAAC,CAAC,oBAAoB,IAAI,wBAAwB,CAAC,EAAE,CAAC;QACzD,OAAO,IAAI,CAAC;IACb,CAAC;IAED,MAAM,WAAW,GAAG,MAAM,CAAC,oBAAoB,EAAE,CAAC,IAAI,EAAE,EAAE,CAAC,sBAAsB,CAAC,IAAY,CAAC,CAAC,CAAC;IACjG,MAAM,eAAe,GAAG,WAAW,CAAC,CAAC,CAAC,YAAY,CAAC,WAAW,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;IACrE,MAAM,CAAC,QAAQ,EAAE,QAAQ,CAAC,GACzB,eAAe,CAAC,MAAM,GAAG,wBAAwB,CAAC,MAAM,GAAG,CAAC,IAAI,eAAe,CAAC,MAAM,GAAG,IAAI;QAC5F,CAAC,CAAE,CAAC,WAAmB,EAAE,eAAe,CAAW;QACnD,CAAC,CAAE,CAAC,oBAA4B,EAAE,wBAAwB,CAAW,CAAC;IAExE,IAAI,aAAa,GAAG,IAAI,IAAI,mBAAmB,CAAC,CAAC,CAAC,mBAAmB,CAAC,IAAwC,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC;IACtH,IAAI,QAAQ,CAAC,MAAM,GAAG,aAAa,EAAE,CAAC;QACrC,aAAa,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,QAAQ,CAAC,MAAM,GAAG,GAAG,CAAC,CAAC;IACpD,CAAC;IAED,IAAI,aAAa,GAAS,QAAQ,CAAC;IACnC,IAAI,aAAa,GAAG,QAAQ,CAAC;IAE7B,KAAK,MAAM,QAAQ,IAAI,cAAc,EAAE,CAAC;QACvC,MAAM,OAAO,GAAG,EAAE,IAAI,EAAE,MAAe,EAAE,QAAQ,EAAE,SAAS,CAAC,QAAQ,EAAE,WAAW,CAAC,EAAE,CAAC;QACtF,MAAM,WAAW,GAAG,YAAY,CAAC,OAAO,CAAC,CAAC;QAE1C,IAAI,WAAW,CAAC,MAAM,GAAG,EAAE,EAAE,CAAC;YAC7B,SAAS;QACV,CAAC;QAED,MAAM,KAAK,GAAG,SAAS,CAAC,GAAG,EAAE,OAAO,CAAC,CAAC;QACtC,MAAM,QAAQ,GAAG,KAAK,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,YAAY,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;QAElE,MAAM,WAAW,GAAG,QAAQ,CAAC,MAAM,GAAG,WAAW,CAAC,MAAM,CAAC;QACzD,IAAI,WAAW,GAAG,GAAG,EAAE,CAAC;YACvB,SAAS;QACV,CAAC;QAED,IAAI,WAAW,CAAC,MAAM,GAAG,aAAa,EAAE,CAAC;YACxC,aAAa,GAAG,OAAO,CAAC;YACxB,aAAa,GAAG,WAAW,CAAC;YAC5B,MAAM;QACP,CAAC;IACF,CAAC;IAED,MAAM,iBAAiB,GAAG,MAAM,CAAC,aAAa,EAAE,CAAC,IAAI,EAAE,EAAE;QACxD,IAAI,CAAC,iBAAiB,CAAC,IAAY,EAAE,IAAI,CAAC,EAAE,CAAC;YAC5C,OAAO,KAAK,CAAC;QACd,CAAC;QACD,IAAI,CAAC,qBAAqB,CAAC,IAAY,CAAC,EAAE,CAAC;YAC1C,OAAO,KAAK,CAAC;QACd,CAAC;QAED,OAAO,IAAI,CAAC;IACb,CAAC,CAAS,CAAC;IAEX,MAAM,SAAS,GACd,YAAY,CAAC,iBAAiB,CAAC,CAAC,MAAM,GAAG,aAAa,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,iBAAiB,CAAC,CAAC,CAAC,aAAa,CAAC;IACvG,OAAO,SAAS,CAAC;AAClB,CAAC,CAAC"}
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"strongly.js","sourceRoot":"","sources":["../../../src/extract-hast/strongly.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,MAAM,EAAE,MAAM,kBAAkB,CAAC;AAC1C,OAAO,EAAE,MAAM,EAAE,MAAM,mBAAmB,CAAC;
|
|
1
|
+
{"version":3,"file":"strongly.js","sourceRoot":"","sources":["../../../src/extract-hast/strongly.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,MAAM,EAAE,MAAM,kBAAkB,CAAC;AAC1C,OAAO,EAAE,MAAM,EAAE,MAAM,mBAAmB,CAAC;AAG3C,MAAM,YAAY,GACjB,iPAAiP,CAAC;AAEnP,MAAM,CAAC,MAAM,mBAAmB,GAAG,CAAC,MAAqB,EAAQ,EAAE;IAClE,MAAM,EAAE,IAAI,EAAE,GAAG,MAAM,CAAC;IACxB,MAAM,SAAS,GAAG,CAAC,SAAS,EAAE,UAAU,EAAE,UAAU,EAAE,MAAM,EAAE,eAAe,CAAC,CAAC;IAC/E,IAAI,eAAe,GAAG,IAAI,CAAC;IAC3B,KAAK,MAAM,QAAQ,IAAI,SAAS,EAAE,CAAC;QAClC,MAAM,OAAO,GAAG,MAAM,CAAC,QAAQ,EAAE,IAAI,CAAC,CAAC;QACvC,IAAI,OAAO,EAAE,CAAC;YACb,eAAe,GAAG,OAAO,CAAC;YAC1B,MAAM;QACP,CAAC;IACF,CAAC;IAED,MAAM,eAAe,GAAG,MAAM,CAAC,eAAe,EAAE,CAAC,IAAI,EAAE,EAAE;QACxD,IAAI,IAAI,CAAC,IAAI,KAAK,SAAS,EAAE,CAAC;YAC7B,OAAO,KAAK,CAAC;QACd,CAAC;QACD,IAAI,IAAI,CAAC,IAAI,KAAK,SAAS,EAAE,CAAC;YAC7B,OAAO,IAAI,CAAC;QACb,CAAC;QACD,MAAM,OAAO,GAAG,IAAe,CAAC;QAChC,MAAM,WAAW,GAAG,GAAG,OAAO,CAAC,OAAO,IAAI,OAAO,CAAC,UAAU,CAAC,EAAE,IAAI,OAAO,CAAC,UAAU,CAAC,SAAS,EAAE,CAAC;QAClG,IAAI,YAAY,CAAC,IAAI,CAAC,WAAW,CAAC,EAAE,CAAC;YACpC,OAAO,KAAK,CAAC;QACd,CAAC;QACD,OAAO,IAAI,CAAC;IACb,CAAC,CAAC,CAAC;IAEH,OAAO,eAAuB,CAAC;AAChC,CAAC,CAAC"}
|
|
@@ -2,4 +2,3 @@ import type { Element } from "hast";
|
|
|
2
2
|
export declare const matchString: (element: Element) => string;
|
|
3
3
|
export declare const classnames: (element: Element) => string[];
|
|
4
4
|
export declare const isStrInclude: (value: unknown, match: string) => boolean;
|
|
5
|
-
export declare const hasAncestors: (element: Element, tagNames: string[], depth?: number) => boolean;
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"utils.js","sourceRoot":"","sources":["../../../src/extract-hast/utils.ts"],"names":[],"mappings":"AAEA,MAAM,CAAC,MAAM,WAAW,GAAG,CAAC,OAAgB,EAAE,EAAE,CAC/C,GAAG,OAAO,CAAC,OAAO,IAAI,OAAO,CAAC,UAAU,CAAC,EAAE,IAAI,UAAU,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC;AAEhF,MAAM,CAAC,MAAM,UAAU,GAAG,CAAC,OAAgB,EAAE,EAAE;IAC9C,IAAI,KAAK,CAAC,OAAO,CAAC,OAAO,CAAC,UAAU,CAAC,SAAS,CAAC,EAAE,CAAC;QACjD,OAAO,OAAO,CAAC,UAAU,CAAC,SAAqB,CAAC;IACjD,CAAC;IACD,OAAO,EAAE,CAAC;AACX,CAAC,CAAC;AAEF,MAAM,CAAC,MAAM,YAAY,GAAG,CAAC,KAAc,EAAE,KAAa,EAAE,EAAE;IAC7D,IAAI,OAAO,KAAK,KAAK,QAAQ,EAAE,CAAC;QAC/B,OAAO,KAAK,CAAC,QAAQ,CAAC,KAAK,CAAC,CAAC;IAC9B,CAAC;IACD,OAAO,KAAK,CAAC;AACd,CAAC,CAAC
|
|
1
|
+
{"version":3,"file":"utils.js","sourceRoot":"","sources":["../../../src/extract-hast/utils.ts"],"names":[],"mappings":"AAEA,MAAM,CAAC,MAAM,WAAW,GAAG,CAAC,OAAgB,EAAE,EAAE,CAC/C,GAAG,OAAO,CAAC,OAAO,IAAI,OAAO,CAAC,UAAU,CAAC,EAAE,IAAI,UAAU,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC;AAEhF,MAAM,CAAC,MAAM,UAAU,GAAG,CAAC,OAAgB,EAAE,EAAE;IAC9C,IAAI,KAAK,CAAC,OAAO,CAAC,OAAO,CAAC,UAAU,CAAC,SAAS,CAAC,EAAE,CAAC;QACjD,OAAO,OAAO,CAAC,UAAU,CAAC,SAAqB,CAAC;IACjD,CAAC;IACD,OAAO,EAAE,CAAC;AACX,CAAC,CAAC;AAEF,MAAM,CAAC,MAAM,YAAY,GAAG,CAAC,KAAc,EAAE,KAAa,EAAE,EAAE;IAC7D,IAAI,OAAO,KAAK,KAAK,QAAQ,EAAE,CAAC;QAC/B,OAAO,KAAK,CAAC,QAAQ,CAAC,KAAK,CAAC,CAAC;IAC9B,CAAC;IACD,OAAO,KAAK,CAAC;AACd,CAAC,CAAC"}
|
|
@@ -13,6 +13,10 @@ export type HtmlToMdastOptions = {
|
|
|
13
13
|
tableAsText?: boolean;
|
|
14
14
|
/** Whether to hide images. */
|
|
15
15
|
hideImage?: boolean;
|
|
16
|
+
/** The language of the HTML. */
|
|
17
|
+
lang?: string;
|
|
18
|
+
/** The URL of the HTML. */
|
|
19
|
+
url?: string;
|
|
16
20
|
};
|
|
17
21
|
/**
|
|
18
22
|
* Converts an HTML string or HAST tree to an MDAST tree.
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"html-to-mdast.js","sourceRoot":"","sources":["../../src/html-to-mdast.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,QAAQ,EAAE,MAAM,qBAAqB,CAAC;AAC/C,OAAO,EAAE,OAAO,EAAE,MAAM,oBAAoB,CAAC;AAG7C,OAAO,EAAmB,WAAW,EAAE,MAAM,gBAAgB,CAAC;AAC9D,OAAO,EAAE,YAAY,EAAE,MAAM,iBAAiB,CAAC;AAC/C,OAAO,EAAE,cAAc,EAAE,MAAM,mCAAmC,CAAC;AACnE,OAAO,EAAE,iBAAiB,EAAE,MAAM,sCAAsC,CAAC;AACzE,OAAO,EAAE,gBAAgB,EAAE,MAAM,qCAAqC,CAAC;AACvE,OAAO,EAAE,gBAAgB,EAAE,MAAM,qCAAqC,CAAC;AACvE,OAAO,EAAE,kBAAkB,EAAE,MAAM,uCAAuC,CAAC;AAC3E,OAAO,EAAE,
|
|
1
|
+
{"version":3,"file":"html-to-mdast.js","sourceRoot":"","sources":["../../src/html-to-mdast.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,QAAQ,EAAE,MAAM,qBAAqB,CAAC;AAC/C,OAAO,EAAE,OAAO,EAAE,MAAM,oBAAoB,CAAC;AAG7C,OAAO,EAAmB,WAAW,EAAE,MAAM,gBAAgB,CAAC;AAC9D,OAAO,EAAE,YAAY,EAAE,MAAM,iBAAiB,CAAC;AAC/C,OAAO,EAAE,cAAc,EAAE,MAAM,mCAAmC,CAAC;AACnE,OAAO,EAAE,iBAAiB,EAAE,MAAM,sCAAsC,CAAC;AACzE,OAAO,EAAE,gBAAgB,EAAE,MAAM,qCAAqC,CAAC;AACvE,OAAO,EAAE,gBAAgB,EAAE,MAAM,qCAAqC,CAAC;AACvE,OAAO,EAAE,kBAAkB,EAAE,MAAM,uCAAuC,CAAC;AAC3E,OAAO,EAAE,WAAW,EAAE,MAAM,+BAA+B,CAAC;AAC5D,OAAO,EAAE,eAAe,EAAE,cAAc,EAAE,cAAc,EAAE,MAAM,oBAAoB,CAAC;AAoBrF;;;;;;;;;;;;;;;;GAgBG;AACH,MAAM,CAAC,MAAM,WAAW,GAAG,CAAC,UAAyB,EAAE,OAA4B,EAAS,EAAE;IAC7F,MAAM,EAAE,UAAU,EAAE,GAAG,EAAE,UAAU,EAAE,IAAI,EAAE,WAAW,EAAE,GAAG,OAAO,IAAI,EAAE,CAAC;IAEzE,MAAM,CAAC,IAAI,EAAE,IAAI,CAAC,GACjB,OAAO,UAAU,KAAK,QAAQ;QAC7B,CAAC,CAAC,CAAC,WAAW,IAAI,cAAc,CAAC,UAAU,CAAC,EAAE,QAAQ,CAAC,UAAU,EAAE,EAAE,QAAQ,EAAE,IAAI,EAAE,CAAC,CAAC;QACvF,CAAC,CAAC,CAAC,WAAW,IAAI,eAAe,CAAC,UAAU,CAAC,EAAE,UAAU,CAAC,CAAC;IAE7D,MAAM,GAAG,GAAG,UAAU,IAAI,cAAc,CAAC,IAAI,CAAC,CAAC;IAE/C,MAAM,aAAa,GAAG,WAAW,CAAC,EAAE,IAAI,EAAE,IAAI,EAAE,GAAG,EAAE,EAAE,UAAU,CAAC,CAAC;IAEnE,MAAM,KAAK,GAAG,OAAO,CAAC,aAAa,EAAE;QACpC,QAAQ,EAAE;YACT,IAAI,EAAE,WAAW;YACjB,GAAG,EAAE,gBAAgB;YACrB,GAAG,EAAE,iBAAiB;YACtB,CAAC,EAAE,cAAc,CAAC,EAAE,MAAM,EAAE,OAAO,EAAE,UAAU,EAAE,CAAC;YAClD,GAAG,EAAE,gBAAgB,CAAC,EAAE,SAAS,EAAE,OAAO,EAAE,SAAS,EAAE,CAAC;YACxD,KAAK,EAAE,kBAAkB,CAAC,EAAE,MAAM,EAAE,OAAO,EAAE,WAAW,EAAE,CAAC;SAC3D;KACD,CAAC,CAAC;IAEH,MAAM,cAAc,GAAG,YAAY,CAAC,KAAK,CAAC,CAAC;IAE3C,OAAO,cAAc,CAAC;AACvB,CAAC,CAAC"}
|
|
@@ -1 +1 @@
|
|
|
1
|
-
export declare const linkReplacer: (markdown: string,
|
|
1
|
+
export declare const linkReplacer: (markdown: string, base: string) => string;
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"link-replacer.js","sourceRoot":"","sources":["../../src/link-replacer.ts"],"names":[],"mappings":"AAAA,MAAM,CAAC,MAAM,YAAY,GAAG,CAAC,QAAgB,EAAE,IAAY,EAAE,EAAE;IAC9D,MAAM,KAAK,GAAG,
|
|
1
|
+
{"version":3,"file":"link-replacer.js","sourceRoot":"","sources":["../../src/link-replacer.ts"],"names":[],"mappings":"AAAA,MAAM,CAAC,MAAM,YAAY,GAAG,CAAC,QAAgB,EAAE,IAAY,EAAE,EAAE;IAC9D,MAAM,KAAK,GAAG,6BAA6B,CAAC;IAC5C,OAAO,QAAQ,CAAC,OAAO,CAAC,KAAK,EAAE,CAAC,KAAK,EAAE,GAAG,EAAE,GAAG,EAAE,IAAI,EAAE,EAAE;QACxD,IAAI,cAAc,CAAC,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC;YAC9B,OAAO,KAAK,CAAC;QACd,CAAC;QACD,IAAI,CAAC;YACJ,MAAM,WAAW,GAAG,IAAI,GAAG,CAAC,GAAG,EAAE,IAAI,CAAC,CAAC,IAAI,CAAC;YAC5C,OAAO,GAAG,GAAG,GAAG,WAAW,GAAG,IAAI,EAAE,CAAC;QACtC,CAAC;QAAC,MAAM,CAAC;YACR,OAAO,KAAK,CAAC;QACd,CAAC;IACF,CAAC,CAAC,CAAC;AACJ,CAAC,CAAC"}
|
|
@@ -1,4 +1,10 @@
|
|
|
1
|
-
import { type Browser
|
|
1
|
+
import { type Browser } from "playwright-core";
|
|
2
|
+
export type LoadHtmlOptions = {
|
|
3
|
+
browser?: Browser;
|
|
4
|
+
timeout?: number;
|
|
5
|
+
waitUntil?: "load" | "domcontentloaded" | "networkidle";
|
|
6
|
+
superBypassMode?: boolean;
|
|
7
|
+
};
|
|
2
8
|
/**
|
|
3
9
|
* Useful function for load the HTML of a URL using Playwright.
|
|
4
10
|
* **Not recommended** for use in production environments.
|
|
@@ -13,4 +19,4 @@ import { type Browser, type BrowserContext } from "playwright";
|
|
|
13
19
|
* console.log(html);
|
|
14
20
|
* ```
|
|
15
21
|
*/
|
|
16
|
-
export declare const loadHtml: (url: string,
|
|
22
|
+
export declare const loadHtml: (url: string, options?: LoadHtmlOptions) => Promise<string>;
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"playwright.js","sourceRoot":"","sources":["../../../src/loaders/playwright.ts"],"names":[],"mappings":"AAAA,OAAO,
|
|
1
|
+
{"version":3,"file":"playwright.js","sourceRoot":"","sources":["../../../src/loaders/playwright.ts"],"names":[],"mappings":"AAAA,OAAO,EAAgB,QAAQ,EAAE,OAAO,EAAE,MAAM,iBAAiB,CAAC;AASlE,MAAM,mBAAmB,GAAG;IAC3B,SAAS,EACR,qHAAqH;IACtH,QAAQ,EAAE,EAAE,KAAK,EAAE,IAAI,EAAE,MAAM,EAAE,IAAI,EAAE;IACvC,iBAAiB,EAAE,CAAC;IACpB,QAAQ,EAAE,KAAK;IACf,QAAQ,EAAE,KAAK;IACf,iBAAiB,EAAE,IAAI;IACvB,MAAM,EAAE,OAAO;IACf,UAAU,EAAE,kBAAkB;CAC9B,CAAC;AAEF;;;;;;;;;;;;;GAaG;AACH,MAAM,CAAC,MAAM,QAAQ,GAAG,KAAK,EAAE,GAAW,EAAE,OAAyB,EAAE,EAAE;IACxE,MAAM,EAAE,OAAO,EAAE,SAAS,EAAE,OAAO,EAAE,eAAe,EAAE,GAAG,OAAO,IAAI,EAAE,CAAC;IACvE,MAAM,QAAQ,GAAG,OAAO,IAAI,CAAC,MAAM,QAAQ,CAAC,MAAM,CAAC,EAAE,QAAQ,EAAE,IAAI,EAAE,CAAC,CAAC,CAAC;IACxE,MAAM,OAAO,GAAG,MAAM,QAAQ,CAAC,UAAU,CAAC,eAAe,CAAC,CAAC,CAAC,mBAAmB,CAAC,CAAC,CAAC,OAAO,CAAC,gBAAgB,CAAC,CAAC,CAAC;IAE7G,IAAI,eAAe,EAAE,CAAC;QACrB,MAAM,OAAO,CAAC,aAAa,CAAC,GAAG,EAAE;YAChC,MAAM,CAAC,cAAc,CAAC,SAAS,EAAE,WAAW,EAAE,EAAE,GAAG,EAAE,GAAG,EAAE,CAAC,SAAS,EAAE,CAAC,CAAC;YACxE,MAAM,CAAC,cAAc,CAAC,SAAS,EAAE,WAAW,EAAE,EAAE,GAAG,EAAE,GAAG,EAAE,CAAC,CAAC,OAAO,EAAE,IAAI,CAAC,EAAE,CAAC,CAAC;YAC9E,MAAM,CAAC,cAAc,CAAC,SAAS,EAAE,SAAS,EAAE,EAAE,GAAG,EAAE,GAAG,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,CAAC,EAAE,CAAC,EAAE,CAAC,CAAC,EAAE,CAAC,CAAC;QAC7E,CAAC,CAAC,CAAC;IACJ,CAAC;IAED,MAAM,IAAI,GAAG,MAAM,OAAO,CAAC,OAAO,EAAE,CAAC;IACrC,IAAI,eAAe,EAAE,CAAC;QACrB,MAAM,IAAI,CAAC,KAAK,CAAC,SAAS,EAAE,CAAC,KAAK,EAAE,EAAE;YACrC,IAAI,KAAK,CAAC,OAAO,EAAE,CAAC,GAAG,EAAE,CAAC,QAAQ,CAAC,kBAAkB,CAAC,EAAE,CAAC;gBACxD,OAAO,KAAK,CAAC,KAAK,EAAE,CAAC;YACtB,CAAC;YACD,OAAO,KAAK,CAAC,QAAQ,EAAE,CAAC;QACzB,CAAC,CAAC,CAAC;IACJ,CAAC;IAED,MAAM,IAAI,CAAC,IAAI,CAAC,GAAG,EAAE,EAAE,SAAS,EAAE,SAAS,IAAI,MAAM,EAAE,OAAO,EAAE,CAAC,CAAC;IAClE,MAAM,IAAI,GAAG,MAAM,IAAI,CAAC,OAAO,EAAE,CAAC;IAClC,MAAM,IAAI,CAAC,KAAK,EAAE,CAAC;IAEnB,IAAI,OAAO,EAAE,CAAC;QACb,MAAM,OAAO,CAAC,KAAK,EAAE,CAAC;IACvB,CAAC;SAAM,CAAC;QACP,MAAM,QAAQ,CAAC,KAAK,EAAE,CAAC;IACxB,CAAC;IAED,OAAO,IAAI,CAAC;AACb,CAAC,CAAC"}
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"custom-br-handler.js","sourceRoot":"","sources":["../../../src/mdast-handlers/custom-br-handler.ts"],"names":[],"mappings":"AAAA,OAAO,EAAe,eAAe,EAAE,MAAM,oBAAoB,CAAC;AAElE,MAAM,CAAC,MAAM,eAAe,GAAW,CAAC,KAAK,EAAE,IAAI,EAAE,EAAE;IACtD,OAAO,eAAe,CAAC,EAAE,CAAC,KAAK,EAAE,IAAI,CAAC,CAAC;AACxC,CAAC,CAAC"}
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"mdast-to-markdown.js","sourceRoot":"","sources":["../../src/mdast-to-markdown.ts"],"names":[],"mappings":"AAEA,OAAO,EAAE,aAAa,EAAE,MAAM,gBAAgB,CAAC;AAC/C,OAAO,EAAE,cAAc,EAAE,MAAM,iBAAiB,CAAC;AACjD,OAAO,EAAqC,UAAU,EAAE,MAAM,wBAAwB,CAAC;AAEvF,OAAO,EAAE,YAAY,EAAE,MAAM,iBAAiB,CAAC;AAC/C,OAAO,EAAE,QAAQ,EAAE,MAAM,qBAAqB,CAAC;AAY/C;;GAEG;AACH,MAAM,CAAC,MAAM,iCAAiC,GAA2B;IACxE,UAAU,EAAE,CAAC,aAAa,EAAE,EAAE,cAAc,EAAE,CAAC;IAC/C,MAAM,EAAE,GAAG;CACX,CAAC;AAEF;;;;;;;;;;;;;;;;;;;;;;;;;;GA0BG;AACH,MAAM,CAAC,MAAM,eAAe,GAAG,CAAC,KAA4B,EAAE,OAAgC,EAAU,EAAE;IACzG,MAAM,EAAE,OAAO,EAAE,GAAG,iBAAiB,EAAE,GAAG,EAAE,GAAG,iCAAiC,EAAE,GAAG,OAAO,EAAE,CAAC;IAE/F,IAAI,QAAQ,GAAG,UAAU,CAAC,QAAQ,CAAC,KAAK,CAAC,EAAE,iBAAiB,CAAC,CAAC;
|
|
1
|
+
{"version":3,"file":"mdast-to-markdown.js","sourceRoot":"","sources":["../../src/mdast-to-markdown.ts"],"names":[],"mappings":"AAEA,OAAO,EAAE,aAAa,EAAE,MAAM,gBAAgB,CAAC;AAC/C,OAAO,EAAE,cAAc,EAAE,MAAM,iBAAiB,CAAC;AACjD,OAAO,EAAqC,UAAU,EAAE,MAAM,wBAAwB,CAAC;AAEvF,OAAO,EAAE,YAAY,EAAE,MAAM,iBAAiB,CAAC;AAC/C,OAAO,EAAE,QAAQ,EAAE,MAAM,qBAAqB,CAAC;AAY/C;;GAEG;AACH,MAAM,CAAC,MAAM,iCAAiC,GAA2B;IACxE,UAAU,EAAE,CAAC,aAAa,EAAE,EAAE,cAAc,EAAE,CAAC;IAC/C,MAAM,EAAE,GAAG;CACX,CAAC;AAEF;;;;;;;;;;;;;;;;;;;;;;;;;;GA0BG;AACH,MAAM,CAAC,MAAM,eAAe,GAAG,CAAC,KAA4B,EAAE,OAAgC,EAAU,EAAE;IACzG,MAAM,EAAE,OAAO,EAAE,GAAG,iBAAiB,EAAE,GAAG,EAAE,GAAG,iCAAiC,EAAE,GAAG,OAAO,EAAE,CAAC;IAE/F,IAAI,QAAQ,GAAG,UAAU,CAAC,QAAQ,CAAC,KAAK,CAAC,EAAE,iBAAiB,CAAC,CAAC,OAAO,CAAC,WAAW,EAAE,EAAE,CAAC,CAAC;IAEvF,IAAI,OAAO,EAAE,CAAC;QACb,QAAQ,GAAG,YAAY,CAAC,QAAQ,EAAE,OAAO,CAAC,CAAC;IAC5C,CAAC;IAED,OAAO,QAAQ,CAAC;AACjB,CAAC,CAAC"}
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"hast-utils.js","sourceRoot":"","sources":["../../../src/utils/hast-utils.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,MAAM,EAAE,SAAS,EAAE,MAAM,kBAAkB,CAAC;AAErD,MAAM,CAAC,MAAM,eAAe,GAAG,CAAC,IAAU,EAAE,EAAE;IAC7C,MAAM,IAAI,GAAG,MAAM,CAAC,MAAM,EAAE,IAAI,CAAC,CAAC;IAClC,IAAI,IAAI,IAAI,OAAO,IAAI,CAAC,UAAU,CAAC,IAAI,KAAK,QAAQ,EAAE,CAAC;QACtD,OAAO,IAAI,CAAC,UAAU,CAAC,IAAI,CAAC;IAC7B,CAAC;IACD,IAAI,IAAI,CAAC,IAAI,KAAK,SAAS,EAAE,CAAC;QAC7B,OAAO;IACR,CAAC;IACD,MAAM,OAAO,GAAG,IAAe,CAAC;IAChC,IAAI,OAAO,CAAC,OAAO,KAAK,MAAM,EAAE,CAAC;QAChC,OAAO;IACR,CAAC;IAED,MAAM,QAAQ,GAAG,OAAO,CAAC,UAAU,CAAC,IAAI,IAAI,OAAO,CAAC,UAAU,CAAC,UAAU,CAAC,CAAC;IAC3E,IAAI,QAAQ,EAAE,CAAC;QACd,OAAO,QAAkB,CAAC;IAC3B,CAAC;IAED,OAAO,SAAS,CAAC;AAClB,CAAC,CAAC;AAEF,MAAM,CAAC,MAAM,cAAc,GAAG,CAAC,GAAW,EAAE,EAAE;IAC7C,MAAM,KAAK,GAAG,GAAG,CAAC,KAAK,CAAC,uBAAuB,CAAC,CAAC;IACjD,IAAI,KAAK,EAAE,CAAC;QACX,OAAO,KAAK,CAAC,CAAC,CAAC,CAAC;IACjB,CAAC;IACD,OAAO,SAAS,CAAC;AAClB,CAAC,CAAC;AAEF,MAAM,CAAC,MAAM,cAAc,GAAG,CAAC,IAAU,EAAsB,EAAE;IAChE,IAAI,IAAI,CAAC,IAAI,KAAK,SAAS,EAAE,CAAC;QAC7B,OAAO,SAAS,CAAC;IAClB,CAAC;IAED,MAAM,iBAAiB,GAAG,CAAC,QAAQ,EAAE,aAAa,CAAC,CAAC;IACpD,MAAM,QAAQ,GAAG,SAAS,CAAC,MAAM,EAAE,IAAI,CAAC,CAAC;IAEzC,KAAK,MAAM,IAAI,IAAI,QAAQ,EAAE,CAAC;QAC7B,MAAM,QAAQ,GAAG,IAAI,CAAC,UAAU,CAAC,QAAQ,IAAI,IAAI,CAAC,UAAU,CAAC,IAAI,CAAC;QAClE,IAAI,OAAO,QAAQ,KAAK,QAAQ,IAAI,iBAAiB,CAAC,QAAQ,CAAC,QAAQ,CAAC,EAAE,CAAC;YAC1E,OAAO,OAAO,IAAI,CAAC,UAAU,CAAC,OAAO,KAAK,QAAQ,CAAC,CAAC,CAAC,IAAI,CAAC,UAAU,CAAC,OAAO,CAAC,CAAC,CAAC,SAAS,CAAC;QAC1F,CAAC;IACF,CAAC;IAED,OAAO,SAAS,CAAC;AAClB,CAAC,CAAC"}
|
|
@@ -0,0 +1,46 @@
|
|
|
1
|
+
// src/utils/hast-utils.ts
|
|
2
|
+
import { select, selectAll } from "hast-util-select";
|
|
3
|
+
var getLangFromHast = (node) => {
|
|
4
|
+
const html = select("html", node);
|
|
5
|
+
if (html && typeof html.properties.lang === "string") {
|
|
6
|
+
return html.properties.lang;
|
|
7
|
+
}
|
|
8
|
+
if (node.type !== "element") {
|
|
9
|
+
return;
|
|
10
|
+
}
|
|
11
|
+
const element = node;
|
|
12
|
+
if (element.tagName !== "html") {
|
|
13
|
+
return;
|
|
14
|
+
}
|
|
15
|
+
const langAttr = element.properties.lang || element.properties["xml:lang"];
|
|
16
|
+
if (langAttr) {
|
|
17
|
+
return langAttr;
|
|
18
|
+
}
|
|
19
|
+
return void 0;
|
|
20
|
+
};
|
|
21
|
+
var getLangFromStr = (str) => {
|
|
22
|
+
const match = str.match(/lang=["']([^"']+)["']/);
|
|
23
|
+
if (match) {
|
|
24
|
+
return match[1];
|
|
25
|
+
}
|
|
26
|
+
return void 0;
|
|
27
|
+
};
|
|
28
|
+
var getUrlFromHast = (node) => {
|
|
29
|
+
if (node.type !== "element") {
|
|
30
|
+
return void 0;
|
|
31
|
+
}
|
|
32
|
+
const metaTagAttributes = ["og:url", "twitter:url"];
|
|
33
|
+
const metaTags = selectAll("meta", node);
|
|
34
|
+
for (const meta of metaTags) {
|
|
35
|
+
const property = meta.properties.property || meta.properties.name;
|
|
36
|
+
if (typeof property === "string" && metaTagAttributes.includes(property)) {
|
|
37
|
+
return typeof meta.properties.content === "string" ? meta.properties.content : void 0;
|
|
38
|
+
}
|
|
39
|
+
}
|
|
40
|
+
return void 0;
|
|
41
|
+
};
|
|
42
|
+
export {
|
|
43
|
+
getLangFromHast,
|
|
44
|
+
getLangFromStr,
|
|
45
|
+
getUrlFromHast
|
|
46
|
+
};
|
package/package.json
CHANGED
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
{
|
|
2
2
|
"name": "webforai",
|
|
3
|
-
"version": "1.
|
|
3
|
+
"version": "1.6.0",
|
|
4
4
|
"description": "A library that provides a web interface for AI",
|
|
5
5
|
"author": "inaridiy",
|
|
6
6
|
"license": "MIT",
|
|
@@ -80,20 +80,30 @@
|
|
|
80
80
|
},
|
|
81
81
|
"peerDependencies": {
|
|
82
82
|
"@cloudflare/puppeteer": ">=0.0.6",
|
|
83
|
-
"playwright": ">=1.4",
|
|
83
|
+
"playwright-core": ">=1.4",
|
|
84
84
|
"puppeteer": ">=22"
|
|
85
85
|
},
|
|
86
|
+
"peerDependenciesMeta": {
|
|
87
|
+
"@cloudflare/puppeteer": {
|
|
88
|
+
"optional": true
|
|
89
|
+
},
|
|
90
|
+
"playwright-core": {
|
|
91
|
+
"optional": false
|
|
92
|
+
},
|
|
93
|
+
"puppeteer": {
|
|
94
|
+
"optional": true
|
|
95
|
+
}
|
|
96
|
+
},
|
|
86
97
|
"dependencies": {
|
|
87
98
|
"@clack/prompts": "^0.7.0",
|
|
88
99
|
"commander": "^12.1.0",
|
|
89
|
-
"
|
|
90
|
-
"hast-util-
|
|
91
|
-
"hast-util-
|
|
92
|
-
"hast-util-to-html": "^9.0.0",
|
|
100
|
+
"hast-util-from-html": "^2.0.3",
|
|
101
|
+
"hast-util-select": "^6.0.2",
|
|
102
|
+
"hast-util-to-html": "^9.0.3",
|
|
93
103
|
"hast-util-to-mdast": "^10.1.0",
|
|
94
104
|
"hast-util-to-string": "^3.0.0",
|
|
95
105
|
"hast-util-to-text": "^4.0.0",
|
|
96
|
-
"mathml-to-latex": "^1.
|
|
106
|
+
"mathml-to-latex": "^1.4.1",
|
|
97
107
|
"mdast-util-gfm": "^3.0.0",
|
|
98
108
|
"mdast-util-math": "^3.0.0",
|
|
99
109
|
"mdast-util-to-markdown": "^2.1.0",
|
|
@@ -105,18 +115,19 @@
|
|
|
105
115
|
"devDependencies": {
|
|
106
116
|
"@cloudflare/puppeteer": "^0.0.6",
|
|
107
117
|
"@tsconfig/recommended": "^1.0.3",
|
|
118
|
+
"fastest-levenshtein": "^1.0.16",
|
|
108
119
|
"@types/hast": "^3.0.2",
|
|
109
120
|
"@types/mdast": "^4.0.2",
|
|
110
|
-
"@types/node": "^20.
|
|
121
|
+
"@types/node": "^20.14.10",
|
|
111
122
|
"arg": "^5.0.2",
|
|
112
123
|
"esbuild": "^0.19.11",
|
|
113
124
|
"glob": "^10.3.10",
|
|
114
125
|
"ncp": "^2.0.0",
|
|
115
126
|
"np": "^9.2.0",
|
|
116
|
-
"playwright": "^1.40.1",
|
|
127
|
+
"playwright-core": "^1.40.1",
|
|
117
128
|
"puppeteer": "^23.2.2",
|
|
118
129
|
"rimraf": "^5.0.5",
|
|
119
|
-
"tsx": "^4.
|
|
130
|
+
"tsx": "^4.19.1",
|
|
120
131
|
"typescript": "^5.4.5"
|
|
121
132
|
},
|
|
122
133
|
"scripts": {
|
|
@@ -1,127 +0,0 @@
|
|
|
1
|
-
"use strict";
|
|
2
|
-
var __create = Object.create;
|
|
3
|
-
var __defProp = Object.defineProperty;
|
|
4
|
-
var __getOwnPropDesc = Object.getOwnPropertyDescriptor;
|
|
5
|
-
var __getOwnPropNames = Object.getOwnPropertyNames;
|
|
6
|
-
var __getProtoOf = Object.getPrototypeOf;
|
|
7
|
-
var __hasOwnProp = Object.prototype.hasOwnProperty;
|
|
8
|
-
var __export = (target, all) => {
|
|
9
|
-
for (var name in all)
|
|
10
|
-
__defProp(target, name, { get: all[name], enumerable: true });
|
|
11
|
-
};
|
|
12
|
-
var __copyProps = (to, from, except, desc) => {
|
|
13
|
-
if (from && typeof from === "object" || typeof from === "function") {
|
|
14
|
-
for (let key of __getOwnPropNames(from))
|
|
15
|
-
if (!__hasOwnProp.call(to, key) && key !== except)
|
|
16
|
-
__defProp(to, key, { get: () => from[key], enumerable: !(desc = __getOwnPropDesc(from, key)) || desc.enumerable });
|
|
17
|
-
}
|
|
18
|
-
return to;
|
|
19
|
-
};
|
|
20
|
-
var __toESM = (mod, isNodeMode, target) => (target = mod != null ? __create(__getProtoOf(mod)) : {}, __copyProps(
|
|
21
|
-
// If the importer is in node compatibility mode or this is not an ESM
|
|
22
|
-
// file that has been converted to a CommonJS file using a Babel-
|
|
23
|
-
// compatible transform (i.e. "__esModule" has not been set), then set
|
|
24
|
-
// "default" to the CommonJS "module.exports" for node compatibility.
|
|
25
|
-
isNodeMode || !mod || !mod.__esModule ? __defProp(target, "default", { value: mod, enumerable: true }) : target,
|
|
26
|
-
mod
|
|
27
|
-
));
|
|
28
|
-
var __toCommonJS = (mod) => __copyProps(__defProp({}, "__esModule", { value: true }), mod);
|
|
29
|
-
var loadHtml_exports = {};
|
|
30
|
-
__export(loadHtml_exports, {
|
|
31
|
-
loadHtml: () => loadHtml
|
|
32
|
-
});
|
|
33
|
-
module.exports = __toCommonJS(loadHtml_exports);
|
|
34
|
-
var import_promises = __toESM(require("node:fs/promises"), 1);
|
|
35
|
-
var import_prompts = require("@clack/prompts");
|
|
36
|
-
var import_picocolors = __toESM(require("picocolors"), 1);
|
|
37
|
-
var import_zx = require("zx");
|
|
38
|
-
const loadHtml = async (sourcePath, loader) => {
|
|
39
|
-
const loadSpinner = (0, import_prompts.spinner)();
|
|
40
|
-
try {
|
|
41
|
-
switch (loader.type) {
|
|
42
|
-
case "local":
|
|
43
|
-
return await loadLocalHtml(sourcePath, loadSpinner);
|
|
44
|
-
case "web":
|
|
45
|
-
switch (loader.loader) {
|
|
46
|
-
case "fetch":
|
|
47
|
-
return await loadFetchHtml(sourcePath, loadSpinner);
|
|
48
|
-
case "playwright":
|
|
49
|
-
return await loadPlaywrightHtml(sourcePath, loadSpinner);
|
|
50
|
-
default:
|
|
51
|
-
throw new Error("Invalid loader");
|
|
52
|
-
}
|
|
53
|
-
default:
|
|
54
|
-
throw new Error("Invalid loader type");
|
|
55
|
-
}
|
|
56
|
-
} catch (error) {
|
|
57
|
-
loadSpinner.stop(import_picocolors.default.red("Content loading failed!"));
|
|
58
|
-
console.error(error);
|
|
59
|
-
process.exit(1);
|
|
60
|
-
}
|
|
61
|
-
};
|
|
62
|
-
const loadLocalHtml = async (sourcePath, loadSpinner) => {
|
|
63
|
-
loadSpinner.start("Retrieving content.");
|
|
64
|
-
try {
|
|
65
|
-
await import_promises.default.access(sourcePath);
|
|
66
|
-
} catch {
|
|
67
|
-
throw new Error("The file does not exist");
|
|
68
|
-
}
|
|
69
|
-
const stats = await import_promises.default.stat(sourcePath);
|
|
70
|
-
if (!stats.isFile()) {
|
|
71
|
-
throw new Error("The source must be a file");
|
|
72
|
-
}
|
|
73
|
-
const content = await import_promises.default.readFile(sourcePath, "utf-8");
|
|
74
|
-
loadSpinner.stop("Content retrieval is complete!");
|
|
75
|
-
return content;
|
|
76
|
-
};
|
|
77
|
-
const loadFetchHtml = async (sourcePath, loadSpinner) => {
|
|
78
|
-
loadSpinner.start("Downloading content.");
|
|
79
|
-
const { loadHtml: loadHtml2 } = await import("../../../loaders/fetch");
|
|
80
|
-
const content = await loadHtml2(sourcePath);
|
|
81
|
-
loadSpinner.stop("Content download is complete!");
|
|
82
|
-
return content;
|
|
83
|
-
};
|
|
84
|
-
const loadPlaywrightHtml = async (sourcePath, loadSpinner) => {
|
|
85
|
-
const launchSpinner = (0, import_prompts.spinner)();
|
|
86
|
-
launchSpinner.start("Launching playwright.");
|
|
87
|
-
const playwright = await import("playwright");
|
|
88
|
-
let browser;
|
|
89
|
-
try {
|
|
90
|
-
browser = await playwright.chromium.launch();
|
|
91
|
-
launchSpinner.stop("Playwright launched!");
|
|
92
|
-
} catch (error) {
|
|
93
|
-
await handlePlaywrightLaunchError(error);
|
|
94
|
-
browser = await playwright.chromium.launch();
|
|
95
|
-
}
|
|
96
|
-
loadSpinner.start("Loading content.");
|
|
97
|
-
const { loadHtml: loadHtml2 } = await import("../../../loaders/playwright");
|
|
98
|
-
const content = await loadHtml2(sourcePath, browser);
|
|
99
|
-
loadSpinner.stop("Content download is complete!");
|
|
100
|
-
await browser.close();
|
|
101
|
-
return content;
|
|
102
|
-
};
|
|
103
|
-
const handlePlaywrightLaunchError = async (error) => {
|
|
104
|
-
if (!error.message.includes("npx playwright install")) {
|
|
105
|
-
throw error;
|
|
106
|
-
}
|
|
107
|
-
const isInstall = await (0, import_prompts.confirm)({
|
|
108
|
-
message: "To continue, you will need to download a browser. Do you wish to continue?"
|
|
109
|
-
});
|
|
110
|
-
if (!isInstall) {
|
|
111
|
-
(0, import_prompts.cancel)("End of process.");
|
|
112
|
-
process.exit(1);
|
|
113
|
-
}
|
|
114
|
-
const installDepsSpinner = (0, import_prompts.spinner)();
|
|
115
|
-
installDepsSpinner.start("Installing browser dependencies.");
|
|
116
|
-
try {
|
|
117
|
-
await import_zx.$`npx playwright install`;
|
|
118
|
-
installDepsSpinner.stop("Browser dependencies installed!");
|
|
119
|
-
} catch (error2) {
|
|
120
|
-
installDepsSpinner.stop(import_picocolors.default.red("Browser dependencies failed to install!"));
|
|
121
|
-
throw error2;
|
|
122
|
-
}
|
|
123
|
-
};
|
|
124
|
-
// Annotate the CommonJS export names for ESM import in node:
|
|
125
|
-
0 && (module.exports = {
|
|
126
|
-
loadHtml
|
|
127
|
-
});
|