extract-webpage 1.2.299 → 1.2.300
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/extract-webpage.cjs.js +1 -1
- package/dist/extract-webpage.cjs.js.map +1 -1
- package/dist/extract-webpage.es.js +1 -1
- package/dist/extract-webpage.es.js.map +1 -1
- package/dist/html-to-content/prism-global.d.ts +2 -0
- package/package.json +4 -4
- package/src/html-to-content/html-utils.ts +3 -1
- package/src/html-to-content/prism-global.ts +22 -0
package/package.json
CHANGED
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
{
|
|
2
2
|
"name": "extract-webpage",
|
|
3
|
-
"version": "1.2.
|
|
3
|
+
"version": "1.2.300",
|
|
4
4
|
"module": "./dist/extract-webpage.es.js",
|
|
5
5
|
"description": "Search, extract, cite, and outline the web for a topic with AI Research Agent.",
|
|
6
6
|
"author": "vtempest <grokthiscontact@gmail.com>",
|
|
@@ -74,11 +74,11 @@
|
|
|
74
74
|
"dependencies": {
|
|
75
75
|
"@huggingface/transformers": "^4.2.0",
|
|
76
76
|
"ai": "^7.0.94",
|
|
77
|
-
"chat-agent-toolkit": "^1.2.
|
|
77
|
+
"chat-agent-toolkit": "^1.2.305",
|
|
78
78
|
"chrono-node": "^2.10.1",
|
|
79
79
|
"drizzle-orm": "^0.45.2",
|
|
80
|
-
"extract-pdf": "^0.1.
|
|
81
|
-
"extract-youtube": "^1.0.
|
|
80
|
+
"extract-pdf": "^0.1.289",
|
|
81
|
+
"extract-youtube": "^1.0.293",
|
|
82
82
|
"html-entities": "^2.6.0",
|
|
83
83
|
"js-yaml": "^5.4.1",
|
|
84
84
|
"jsdom": "^30.0.1",
|
|
@@ -139,7 +139,9 @@ export function convertURLToAbsoluteURL(base, relative) {
|
|
|
139
139
|
}
|
|
140
140
|
|
|
141
141
|
import { marked } from "marked";
|
|
142
|
-
import
|
|
142
|
+
// Must precede every `prismjs/components/*` import: the grammar scripts read
|
|
143
|
+
// `Prism` off the global object, which this module publishes (see its docs).
|
|
144
|
+
import Prism from "./prism-global";
|
|
143
145
|
import "prismjs/components/prism-markup.js";
|
|
144
146
|
import "prismjs/components/prism-css.js";
|
|
145
147
|
import "prismjs/components/prism-javascript.js";
|
|
@@ -0,0 +1,22 @@
|
|
|
1
|
+
/**
|
|
2
|
+
* @module html-to-content/prism-global
|
|
3
|
+
* @description Publishes Prism on `globalThis` so the `prismjs/components/*`
|
|
4
|
+
* grammar scripts can find it.
|
|
5
|
+
*
|
|
6
|
+
* Those grammar files are plain browser scripts, not modules — `prism-markup.js`
|
|
7
|
+
* literally opens with `Prism.languages.markup = {...}`, resolving `Prism` as a
|
|
8
|
+
* free variable off the global object. Prism's core only publishes that global
|
|
9
|
+
* when it can see `window` (browser), a `WorkerGlobalScope` `self`, or Node's
|
|
10
|
+
* `global`. On Cloudflare Workers / edge runtimes none of the three exist, so
|
|
11
|
+
* the first grammar file throws `ReferenceError: Prism is not defined` while the
|
|
12
|
+
* module graph is still evaluating, taking the whole SSR render down with it.
|
|
13
|
+
*
|
|
14
|
+
* Import this module *before* any `prismjs/components/*` import. ES modules are
|
|
15
|
+
* evaluated in import order, so this file's body runs — and the global lands —
|
|
16
|
+
* before the grammars reference it.
|
|
17
|
+
*/
|
|
18
|
+
import Prism from "prismjs";
|
|
19
|
+
|
|
20
|
+
(globalThis as typeof globalThis & { Prism?: unknown }).Prism ??= Prism;
|
|
21
|
+
|
|
22
|
+
export default Prism;
|