mdream 1.0.3 → 1.0.5
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/browser.d.mts +0 -3
- package/dist/browser.mjs +0 -2
- package/dist/edge.d.mts +0 -3
- package/dist/edge.mjs +0 -2
- package/dist/iife.js +17 -18
- package/dist/index.d.mts +0 -2
- package/dist/index.mjs +0 -2
- package/dist/worker.d.mts +0 -4
- package/dist/worker.mjs +0 -13
- package/package.json +15 -15
- package/wasm/mdream_edge.js +16 -17
- package/wasm/mdream_edge_bg.wasm +0 -0
- package/wasm/package.json +1 -1
package/dist/index.d.mts
CHANGED
|
@@ -1,4 +1,3 @@
|
|
|
1
|
-
//#region src/index.d.ts
|
|
2
1
|
interface CleanOptions {
|
|
3
2
|
/** Strip tracking query parameters (utm_*, fbclid, gclid, etc.) from URLs */
|
|
4
3
|
urls?: boolean;
|
|
@@ -73,5 +72,4 @@ interface MdreamOptions {
|
|
|
73
72
|
}
|
|
74
73
|
declare function htmlToMarkdown(html: string, options?: Partial<MdreamOptions>): string;
|
|
75
74
|
declare function streamHtmlToMarkdown(htmlStream: ReadableStream<Uint8Array | string> | null, options?: Partial<MdreamOptions>): AsyncIterable<string>;
|
|
76
|
-
//#endregion
|
|
77
75
|
export { CleanOptions, ExtractedElement, FrontmatterConfig, MdreamOptions, TagOverride, htmlToMarkdown, streamHtmlToMarkdown };
|
package/dist/index.mjs
CHANGED
|
@@ -1,5 +1,4 @@
|
|
|
1
1
|
import { MarkdownStream, htmlToMarkdown as htmlToMarkdown$1 } from "../napi/index.mjs";
|
|
2
|
-
//#region src/index.ts
|
|
3
2
|
const MINIMAL_FILTER_DEFAULT = { exclude: [
|
|
4
3
|
"form",
|
|
5
4
|
"fieldset",
|
|
@@ -110,5 +109,4 @@ async function* streamHtmlToMarkdown(htmlStream, options = {}) {
|
|
|
110
109
|
reader.releaseLock();
|
|
111
110
|
}
|
|
112
111
|
}
|
|
113
|
-
//#endregion
|
|
114
112
|
export { htmlToMarkdown, streamHtmlToMarkdown };
|
package/dist/worker.d.mts
CHANGED
|
@@ -1,4 +1,3 @@
|
|
|
1
|
-
//#region src/index.d.ts
|
|
2
1
|
interface CleanOptions {
|
|
3
2
|
/** Strip tracking query parameters (utm_*, fbclid, gclid, etc.) from URLs */
|
|
4
3
|
urls?: boolean;
|
|
@@ -71,8 +70,6 @@ interface MdreamOptions {
|
|
|
71
70
|
/** Tag overrides. String values act as aliases */
|
|
72
71
|
tagOverrides?: Record<string, TagOverride | string>;
|
|
73
72
|
}
|
|
74
|
-
//#endregion
|
|
75
|
-
//#region src/worker.d.ts
|
|
76
73
|
/**
|
|
77
74
|
* Initialize the mdream web worker. Must be called before htmlToMarkdown.
|
|
78
75
|
* @param wasmUrl - URL to the mdream_edge_bg.wasm file
|
|
@@ -87,5 +84,4 @@ declare function htmlToMarkdown(html: string, options?: Partial<MdreamOptions>):
|
|
|
87
84
|
* Terminate the web worker and free resources.
|
|
88
85
|
*/
|
|
89
86
|
declare function terminateWorker(): void;
|
|
90
|
-
//#endregion
|
|
91
87
|
export { htmlToMarkdown, initWorker, terminateWorker };
|
package/dist/worker.mjs
CHANGED
|
@@ -1,4 +1,3 @@
|
|
|
1
|
-
//#region src/worker.ts
|
|
2
1
|
let _worker = null;
|
|
3
2
|
let _ready = null;
|
|
4
3
|
let _idCounter = 0;
|
|
@@ -48,10 +47,6 @@ function onMessage(e) {
|
|
|
48
47
|
else pending.reject(new Error(msg.message));
|
|
49
48
|
}
|
|
50
49
|
}
|
|
51
|
-
/**
|
|
52
|
-
* Initialize the mdream web worker. Must be called before htmlToMarkdown.
|
|
53
|
-
* @param wasmUrl - URL to the mdream_edge_bg.wasm file
|
|
54
|
-
*/
|
|
55
50
|
function initWorker(wasmUrl) {
|
|
56
51
|
if (_ready) return _ready;
|
|
57
52
|
_ready = new Promise((resolve, reject) => {
|
|
@@ -71,10 +66,6 @@ function initWorker(wasmUrl) {
|
|
|
71
66
|
});
|
|
72
67
|
return _ready;
|
|
73
68
|
}
|
|
74
|
-
/**
|
|
75
|
-
* Convert HTML to markdown using the web worker.
|
|
76
|
-
* Call initWorker() first.
|
|
77
|
-
*/
|
|
78
69
|
function htmlToMarkdown(html, options) {
|
|
79
70
|
if (!_worker || !_ready) return Promise.reject(/* @__PURE__ */ new Error("Call initWorker() before htmlToMarkdown()"));
|
|
80
71
|
return _ready.then(() => {
|
|
@@ -93,9 +84,6 @@ function htmlToMarkdown(html, options) {
|
|
|
93
84
|
});
|
|
94
85
|
});
|
|
95
86
|
}
|
|
96
|
-
/**
|
|
97
|
-
* Terminate the web worker and free resources.
|
|
98
|
-
*/
|
|
99
87
|
function terminateWorker() {
|
|
100
88
|
if (_worker) {
|
|
101
89
|
_worker.terminate();
|
|
@@ -105,5 +93,4 @@ function terminateWorker() {
|
|
|
105
93
|
for (const [, pending] of _pending) pending.reject(/* @__PURE__ */ new Error("Worker terminated"));
|
|
106
94
|
_pending.clear();
|
|
107
95
|
}
|
|
108
|
-
//#endregion
|
|
109
96
|
export { htmlToMarkdown, initWorker, terminateWorker };
|
package/package.json
CHANGED
|
@@ -1,7 +1,7 @@
|
|
|
1
1
|
{
|
|
2
2
|
"name": "mdream",
|
|
3
3
|
"type": "module",
|
|
4
|
-
"version": "1.0.
|
|
4
|
+
"version": "1.0.5",
|
|
5
5
|
"description": "Ultra-performant HTML to Markdown Convertor Optimized for LLMs and llm.txt artifacts.",
|
|
6
6
|
"author": {
|
|
7
7
|
"name": "Harlan Wilton",
|
|
@@ -51,22 +51,22 @@
|
|
|
51
51
|
"browser": "./dist/browser.mjs",
|
|
52
52
|
"dependencies": {},
|
|
53
53
|
"optionalDependencies": {
|
|
54
|
-
"@mdream/rust-
|
|
55
|
-
"@mdream/rust-android-arm64": "1.0.
|
|
56
|
-
"@mdream/rust-freebsd-x64": "1.0.
|
|
57
|
-
"@mdream/rust-
|
|
58
|
-
"@mdream/rust-darwin-
|
|
59
|
-
"@mdream/rust-linux-arm-gnueabihf": "1.0.
|
|
60
|
-
"@mdream/rust-linux-arm64-
|
|
61
|
-
"@mdream/rust-linux-arm64-
|
|
62
|
-
"@mdream/rust-linux-x64-
|
|
63
|
-
"@mdream/rust-
|
|
64
|
-
"@mdream/rust-
|
|
65
|
-
"@mdream/rust-win32-arm64-msvc": "1.0.
|
|
66
|
-
"@mdream/rust-win32-x64-msvc": "1.0.
|
|
54
|
+
"@mdream/rust-android-arm-eabi": "1.0.5",
|
|
55
|
+
"@mdream/rust-android-arm64": "1.0.5",
|
|
56
|
+
"@mdream/rust-freebsd-x64": "1.0.5",
|
|
57
|
+
"@mdream/rust-darwin-x64": "1.0.5",
|
|
58
|
+
"@mdream/rust-darwin-arm64": "1.0.5",
|
|
59
|
+
"@mdream/rust-linux-arm-gnueabihf": "1.0.5",
|
|
60
|
+
"@mdream/rust-linux-arm64-gnu": "1.0.5",
|
|
61
|
+
"@mdream/rust-linux-arm64-musl": "1.0.5",
|
|
62
|
+
"@mdream/rust-linux-x64-gnu": "1.0.5",
|
|
63
|
+
"@mdream/rust-linux-x64-musl": "1.0.5",
|
|
64
|
+
"@mdream/rust-wasm32-wasi": "1.0.5",
|
|
65
|
+
"@mdream/rust-win32-arm64-msvc": "1.0.5",
|
|
66
|
+
"@mdream/rust-win32-x64-msvc": "1.0.5"
|
|
67
67
|
},
|
|
68
68
|
"devDependencies": {
|
|
69
|
-
"@mdream/js": "1.0.
|
|
69
|
+
"@mdream/js": "1.0.5"
|
|
70
70
|
},
|
|
71
71
|
"scripts": {
|
|
72
72
|
"flame": "pnpm build && unbuild bench/bundle && clinic flame -- node bench/bundle/dist/string.mjs 10",
|
package/wasm/mdream_edge.js
CHANGED
|
@@ -99,30 +99,29 @@ export function htmlToMarkdownResult(html, options) {
|
|
|
99
99
|
const ret = wasm.htmlToMarkdownResult(ptr0, len0, addHeapObject(options));
|
|
100
100
|
return takeObject(ret);
|
|
101
101
|
}
|
|
102
|
-
|
|
103
102
|
function __wbg_get_imports() {
|
|
104
103
|
const import0 = {
|
|
105
104
|
__proto__: null,
|
|
106
|
-
|
|
105
|
+
__wbg___wbindgen_boolean_get_6ea149f0a8dcc5ff: function(arg0) {
|
|
107
106
|
const v = getObject(arg0);
|
|
108
107
|
const ret = typeof(v) === 'boolean' ? v : undefined;
|
|
109
108
|
return isLikeNone(ret) ? 0xFFFFFF : ret ? 1 : 0;
|
|
110
109
|
},
|
|
111
|
-
|
|
110
|
+
__wbg___wbindgen_is_null_52ff4ec04186736f: function(arg0) {
|
|
112
111
|
const ret = getObject(arg0) === null;
|
|
113
112
|
return ret;
|
|
114
113
|
},
|
|
115
|
-
|
|
114
|
+
__wbg___wbindgen_is_undefined_29a43b4d42920abd: function(arg0) {
|
|
116
115
|
const ret = getObject(arg0) === undefined;
|
|
117
116
|
return ret;
|
|
118
117
|
},
|
|
119
|
-
|
|
118
|
+
__wbg___wbindgen_number_get_c7f42aed0525c451: function(arg0, arg1) {
|
|
120
119
|
const obj = getObject(arg1);
|
|
121
120
|
const ret = typeof(obj) === 'number' ? obj : undefined;
|
|
122
121
|
getDataViewMemory0().setFloat64(arg0 + 8 * 1, isLikeNone(ret) ? 0 : ret, true);
|
|
123
122
|
getDataViewMemory0().setInt32(arg0 + 4 * 0, !isLikeNone(ret), true);
|
|
124
123
|
},
|
|
125
|
-
|
|
124
|
+
__wbg___wbindgen_string_get_7ed5322991caaec5: function(arg0, arg1) {
|
|
126
125
|
const obj = getObject(arg1);
|
|
127
126
|
const ret = typeof(obj) === 'string' ? obj : undefined;
|
|
128
127
|
var ptr1 = isLikeNone(ret) ? 0 : passStringToWasm0(ret, wasm.__wbindgen_export, wasm.__wbindgen_export2);
|
|
@@ -130,46 +129,46 @@ function __wbg_get_imports() {
|
|
|
130
129
|
getDataViewMemory0().setInt32(arg0 + 4 * 1, len1, true);
|
|
131
130
|
getDataViewMemory0().setInt32(arg0 + 4 * 0, ptr1, true);
|
|
132
131
|
},
|
|
133
|
-
|
|
132
|
+
__wbg___wbindgen_throw_6b64449b9b9ed33c: function(arg0, arg1) {
|
|
134
133
|
throw new Error(getStringFromWasm0(arg0, arg1));
|
|
135
134
|
},
|
|
136
|
-
|
|
135
|
+
__wbg_entries_e0b73aa8571ddb56: function(arg0) {
|
|
137
136
|
const ret = Object.entries(getObject(arg0));
|
|
138
137
|
return addHeapObject(ret);
|
|
139
138
|
},
|
|
140
|
-
|
|
139
|
+
__wbg_from_0dbf29f09e7fb200: function(arg0) {
|
|
141
140
|
const ret = Array.from(getObject(arg0));
|
|
142
141
|
return addHeapObject(ret);
|
|
143
142
|
},
|
|
144
|
-
|
|
143
|
+
__wbg_get_6011fa3a58f61074: function() { return handleError(function (arg0, arg1) {
|
|
145
144
|
const ret = Reflect.get(getObject(arg0), getObject(arg1));
|
|
146
145
|
return addHeapObject(ret);
|
|
147
146
|
}, arguments); },
|
|
148
|
-
|
|
147
|
+
__wbg_get_8360291721e2339f: function(arg0, arg1) {
|
|
149
148
|
const ret = getObject(arg0)[arg1 >>> 0];
|
|
150
149
|
return addHeapObject(ret);
|
|
151
150
|
},
|
|
152
|
-
|
|
151
|
+
__wbg_isArray_c3109d14ffc06469: function(arg0) {
|
|
153
152
|
const ret = Array.isArray(getObject(arg0));
|
|
154
153
|
return ret;
|
|
155
154
|
},
|
|
156
|
-
|
|
155
|
+
__wbg_length_3d4ecd04bd8d22f1: function(arg0) {
|
|
157
156
|
const ret = getObject(arg0).length;
|
|
158
157
|
return ret;
|
|
159
158
|
},
|
|
160
|
-
|
|
159
|
+
__wbg_new_682678e2f47e32bc: function() {
|
|
161
160
|
const ret = new Array();
|
|
162
161
|
return addHeapObject(ret);
|
|
163
162
|
},
|
|
164
|
-
|
|
163
|
+
__wbg_new_aa8d0fa9762c29bd: function() {
|
|
165
164
|
const ret = new Object();
|
|
166
165
|
return addHeapObject(ret);
|
|
167
166
|
},
|
|
168
|
-
|
|
167
|
+
__wbg_push_471a5b068a5295f6: function(arg0, arg1) {
|
|
169
168
|
const ret = getObject(arg0).push(getObject(arg1));
|
|
170
169
|
return ret;
|
|
171
170
|
},
|
|
172
|
-
|
|
171
|
+
__wbg_set_022bee52d0b05b19: function() { return handleError(function (arg0, arg1, arg2) {
|
|
173
172
|
const ret = Reflect.set(getObject(arg0), getObject(arg1), getObject(arg2));
|
|
174
173
|
return ret;
|
|
175
174
|
}, arguments); },
|
package/wasm/mdream_edge_bg.wasm
CHANGED
|
Binary file
|
package/wasm/package.json
CHANGED