@plurnk/plurnk-mimetypes-text-html 1.0.2 → 1.0.4
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/TextHtml.d.ts.map
CHANGED
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"TextHtml.d.ts","sourceRoot":"","sources":["../src/TextHtml.ts"],"names":[],"mappings":"AAAA,OAAO,EACH,WAAW,EAGd,MAAM,0BAA0B,CAAC;AAClC,OAAO,KAAK,EACR,cAAc,EACd,UAAU,EACV,YAAY,EACZ,UAAU,EACb,MAAM,0BAA0B,CAAC;AAgClC,MAAM,CAAC,OAAO,OAAO,QAAS,SAAQ,WAAW;IACpC,UAAU,CAAC,OAAO,EAAE,MAAM,GAAG,UAAU,GAAG,UAAU,EAAE,CAwB9D;IAUQ,QAAQ,CAAC,OAAO,EAAE,cAAc,GAAG,OAAO,CAiBlD;IAWQ,OAAO,CAAC,OAAO,EAAE,cAAc,GAAG,MAAM,GAAG,SAAS,CAK5D;
|
|
1
|
+
{"version":3,"file":"TextHtml.d.ts","sourceRoot":"","sources":["../src/TextHtml.ts"],"names":[],"mappings":"AAAA,OAAO,EACH,WAAW,EAGd,MAAM,0BAA0B,CAAC;AAClC,OAAO,KAAK,EACR,cAAc,EACd,UAAU,EACV,YAAY,EACZ,UAAU,EACb,MAAM,0BAA0B,CAAC;AAgClC,MAAM,CAAC,OAAO,OAAO,QAAS,SAAQ,WAAW;IACpC,UAAU,CAAC,OAAO,EAAE,MAAM,GAAG,UAAU,GAAG,UAAU,EAAE,CAwB9D;IAUQ,QAAQ,CAAC,OAAO,EAAE,cAAc,GAAG,OAAO,CAiBlD;IAWQ,OAAO,CAAC,OAAO,EAAE,cAAc,GAAG,MAAM,GAAG,SAAS,CAK5D;IASD,UAAmB,MAAM,CAAC,OAAO,EAAE,cAAc,GAAG,MAAM,CAKzD;IAQc,KAAK,CAChB,OAAO,EAAE,cAAc,EACvB,OAAO,EAAE,YAAY,EACrB,OAAO,EAAE,MAAM,EACf,KAAK,CAAC,EAAE,MAAM,GACf,OAAO,CAAC,UAAU,EAAE,CAAC,CAgCvB;CACJ"}
|
package/dist/TextHtml.js
CHANGED
|
@@ -72,14 +72,15 @@ export default class TextHtml extends BaseHandler {
|
|
|
72
72
|
// Route the regex/glob query surface (and, transitively, the framework's
|
|
73
73
|
// content()??toText() embed-source) through the SAME markdown projection,
|
|
74
74
|
// so body matchers scan the readable text, not raw `<div class>` markup.
|
|
75
|
-
// xpath is unaffected — query() overrides it to hit the real DOM.
|
|
76
|
-
//
|
|
77
|
-
//
|
|
75
|
+
// xpath is unaffected — query() overrides it to hit the real DOM. A page
|
|
76
|
+
// with no readable content projects to EMPTY text (regex/glob match nothing,
|
|
77
|
+
// honestly) — never the raw body: that raw fall-through is what fed 2MB of
|
|
78
|
+
// DOM to the embedder and wedged the daemon (#412). Empty when empty.
|
|
78
79
|
toText(content) {
|
|
79
80
|
const html = typeof content === "string"
|
|
80
81
|
? content
|
|
81
82
|
: new TextDecoder("utf-8").decode(content);
|
|
82
|
-
return htmlToMarkdown(html) ??
|
|
83
|
+
return htmlToMarkdown(html) ?? "";
|
|
83
84
|
}
|
|
84
85
|
// Override xpath dispatch. parse5's tree isn't xpath-traversable, so we
|
|
85
86
|
// re-parse via @xmldom/xmldom (which produces a real DOM that the `xpath`
|
package/dist/TextHtml.js.map
CHANGED
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"TextHtml.js","sourceRoot":"","sources":["../src/TextHtml.ts"],"names":[],"mappings":"AAAA,OAAO,EACH,WAAW,EACX,sBAAsB,EACtB,sBAAsB,GACzB,MAAM,0BAA0B,CAAC;AAOlC,OAAO,EAAE,KAAK,EAAE,MAAM,QAAQ,CAAC;AAE/B,OAAO,EAAE,SAAS,EAAE,MAAM,gBAAgB,CAAC;AAC3C,OAAO,KAAK,KAAK,MAAM,OAAO,CAAC;AAC/B,OAAO,EAAE,cAAc,EAAE,MAAM,qBAAqB,CAAC;AAyBrD,MAAM,YAAY,GAAG,IAAI,GAAG,CAAC,CAAC,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,CAAC,CAAC,CAAC;AAEnE,MAAM,CAAC,OAAO,OAAO,QAAS,SAAQ,WAAW;IACpC,UAAU,CAAC,OAA4B;QAC5C,MAAM,IAAI,GAAG,OAAO,OAAO,KAAK,QAAQ;YACpC,CAAC,CAAC,OAAO;YACT,CAAC,CAAC,IAAI,WAAW,CAAC,OAAO,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC;QAE/C,MAAM,GAAG,GAAG,KAAK,CAAC,IAAI,EAAE,EAAE,sBAAsB,EAAE,IAAI,EAAE,CAAC,CAAC;QAC1D,MAAM,OAAO,GAAiB,EAAE,CAAC;QACjC,MAAM,SAAS,GAA0C,iBAAiB,CAAC,GAAG,EAAE,OAAO,CAAC,CAAC;QAEzF,wEAAwE;QACxE,qEAAqE;QACrE,oEAAoE;QACpE,MAAM,KAAK,GAAG,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,KAAK,SAAS,IAAI,CAAC,CAAC,KAAK,KAAK,CAAC,CAAC,CAAC;QACzE,IAAI,SAAS,KAAK,IAAI,IAAI,CAAC,KAAK,EAAE,CAAC;YAC/B,OAAO,CAAC,OAAO,CAAC;gBACZ,IAAI,EAAE,SAAS,CAAC,IAAI;gBACpB,IAAI,EAAE,SAAS;gBACf,KAAK,EAAE,CAAC;gBACR,IAAI,EAAE,SAAS,CAAC,IAAI;gBACpB,OAAO,EAAE,SAAS,CAAC,IAAI;aAC1B,CAAC,CAAC;QACP,CAAC;QAED,OAAO,OAAO,CAAC;IACnB,CAAC;IAED,yEAAyE;IACzE,mEAAmE;IACnE,wEAAwE;IACxE,sEAAsE;IACtE,iEAAiE;IACjE,oEAAoE;IACpE,qEAAqE;IACrE,0CAA0C;IACjC,QAAQ,CAAC,OAAuB;QACrC,MAAM,IAAI,GAAG,OAAO,OAAO,KAAK,QAAQ;YACpC,CAAC,CAAC,OAAO;YACT,CAAC,CAAC,IAAI,WAAW,CAAC,OAAO,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC;QAC/C,IAAI,GAAG,CAAC;QACR,IAAI,CAAC;YACD,GAAG,GAAG,KAAK,CAAC,IAAI,EAAE,EAAE,sBAAsB,EAAE,IAAI,EAAE,CAAC,CAAC;QACxD,CAAC;QAAC,MAAM,CAAC;YACL,OAAO,IAAI,CAAC;QAChB,CAAC;QACD,MAAM,IAAI,GAA4B;YAClC,IAAI,EAAE,UAAU;YAChB,IAAI,EAAE,CAAC;YACP,OAAO,EAAE,CAAC;YACV,QAAQ,EAAE,eAAe,CAAC,GAAG,CAAC;SACjC,CAAC;QACF,OAAO,IAAI,CAAC;IAChB,CAAC;IAED,wEAAwE;IACxE,2EAA2E;IAC3E,wEAAwE;IACxE,sEAAsE;IACtE,0EAA0E;IAC1E,sEAAsE;IACtE,oEAAoE;IACpE,4EAA4E;IAC5E,uBAAuB;IACd,OAAO,CAAC,OAAuB;QACpC,MAAM,IAAI,GAAG,OAAO,OAAO,KAAK,QAAQ;YACpC,CAAC,CAAC,OAAO;YACT,CAAC,CAAC,IAAI,WAAW,CAAC,OAAO,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC;QAC/C,OAAO,cAAc,CAAC,IAAI,CAAC,CAAC;IAChC,CAAC;IAED,yEAAyE;IACzE,0EAA0E;IAC1E,yEAAyE;IACzE,2EAA2E;IAC3E,wEAAwE;IACxE,sDAAsD;IACnC,MAAM,CAAC,OAAuB;QAC7C,MAAM,IAAI,GAAG,OAAO,OAAO,KAAK,QAAQ;YACpC,CAAC,CAAC,OAAO;YACT,CAAC,CAAC,IAAI,WAAW,CAAC,OAAO,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC;QAC/C,OAAO,cAAc,CAAC,IAAI,CAAC,IAAI,IAAI,CAAC;IACxC,CAAC;IAED,wEAAwE;IACxE,0EAA0E;IAC1E,sEAAsE;IACtE,uEAAuE;IACvE,yEAAyE;IACzE,yEAAyE;IAChE,KAAK,CAAC,KAAK,CAChB,OAAuB,EACvB,OAAqB,EACrB,OAAe,EACf,KAAc;QAEd,IAAI,OAAO,KAAK,OAAO,EAAE,CAAC;YACtB,MAAM,IAAI,GAAG,OAAO,OAAO,KAAK,QAAQ;gBACpC,CAAC,CAAC,OAAO;gBACT,CAAC,CAAC,IAAI,WAAW,CAAC,OAAO,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC;YAE/C,2DAA2D;YAC3D,kEAAkE;YAClE,iEAAiE;YACjE,6DAA6D;YAC7D,6DAA6D;YAC7D,+DAA+D;YAC/D,6DAA6D;YAC7D,+DAA+D;YAC/D,qDAAqD;YACrD,IAAI,GAAG,CAAC;YACR,IAAI,CAAC;gBACD,GAAG,GAAG,IAAI,SAAS,EAAE,CAAC,eAAe,CAAC,IAAI,EAAE,UAAU,CAAC,CAAC;YAC5D,CAAC;YAAC,OAAO,KAAK,EAAE,CAAC;gBACb,MAAM,IAAI,sBAAsB,CAAC,EAAE,QAAQ,EAAE,IAAI,CAAC,QAAQ,EAAE,KAAK,EAAE,CAAC,CAAC;YACzE,CAAC;YAED,IAAI,MAA8B,CAAC;YACnC,IAAI,CAAC;gBACD,MAAM,GAAG,KAAK,CAAC,MAAM,CAAC,OAAO,EAAE,GAAsB,CAAC,CAAC;YAC3D,CAAC;YAAC,OAAO,KAAK,EAAE,CAAC;gBACb,MAAM,IAAI,sBAAsB,CAAC,EAAE,OAAO,EAAE,OAAO,EAAE,UAAU,EAAE,OAAO,EAAE,KAAK,EAAE,CAAC,CAAC;YACvF,CAAC;YAED,OAAO,gBAAgB,CAAC,OAAO,EAAE,MAAM,CAAC,CAAC;QAC7C,CAAC;QACD,OAAO,KAAK,CAAC,KAAK,CAAC,OAAO,EAAE,OAAO,EAAE,OAAO,EAAE,KAAK,CAAC,CAAC;IACzD,CAAC;CACJ;AAED,+EAA+E;AAC/E,2EAA2E;AAC3E,8BAA8B;AAC9B,SAAS,gBAAgB,CAAC,OAAe,EAAE,MAA8B;IACrE,IAAI,KAAK,CAAC,OAAO,CAAC,MAAM,CAAC,EAAE,CAAC;QACxB,OAAO,MAAM,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC,EAAc,EAAE;YACtC,MAAM,IAAI,GAAG,QAAQ,CAAC,IAAI,CAAC,CAAC;YAC5B,MAAM,OAAO,GAAG,IAAI,KAAK,SAAS,CAAC,CAAC,CAAC,OAAO,CAAC,IAAI,EAAE,IAAI,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC;YAChE,OAAO;gBACH,OAAO,EAAE,aAAa,CAAC,IAAI,CAAC;gBAC5B,QAAQ,EAAE,MAAM,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,IAAI,OAAO,KAAK,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC,CAAC,SAAS;gBAClE,GAAG,CAAC,IAAI,KAAK,SAAS,IAAI,EAAE,KAAK,EAAE,CAAC,EAAE,IAAI,EAAE,OAAO,EAAE,OAAO,IAAI,IAAI,EAAE,CAAC,EAAE,CAAC;aAC7E,CAAC;QACN,CAAC,CAAC,CAAC;IACP,CAAC;IACD,IAAI,MAAM,KAAK,IAAI,IAAI,MAAM,KAAK,SAAS;QAAE,OAAO,EAAE,CAAC;IACvD,4EAA4E;IAC5E,yDAAyD;IACzD,OAAO,CAAC,EAAE,OAAO,EAAE,OAAO,MAAM,KAAK,QAAQ,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,MAAM,CAAC,MAAM,CAAC,EAAE,CAAC,CAAC;AAC/E,CAAC;AAED,+EAA+E;AAC/E,8EAA8E;AAC9E,iCAAiC;AACjC,SAAS,QAAQ,CAAC,IAAU;IACxB,MAAM,MAAM,GAAI,IAAgC,CAAC,UAAU,CAAC;IAC5D,IAAI,OAAO,MAAM,KAAK,QAAQ,IAAI,MAAM,GAAG,CAAC;QAAE,OAAO,MAAM,CAAC;IAC5D,MAAM,KAAK,GAAI,IAAa,CAAC,YAAY;WACjC,IAAqC,CAAC,UAAU,IAAI,IAAI,CAAC;IACjE,MAAM,EAAE,GAAI,KAAwC,EAAE,UAAU,CAAC;IACjE,OAAO,OAAO,EAAE,KAAK,QAAQ,IAAI,EAAE,GAAG,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,SAAS,CAAC;AAC7D,CAAC;AAED,MAAM,iBAAiB,GAAG,CAAC,CAAC;AAC5B,8EAA8E;AAC9E,4EAA4E;AAC5E,0BAA0B;AAC1B,SAAS,OAAO,CAAC,IAAU,EAAE,SAAiB;IAC1C,MAAM,UAAU,GAAG,aAAa,CAAC,IAAI,CAAC,CAAC;IACvC,OAAO,SAAS,GAAG,CAAC,UAAU,CAAC,KAAK,CAAC,KAAK,CAAC,EAAE,MAAM,IAAI,CAAC,CAAC,CAAC;AAC9D,CAAC;AAED,4EAA4E;AAC5E,4EAA4E;AAC5E,6EAA6E;AAC7E,wEAAwE;AACxE,yCAAyC;AACzC,MAAM,cAAc,GAAG,CAAC,CAAC;AACzB,MAAM,SAAS,GAAG,CAAC,CAAC;AACpB,MAAM,kBAAkB,GAAG,CAAC,CAAC;AAC7B,MAAM,2BAA2B,GAAG,CAAC,CAAC;AACtC,MAAM,YAAY,GAAG,CAAC,CAAC;AACvB,SAAS,aAAa,CAAC,IAAU;IAC7B,MAAM,EAAE,GAAG,IAAI,CAAC,QAAQ,CAAC;IACzB,IAAI,EAAE,KAAK,cAAc;QAAE,OAAQ,IAAa,CAAC,KAAK,CAAC;IACvD,IAAI,EAAE,KAAK,SAAS,IAAI,EAAE,KAAK,kBAAkB;QAAE,OAAQ,IAAa,CAAC,IAAI,CAAC;IAC9E,IAAI,EAAE,KAAK,YAAY;QAAE,OAAQ,IAAgB,CAAC,IAAI,CAAC;IACvD,IAAI,EAAE,KAAK,2BAA2B;QAAE,OAAQ,IAA8B,CAAC,IAAI,CAAC;IACpF,OAAQ,IAA8C,CAAC,QAAQ,EAAE,CAAC;AACtE,CAAC;AAED,4EAA4E;AAC5E,uEAAuE;AACvE,wBAAwB;AACxB,SAAS,iBAAiB,CACtB,IAAgB,EAChB,GAAiB;IAEjB,IAAI,KAAK,GAA0C,IAAI,CAAC;IACxD,sEAAsE;IACtE,yEAAyE;IACzE,iEAAiE;IACjE,MAAM,IAAI,GAA2C,EAAE,CAAC;IAExD,SAAS,IAAI,CAAC,IAA4B;QACtC,IAAI,CAAC,SAAS,CAAC,IAAI,CAAC,EAAE,CAAC;YACnB,IAAI,aAAa,CAAC,IAAI,CAAC,EAAE,CAAC;gBACtB,KAAK,MAAM,KAAK,IAAI,IAAI,CAAC,UAAU;oBAAE,IAAI,CAAC,KAAK,CAAC,CAAC;YACrD,CAAC;YACD,OAAO;QACX,CAAC;QAED,MAAM,GAAG,GAAG,IAAI,CAAC,OAAO,CAAC;QACzB,IAAI,GAAG,KAAK,OAAO,IAAI,KAAK,KAAK,IAAI,EAAE,CAAC;YACpC,MAAM,IAAI,GAAG,WAAW,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,CAAC;YACtC,IAAI,IAAI,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;gBAClB,KAAK,GAAG;oBACJ,IAAI,EAAE,IAAI;oBACV,IAAI,EAAE,IAAI,CAAC,kBAAkB,EAAE,SAAS,IAAI,CAAC;iBAChD,CAAC;YACN,CAAC;QACL,CAAC;aAAM,IAAI,YAAY,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,CAAC;YAC/B,MAAM,IAAI,GAAG,WAAW,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,CAAC;YACtC,MAAM,GAAG,GAAG,IAAI,CAAC,kBAAkB,CAAC;YACpC,IAAI,IAAI,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;gBAClB,MAAM,KAAK,GAAG,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC;gBAC7B,OAAO,IAAI,CAAC,MAAM,GAAG,CAAC,IAAI,IAAI,CAAC,IAAI,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,KAAK,IAAI,KAAK,EAAE,CAAC;oBAC7D,IAAI,CAAC,GAAG,EAAE,CAAC;gBACf,CAAC;gBACD,MAAM,SAAS,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;gBACpD,GAAG,CAAC,IAAI,CAAC;oBACL,IAAI,EAAE,IAAI;oBACV,IAAI,EAAE,SAAS;oBACf,KAAK;oBACL,IAAI,EAAE,GAAG,EAAE,SAAS,IAAI,CAAC;oBACzB,OAAO,EAAE,GAAG,EAAE,OAAO,IAAI,GAAG,EAAE,SAAS,IAAI,CAAC;oBAC5C,GAAG,CAAC,GAAG,IAAI,EAAE,MAAM,EAAE,GAAG,CAAC,QAAQ,EAAE,SAAS,EAAE,GAAG,CAAC,MAAM,EAAE,CAAC;oBAC3D,GAAG,CAAC,SAAS,CAAC,MAAM,GAAG,CAAC,IAAI,EAAE,SAAS,EAAE,CAAC;iBAC7C,CAAC,CAAC;gBACH,IAAI,CAAC,IAAI,CAAC,EAAE,KAAK,EAAE,IAAI,EAAE,IAAI,EAAE,CAAC,CAAC;YACrC,CAAC;QACL,CAAC;aAAM,IAAI,GAAG,KAAK,KAAK,EAAE,CAAC;YACvB,MAAM,SAAS,GAAG,gBAAgB,CAAC,IAAI,EAAE,MAAM,CAAC,CAAC;YACjD,MAAM,QAAQ,GAAG,SAAS,CAAC,CAAC,CAAC,eAAe,CAAC,SAAS,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC;YACjE,MAAM,GAAG,GAAG,IAAI,CAAC,kBAAkB,CAAC;YACpC,MAAM,SAAS,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;YACpD,GAAG,CAAC,IAAI,CAAC;gBACL,IAAI,EAAE,QAAQ;gBACd,IAAI,EAAE,QAAQ;gBACd,IAAI,EAAE,GAAG,EAAE,SAAS,IAAI,CAAC;gBACzB,OAAO,EAAE,GAAG,EAAE,OAAO,IAAI,GAAG,EAAE,SAAS,IAAI,CAAC;gBAC5C,GAAG,CAAC,GAAG,IAAI,EAAE,MAAM,EAAE,GAAG,CAAC,QAAQ,EAAE,SAAS,EAAE,GAAG,CAAC,MAAM,EAAE,CAAC;gBAC3D,GAAG,CAAC,SAAS,CAAC,MAAM,GAAG,CAAC,IAAI,EAAE,SAAS,EAAE,CAAC;aAC7C,CAAC,CAAC;YACH,OAAO;QACX,CAAC;QAED,KAAK,MAAM,KAAK,IAAI,IAAI,CAAC,UAAU;YAAE,IAAI,CAAC,KAAK,CAAC,CAAC;IACrD,CAAC;IAED,IAAI,CAAC,IAAI,CAAC,CAAC;IACX,OAAO,KAAK,CAAC;AACjB,CAAC;AAED,SAAS,eAAe,CAAC,MAAe;IACpC,MAAM,SAAS,GAAG,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,KAAK,OAAO,CAAC,CAAC;IAC/D,IAAI,CAAC,SAAS;QAAE,OAAO,MAAM,CAAC;IAC9B,MAAM,KAAK,GAAG,SAAS,CAAC,KAAK,CAAC,KAAK,CAAC,wBAAwB,CAAC,CAAC;IAC9D,OAAO,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC;AACrC,CAAC;AAED,SAAS,gBAAgB,CAAC,MAAkB,EAAE,OAAe;IACzD,KAAK,MAAM,KAAK,IAAI,MAAM,CAAC,UAAU,EAAE,CAAC;QACpC,IAAI,SAAS,CAAC,KAAK,CAAC,IAAI,KAAK,CAAC,OAAO,KAAK,OAAO;YAAE,OAAO,KAAK,CAAC;IACpE,CAAC;IACD,OAAO,IAAI,CAAC;AAChB,CAAC;AAED,SAAS,WAAW,CAAC,IAA4B;IAC7C,IAAI,UAAU,CAAC,IAAI,CAAC;QAAE,OAAO,IAAI,CAAC,KAAK,CAAC;IACxC,IAAI,CAAC,aAAa,CAAC,IAAI,CAAC;QAAE,OAAO,EAAE,CAAC;IACpC,IAAI,GAAG,GAAG,EAAE,CAAC;IACb,KAAK,MAAM,KAAK,IAAI,IAAI,CAAC,UAAU;QAAE,GAAG,IAAI,WAAW,CAAC,KAAK,CAAC,CAAC;IAC/D,OAAO,GAAG,CAAC;AACf,CAAC;AAED,SAAS,SAAS,CAAC,IAA4B;IAC3C,OAAQ,IAAgB,CAAC,OAAO,KAAK,SAAS,IAAK,IAAgB,CAAC,KAAK,KAAK,SAAS,CAAC;AAC5F,CAAC;AAED,SAAS,UAAU,CAAC,IAA4B;IAC5C,OAAQ,IAA8B,CAAC,QAAQ,KAAK,OAAO,CAAC;AAChE,CAAC;AAED,SAAS,aAAa,CAAC,IAAa;IAChC,OAAO,KAAK,CAAC,OAAO,CAAE,IAAiC,CAAC,UAAU,CAAC,CAAC;AACxE,CAAC;AAED,6EAA6E;AAC7E,8EAA8E;AAC9E,8EAA8E;AAC9E,+EAA+E;AAC/E,qCAAqC;AACrC,SAAS,eAAe,CAAC,MAAkB;IACvC,MAAM,GAAG,GAAc,EAAE,CAAC;IAC1B,IAAI,CAAC,aAAa,CAAC,MAAM,CAAC;QAAE,OAAO,GAAG,CAAC;IACvC,KAAK,MAAM,KAAK,IAAI,MAAM,CAAC,UAAU,EAAE,CAAC;QACpC,IAAI,UAAU,CAAC,KAAK,CAAC,EAAE,CAAC;YACpB,MAAM,IAAI,GAAG,KAAK,CAAC,KAAK,CAAC;YACzB,IAAI,IAAI,CAAC,MAAM,KAAK,CAAC;gBAAE,SAAS;YAChC,GAAG,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,OAAO,EAAE,IAAI,EAAE,CAAC,CAAC;YAClC,SAAS;QACb,CAAC;QACD,IAAI,SAAS,CAAC,KAAK,CAAC,EAAE,CAAC;YACnB,GAAG,CAAC,IAAI,CAAC,aAAa,CAAC,KAAK,CAAC,CAAC,CAAC;YAC/B,SAAS;QACb,CAAC;QACD,2DAA2D;IAC/D,CAAC;IACD,OAAO,GAAG,CAAC;AACf,CAAC;AAED,SAAS,aAAa,CAAC,EAAW;IAC9B,MAAM,GAAG,GAAG,EAAE,CAAC,kBAAkB,CAAC;IAClC,MAAM,IAAI,GAA4B;QAClC,IAAI,EAAE,EAAE,CAAC,OAAO;QAChB,IAAI,EAAE,GAAG,EAAE,SAAS,IAAI,CAAC;QACzB,OAAO,EAAE,GAAG,EAAE,OAAO,IAAI,GAAG,EAAE,SAAS,IAAI,CAAC;KAC/C,CAAC;IACF,IAAI,EAAE,CAAC,KAAK,IAAI,EAAE,CAAC,KAAK,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QAClC,MAAM,KAAK,GAA2B,EAAE,CAAC;QACzC,KAAK,MAAM,CAAC,IAAI,EAAE,CAAC,KAAK;YAAE,KAAK,CAAC,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,KAAK,CAAC;QAClD,IAAI,CAAC,KAAK,GAAG,KAAK,CAAC;IACvB,CAAC;IACD,MAAM,QAAQ,GAAG,eAAe,CAAC,EAAE,CAAC,CAAC;IACrC,uEAAuE;IACvE,sBAAsB;IACtB,IAAI,QAAQ,CAAC,MAAM,KAAK,CAAC;WAClB,OAAO,QAAQ,CAAC,CAAC,CAAC,KAAK,QAAQ;WAC/B,QAAQ,CAAC,CAAC,CAAC,KAAK,IAAI;WACnB,QAAQ,CAAC,CAAC,CAAsB,CAAC,IAAI,KAAK,OAAO,EAAE,CAAC;QACxD,IAAI,CAAC,IAAI,GAAI,QAAQ,CAAC,CAAC,CAAsB,CAAC,IAAI,CAAC;IACvD,CAAC;SAAM,IAAI,QAAQ,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QAC7B,IAAI,CAAC,QAAQ,GAAG,QAAQ,CAAC;IAC7B,CAAC;IACD,OAAO,IAAI,CAAC;AAChB,CAAC"}
|
|
1
|
+
{"version":3,"file":"TextHtml.js","sourceRoot":"","sources":["../src/TextHtml.ts"],"names":[],"mappings":"AAAA,OAAO,EACH,WAAW,EACX,sBAAsB,EACtB,sBAAsB,GACzB,MAAM,0BAA0B,CAAC;AAOlC,OAAO,EAAE,KAAK,EAAE,MAAM,QAAQ,CAAC;AAE/B,OAAO,EAAE,SAAS,EAAE,MAAM,gBAAgB,CAAC;AAC3C,OAAO,KAAK,KAAK,MAAM,OAAO,CAAC;AAC/B,OAAO,EAAE,cAAc,EAAE,MAAM,qBAAqB,CAAC;AAyBrD,MAAM,YAAY,GAAG,IAAI,GAAG,CAAC,CAAC,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,CAAC,CAAC,CAAC;AAEnE,MAAM,CAAC,OAAO,OAAO,QAAS,SAAQ,WAAW;IACpC,UAAU,CAAC,OAA4B;QAC5C,MAAM,IAAI,GAAG,OAAO,OAAO,KAAK,QAAQ;YACpC,CAAC,CAAC,OAAO;YACT,CAAC,CAAC,IAAI,WAAW,CAAC,OAAO,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC;QAE/C,MAAM,GAAG,GAAG,KAAK,CAAC,IAAI,EAAE,EAAE,sBAAsB,EAAE,IAAI,EAAE,CAAC,CAAC;QAC1D,MAAM,OAAO,GAAiB,EAAE,CAAC;QACjC,MAAM,SAAS,GAA0C,iBAAiB,CAAC,GAAG,EAAE,OAAO,CAAC,CAAC;QAEzF,wEAAwE;QACxE,qEAAqE;QACrE,oEAAoE;QACpE,MAAM,KAAK,GAAG,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,KAAK,SAAS,IAAI,CAAC,CAAC,KAAK,KAAK,CAAC,CAAC,CAAC;QACzE,IAAI,SAAS,KAAK,IAAI,IAAI,CAAC,KAAK,EAAE,CAAC;YAC/B,OAAO,CAAC,OAAO,CAAC;gBACZ,IAAI,EAAE,SAAS,CAAC,IAAI;gBACpB,IAAI,EAAE,SAAS;gBACf,KAAK,EAAE,CAAC;gBACR,IAAI,EAAE,SAAS,CAAC,IAAI;gBACpB,OAAO,EAAE,SAAS,CAAC,IAAI;aAC1B,CAAC,CAAC;QACP,CAAC;QAED,OAAO,OAAO,CAAC;IACnB,CAAC;IAED,yEAAyE;IACzE,mEAAmE;IACnE,wEAAwE;IACxE,sEAAsE;IACtE,iEAAiE;IACjE,oEAAoE;IACpE,qEAAqE;IACrE,0CAA0C;IACjC,QAAQ,CAAC,OAAuB;QACrC,MAAM,IAAI,GAAG,OAAO,OAAO,KAAK,QAAQ;YACpC,CAAC,CAAC,OAAO;YACT,CAAC,CAAC,IAAI,WAAW,CAAC,OAAO,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC;QAC/C,IAAI,GAAG,CAAC;QACR,IAAI,CAAC;YACD,GAAG,GAAG,KAAK,CAAC,IAAI,EAAE,EAAE,sBAAsB,EAAE,IAAI,EAAE,CAAC,CAAC;QACxD,CAAC;QAAC,MAAM,CAAC;YACL,OAAO,IAAI,CAAC;QAChB,CAAC;QACD,MAAM,IAAI,GAA4B;YAClC,IAAI,EAAE,UAAU;YAChB,IAAI,EAAE,CAAC;YACP,OAAO,EAAE,CAAC;YACV,QAAQ,EAAE,eAAe,CAAC,GAAG,CAAC;SACjC,CAAC;QACF,OAAO,IAAI,CAAC;IAChB,CAAC;IAED,wEAAwE;IACxE,2EAA2E;IAC3E,wEAAwE;IACxE,sEAAsE;IACtE,0EAA0E;IAC1E,sEAAsE;IACtE,oEAAoE;IACpE,4EAA4E;IAC5E,uBAAuB;IACd,OAAO,CAAC,OAAuB;QACpC,MAAM,IAAI,GAAG,OAAO,OAAO,KAAK,QAAQ;YACpC,CAAC,CAAC,OAAO;YACT,CAAC,CAAC,IAAI,WAAW,CAAC,OAAO,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC;QAC/C,OAAO,cAAc,CAAC,IAAI,CAAC,CAAC;IAChC,CAAC;IAED,yEAAyE;IACzE,0EAA0E;IAC1E,yEAAyE;IACzE,yEAAyE;IACzE,6EAA6E;IAC7E,2EAA2E;IAC3E,sEAAsE;IACnD,MAAM,CAAC,OAAuB;QAC7C,MAAM,IAAI,GAAG,OAAO,OAAO,KAAK,QAAQ;YACpC,CAAC,CAAC,OAAO;YACT,CAAC,CAAC,IAAI,WAAW,CAAC,OAAO,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC;QAC/C,OAAO,cAAc,CAAC,IAAI,CAAC,IAAI,EAAE,CAAC;IACtC,CAAC;IAED,wEAAwE;IACxE,0EAA0E;IAC1E,sEAAsE;IACtE,uEAAuE;IACvE,yEAAyE;IACzE,yEAAyE;IAChE,KAAK,CAAC,KAAK,CAChB,OAAuB,EACvB,OAAqB,EACrB,OAAe,EACf,KAAc;QAEd,IAAI,OAAO,KAAK,OAAO,EAAE,CAAC;YACtB,MAAM,IAAI,GAAG,OAAO,OAAO,KAAK,QAAQ;gBACpC,CAAC,CAAC,OAAO;gBACT,CAAC,CAAC,IAAI,WAAW,CAAC,OAAO,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC;YAE/C,2DAA2D;YAC3D,kEAAkE;YAClE,iEAAiE;YACjE,6DAA6D;YAC7D,6DAA6D;YAC7D,+DAA+D;YAC/D,6DAA6D;YAC7D,+DAA+D;YAC/D,qDAAqD;YACrD,IAAI,GAAG,CAAC;YACR,IAAI,CAAC;gBACD,GAAG,GAAG,IAAI,SAAS,EAAE,CAAC,eAAe,CAAC,IAAI,EAAE,UAAU,CAAC,CAAC;YAC5D,CAAC;YAAC,OAAO,KAAK,EAAE,CAAC;gBACb,MAAM,IAAI,sBAAsB,CAAC,EAAE,QAAQ,EAAE,IAAI,CAAC,QAAQ,EAAE,KAAK,EAAE,CAAC,CAAC;YACzE,CAAC;YAED,IAAI,MAA8B,CAAC;YACnC,IAAI,CAAC;gBACD,MAAM,GAAG,KAAK,CAAC,MAAM,CAAC,OAAO,EAAE,GAAsB,CAAC,CAAC;YAC3D,CAAC;YAAC,OAAO,KAAK,EAAE,CAAC;gBACb,MAAM,IAAI,sBAAsB,CAAC,EAAE,OAAO,EAAE,OAAO,EAAE,UAAU,EAAE,OAAO,EAAE,KAAK,EAAE,CAAC,CAAC;YACvF,CAAC;YAED,OAAO,gBAAgB,CAAC,OAAO,EAAE,MAAM,CAAC,CAAC;QAC7C,CAAC;QACD,OAAO,KAAK,CAAC,KAAK,CAAC,OAAO,EAAE,OAAO,EAAE,OAAO,EAAE,KAAK,CAAC,CAAC;IACzD,CAAC;CACJ;AAED,+EAA+E;AAC/E,2EAA2E;AAC3E,8BAA8B;AAC9B,SAAS,gBAAgB,CAAC,OAAe,EAAE,MAA8B;IACrE,IAAI,KAAK,CAAC,OAAO,CAAC,MAAM,CAAC,EAAE,CAAC;QACxB,OAAO,MAAM,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC,EAAc,EAAE;YACtC,MAAM,IAAI,GAAG,QAAQ,CAAC,IAAI,CAAC,CAAC;YAC5B,MAAM,OAAO,GAAG,IAAI,KAAK,SAAS,CAAC,CAAC,CAAC,OAAO,CAAC,IAAI,EAAE,IAAI,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC;YAChE,OAAO;gBACH,OAAO,EAAE,aAAa,CAAC,IAAI,CAAC;gBAC5B,QAAQ,EAAE,MAAM,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,IAAI,OAAO,KAAK,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC,CAAC,SAAS;gBAClE,GAAG,CAAC,IAAI,KAAK,SAAS,IAAI,EAAE,KAAK,EAAE,CAAC,EAAE,IAAI,EAAE,OAAO,EAAE,OAAO,IAAI,IAAI,EAAE,CAAC,EAAE,CAAC;aAC7E,CAAC;QACN,CAAC,CAAC,CAAC;IACP,CAAC;IACD,IAAI,MAAM,KAAK,IAAI,IAAI,MAAM,KAAK,SAAS;QAAE,OAAO,EAAE,CAAC;IACvD,4EAA4E;IAC5E,yDAAyD;IACzD,OAAO,CAAC,EAAE,OAAO,EAAE,OAAO,MAAM,KAAK,QAAQ,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,MAAM,CAAC,MAAM,CAAC,EAAE,CAAC,CAAC;AAC/E,CAAC;AAED,+EAA+E;AAC/E,8EAA8E;AAC9E,iCAAiC;AACjC,SAAS,QAAQ,CAAC,IAAU;IACxB,MAAM,MAAM,GAAI,IAAgC,CAAC,UAAU,CAAC;IAC5D,IAAI,OAAO,MAAM,KAAK,QAAQ,IAAI,MAAM,GAAG,CAAC;QAAE,OAAO,MAAM,CAAC;IAC5D,MAAM,KAAK,GAAI,IAAa,CAAC,YAAY;WACjC,IAAqC,CAAC,UAAU,IAAI,IAAI,CAAC;IACjE,MAAM,EAAE,GAAI,KAAwC,EAAE,UAAU,CAAC;IACjE,OAAO,OAAO,EAAE,KAAK,QAAQ,IAAI,EAAE,GAAG,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,SAAS,CAAC;AAC7D,CAAC;AAED,MAAM,iBAAiB,GAAG,CAAC,CAAC;AAC5B,8EAA8E;AAC9E,4EAA4E;AAC5E,0BAA0B;AAC1B,SAAS,OAAO,CAAC,IAAU,EAAE,SAAiB;IAC1C,MAAM,UAAU,GAAG,aAAa,CAAC,IAAI,CAAC,CAAC;IACvC,OAAO,SAAS,GAAG,CAAC,UAAU,CAAC,KAAK,CAAC,KAAK,CAAC,EAAE,MAAM,IAAI,CAAC,CAAC,CAAC;AAC9D,CAAC;AAED,4EAA4E;AAC5E,4EAA4E;AAC5E,6EAA6E;AAC7E,wEAAwE;AACxE,yCAAyC;AACzC,MAAM,cAAc,GAAG,CAAC,CAAC;AACzB,MAAM,SAAS,GAAG,CAAC,CAAC;AACpB,MAAM,kBAAkB,GAAG,CAAC,CAAC;AAC7B,MAAM,2BAA2B,GAAG,CAAC,CAAC;AACtC,MAAM,YAAY,GAAG,CAAC,CAAC;AACvB,SAAS,aAAa,CAAC,IAAU;IAC7B,MAAM,EAAE,GAAG,IAAI,CAAC,QAAQ,CAAC;IACzB,IAAI,EAAE,KAAK,cAAc;QAAE,OAAQ,IAAa,CAAC,KAAK,CAAC;IACvD,IAAI,EAAE,KAAK,SAAS,IAAI,EAAE,KAAK,kBAAkB;QAAE,OAAQ,IAAa,CAAC,IAAI,CAAC;IAC9E,IAAI,EAAE,KAAK,YAAY;QAAE,OAAQ,IAAgB,CAAC,IAAI,CAAC;IACvD,IAAI,EAAE,KAAK,2BAA2B;QAAE,OAAQ,IAA8B,CAAC,IAAI,CAAC;IACpF,OAAQ,IAA8C,CAAC,QAAQ,EAAE,CAAC;AACtE,CAAC;AAED,4EAA4E;AAC5E,uEAAuE;AACvE,wBAAwB;AACxB,SAAS,iBAAiB,CACtB,IAAgB,EAChB,GAAiB;IAEjB,IAAI,KAAK,GAA0C,IAAI,CAAC;IACxD,sEAAsE;IACtE,yEAAyE;IACzE,iEAAiE;IACjE,MAAM,IAAI,GAA2C,EAAE,CAAC;IAExD,SAAS,IAAI,CAAC,IAA4B;QACtC,IAAI,CAAC,SAAS,CAAC,IAAI,CAAC,EAAE,CAAC;YACnB,IAAI,aAAa,CAAC,IAAI,CAAC,EAAE,CAAC;gBACtB,KAAK,MAAM,KAAK,IAAI,IAAI,CAAC,UAAU;oBAAE,IAAI,CAAC,KAAK,CAAC,CAAC;YACrD,CAAC;YACD,OAAO;QACX,CAAC;QAED,MAAM,GAAG,GAAG,IAAI,CAAC,OAAO,CAAC;QACzB,IAAI,GAAG,KAAK,OAAO,IAAI,KAAK,KAAK,IAAI,EAAE,CAAC;YACpC,MAAM,IAAI,GAAG,WAAW,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,CAAC;YACtC,IAAI,IAAI,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;gBAClB,KAAK,GAAG;oBACJ,IAAI,EAAE,IAAI;oBACV,IAAI,EAAE,IAAI,CAAC,kBAAkB,EAAE,SAAS,IAAI,CAAC;iBAChD,CAAC;YACN,CAAC;QACL,CAAC;aAAM,IAAI,YAAY,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,CAAC;YAC/B,MAAM,IAAI,GAAG,WAAW,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,CAAC;YACtC,MAAM,GAAG,GAAG,IAAI,CAAC,kBAAkB,CAAC;YACpC,IAAI,IAAI,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;gBAClB,MAAM,KAAK,GAAG,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC;gBAC7B,OAAO,IAAI,CAAC,MAAM,GAAG,CAAC,IAAI,IAAI,CAAC,IAAI,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,KAAK,IAAI,KAAK,EAAE,CAAC;oBAC7D,IAAI,CAAC,GAAG,EAAE,CAAC;gBACf,CAAC;gBACD,MAAM,SAAS,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;gBACpD,GAAG,CAAC,IAAI,CAAC;oBACL,IAAI,EAAE,IAAI;oBACV,IAAI,EAAE,SAAS;oBACf,KAAK;oBACL,IAAI,EAAE,GAAG,EAAE,SAAS,IAAI,CAAC;oBACzB,OAAO,EAAE,GAAG,EAAE,OAAO,IAAI,GAAG,EAAE,SAAS,IAAI,CAAC;oBAC5C,GAAG,CAAC,GAAG,IAAI,EAAE,MAAM,EAAE,GAAG,CAAC,QAAQ,EAAE,SAAS,EAAE,GAAG,CAAC,MAAM,EAAE,CAAC;oBAC3D,GAAG,CAAC,SAAS,CAAC,MAAM,GAAG,CAAC,IAAI,EAAE,SAAS,EAAE,CAAC;iBAC7C,CAAC,CAAC;gBACH,IAAI,CAAC,IAAI,CAAC,EAAE,KAAK,EAAE,IAAI,EAAE,IAAI,EAAE,CAAC,CAAC;YACrC,CAAC;QACL,CAAC;aAAM,IAAI,GAAG,KAAK,KAAK,EAAE,CAAC;YACvB,MAAM,SAAS,GAAG,gBAAgB,CAAC,IAAI,EAAE,MAAM,CAAC,CAAC;YACjD,MAAM,QAAQ,GAAG,SAAS,CAAC,CAAC,CAAC,eAAe,CAAC,SAAS,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC;YACjE,MAAM,GAAG,GAAG,IAAI,CAAC,kBAAkB,CAAC;YACpC,MAAM,SAAS,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;YACpD,GAAG,CAAC,IAAI,CAAC;gBACL,IAAI,EAAE,QAAQ;gBACd,IAAI,EAAE,QAAQ;gBACd,IAAI,EAAE,GAAG,EAAE,SAAS,IAAI,CAAC;gBACzB,OAAO,EAAE,GAAG,EAAE,OAAO,IAAI,GAAG,EAAE,SAAS,IAAI,CAAC;gBAC5C,GAAG,CAAC,GAAG,IAAI,EAAE,MAAM,EAAE,GAAG,CAAC,QAAQ,EAAE,SAAS,EAAE,GAAG,CAAC,MAAM,EAAE,CAAC;gBAC3D,GAAG,CAAC,SAAS,CAAC,MAAM,GAAG,CAAC,IAAI,EAAE,SAAS,EAAE,CAAC;aAC7C,CAAC,CAAC;YACH,OAAO;QACX,CAAC;QAED,KAAK,MAAM,KAAK,IAAI,IAAI,CAAC,UAAU;YAAE,IAAI,CAAC,KAAK,CAAC,CAAC;IACrD,CAAC;IAED,IAAI,CAAC,IAAI,CAAC,CAAC;IACX,OAAO,KAAK,CAAC;AACjB,CAAC;AAED,SAAS,eAAe,CAAC,MAAe;IACpC,MAAM,SAAS,GAAG,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,KAAK,OAAO,CAAC,CAAC;IAC/D,IAAI,CAAC,SAAS;QAAE,OAAO,MAAM,CAAC;IAC9B,MAAM,KAAK,GAAG,SAAS,CAAC,KAAK,CAAC,KAAK,CAAC,wBAAwB,CAAC,CAAC;IAC9D,OAAO,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC;AACrC,CAAC;AAED,SAAS,gBAAgB,CAAC,MAAkB,EAAE,OAAe;IACzD,KAAK,MAAM,KAAK,IAAI,MAAM,CAAC,UAAU,EAAE,CAAC;QACpC,IAAI,SAAS,CAAC,KAAK,CAAC,IAAI,KAAK,CAAC,OAAO,KAAK,OAAO;YAAE,OAAO,KAAK,CAAC;IACpE,CAAC;IACD,OAAO,IAAI,CAAC;AAChB,CAAC;AAED,SAAS,WAAW,CAAC,IAA4B;IAC7C,IAAI,UAAU,CAAC,IAAI,CAAC;QAAE,OAAO,IAAI,CAAC,KAAK,CAAC;IACxC,IAAI,CAAC,aAAa,CAAC,IAAI,CAAC;QAAE,OAAO,EAAE,CAAC;IACpC,IAAI,GAAG,GAAG,EAAE,CAAC;IACb,KAAK,MAAM,KAAK,IAAI,IAAI,CAAC,UAAU;QAAE,GAAG,IAAI,WAAW,CAAC,KAAK,CAAC,CAAC;IAC/D,OAAO,GAAG,CAAC;AACf,CAAC;AAED,SAAS,SAAS,CAAC,IAA4B;IAC3C,OAAQ,IAAgB,CAAC,OAAO,KAAK,SAAS,IAAK,IAAgB,CAAC,KAAK,KAAK,SAAS,CAAC;AAC5F,CAAC;AAED,SAAS,UAAU,CAAC,IAA4B;IAC5C,OAAQ,IAA8B,CAAC,QAAQ,KAAK,OAAO,CAAC;AAChE,CAAC;AAED,SAAS,aAAa,CAAC,IAAa;IAChC,OAAO,KAAK,CAAC,OAAO,CAAE,IAAiC,CAAC,UAAU,CAAC,CAAC;AACxE,CAAC;AAED,6EAA6E;AAC7E,8EAA8E;AAC9E,8EAA8E;AAC9E,+EAA+E;AAC/E,qCAAqC;AACrC,SAAS,eAAe,CAAC,MAAkB;IACvC,MAAM,GAAG,GAAc,EAAE,CAAC;IAC1B,IAAI,CAAC,aAAa,CAAC,MAAM,CAAC;QAAE,OAAO,GAAG,CAAC;IACvC,KAAK,MAAM,KAAK,IAAI,MAAM,CAAC,UAAU,EAAE,CAAC;QACpC,IAAI,UAAU,CAAC,KAAK,CAAC,EAAE,CAAC;YACpB,MAAM,IAAI,GAAG,KAAK,CAAC,KAAK,CAAC;YACzB,IAAI,IAAI,CAAC,MAAM,KAAK,CAAC;gBAAE,SAAS;YAChC,GAAG,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,OAAO,EAAE,IAAI,EAAE,CAAC,CAAC;YAClC,SAAS;QACb,CAAC;QACD,IAAI,SAAS,CAAC,KAAK,CAAC,EAAE,CAAC;YACnB,GAAG,CAAC,IAAI,CAAC,aAAa,CAAC,KAAK,CAAC,CAAC,CAAC;YAC/B,SAAS;QACb,CAAC;QACD,2DAA2D;IAC/D,CAAC;IACD,OAAO,GAAG,CAAC;AACf,CAAC;AAED,SAAS,aAAa,CAAC,EAAW;IAC9B,MAAM,GAAG,GAAG,EAAE,CAAC,kBAAkB,CAAC;IAClC,MAAM,IAAI,GAA4B;QAClC,IAAI,EAAE,EAAE,CAAC,OAAO;QAChB,IAAI,EAAE,GAAG,EAAE,SAAS,IAAI,CAAC;QACzB,OAAO,EAAE,GAAG,EAAE,OAAO,IAAI,GAAG,EAAE,SAAS,IAAI,CAAC;KAC/C,CAAC;IACF,IAAI,EAAE,CAAC,KAAK,IAAI,EAAE,CAAC,KAAK,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QAClC,MAAM,KAAK,GAA2B,EAAE,CAAC;QACzC,KAAK,MAAM,CAAC,IAAI,EAAE,CAAC,KAAK;YAAE,KAAK,CAAC,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,KAAK,CAAC;QAClD,IAAI,CAAC,KAAK,GAAG,KAAK,CAAC;IACvB,CAAC;IACD,MAAM,QAAQ,GAAG,eAAe,CAAC,EAAE,CAAC,CAAC;IACrC,uEAAuE;IACvE,sBAAsB;IACtB,IAAI,QAAQ,CAAC,MAAM,KAAK,CAAC;WAClB,OAAO,QAAQ,CAAC,CAAC,CAAC,KAAK,QAAQ;WAC/B,QAAQ,CAAC,CAAC,CAAC,KAAK,IAAI;WACnB,QAAQ,CAAC,CAAC,CAAsB,CAAC,IAAI,KAAK,OAAO,EAAE,CAAC;QACxD,IAAI,CAAC,IAAI,GAAI,QAAQ,CAAC,CAAC,CAAsB,CAAC,IAAI,CAAC;IACvD,CAAC;SAAM,IAAI,QAAQ,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QAC7B,IAAI,CAAC,QAAQ,GAAG,QAAQ,CAAC;IAC7B,CAAC;IACD,OAAO,IAAI,CAAC;AAChB,CAAC"}
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"htmlToMarkdown.d.ts","sourceRoot":"","sources":["../src/htmlToMarkdown.ts"],"names":[],"mappings":"
|
|
1
|
+
{"version":3,"file":"htmlToMarkdown.d.ts","sourceRoot":"","sources":["../src/htmlToMarkdown.ts"],"names":[],"mappings":"AA0BA,wBAAgB,cAAc,CAAC,IAAI,EAAE,MAAM,GAAG,MAAM,GAAG,SAAS,CA8B/D"}
|
package/dist/htmlToMarkdown.js
CHANGED
|
@@ -10,9 +10,12 @@ import TurndownService from "turndown";
|
|
|
10
10
|
// Pipeline (SPEC §18): main-content extraction via @mozilla/readability over a
|
|
11
11
|
// linkedom DOM, then HTML→markdown via turndown. Readability strips nav, ads,
|
|
12
12
|
// and chrome and returns the article body; turndown renders it as markdown.
|
|
13
|
-
// When Readability finds no article (
|
|
14
|
-
//
|
|
15
|
-
//
|
|
13
|
+
// When Readability finds no article (app shells, forms, fragments, very short
|
|
14
|
+
// HTML) the readable projection is genuinely ABSENT — return undefined. We do
|
|
15
|
+
// NOT degrade to markdown-of-the-whole-body: that turned a 2MB JS-shell page
|
|
16
|
+
// (e.g. youtube.com/watch) into 2MB of "markdown" and fed it to the embedder,
|
|
17
|
+
// wedging the daemon (#412). Empty content when the content is empty; raw HTML
|
|
18
|
+
// has exactly one destiny (article extraction) and no escape hatch downstream.
|
|
16
19
|
const turndown = new TurndownService({
|
|
17
20
|
headingStyle: "atx",
|
|
18
21
|
bulletListMarker: "-",
|
|
@@ -35,19 +38,32 @@ export function htmlToMarkdown(html) {
|
|
|
35
38
|
// Readability is best-effort denoising — its failure is not ours.
|
|
36
39
|
articleHtml = undefined;
|
|
37
40
|
}
|
|
41
|
+
// No article (app shells, forms, snippets) → project the READABLE TEXT, not
|
|
42
|
+
// the raw DOM: strip script/style/noscript/template first, then render what
|
|
43
|
+
// remains. A 2MB youtube.com/watch page is dominated by <script> JSON blobs
|
|
44
|
+
// — stripped, its structural markup renders to almost nothing (#412 freeze
|
|
45
|
+
// cured). A <p> of search-result prose keeps its text. The axis is
|
|
46
|
+
// readable-text-vs-noise, NOT size: a genuinely long article is legitimately
|
|
47
|
+
// large and stays whole. Never the raw markup, never a size cap.
|
|
38
48
|
if (articleHtml === null || articleHtml === undefined || articleHtml.trim().length === 0) {
|
|
39
49
|
articleHtml = readableBody(document);
|
|
40
50
|
}
|
|
41
51
|
const markdown = turndown.turndown(articleHtml).trim();
|
|
42
52
|
return markdown.length === 0 ? undefined : markdown;
|
|
43
53
|
}
|
|
44
|
-
//
|
|
45
|
-
// content
|
|
46
|
-
//
|
|
54
|
+
// The readable body with noise removed: script/style/noscript/template carry no
|
|
55
|
+
// reading content but dominate app-shell byte weight — dropping them is what
|
|
56
|
+
// keeps turndown (and the downstream embedder) proportional to real text, not
|
|
57
|
+
// raw page size. Falls to the document element for unwrapped fragments.
|
|
47
58
|
function readableBody(document) {
|
|
59
|
+
const el = document.documentElement;
|
|
60
|
+
if (el === null)
|
|
61
|
+
return "";
|
|
62
|
+
// Strip from the whole element so noise is gone whether the content sits in
|
|
63
|
+
// <body> or loose under <html> (linkedom parks bare fragments outside body).
|
|
64
|
+
for (const node of el.querySelectorAll("script, style, noscript, template"))
|
|
65
|
+
node.remove();
|
|
48
66
|
const bodyHtml = document.body?.innerHTML ?? "";
|
|
49
|
-
|
|
50
|
-
return bodyHtml;
|
|
51
|
-
return document.documentElement?.innerHTML ?? "";
|
|
67
|
+
return bodyHtml.trim().length > 0 ? bodyHtml : el.innerHTML;
|
|
52
68
|
}
|
|
53
69
|
//# sourceMappingURL=htmlToMarkdown.js.map
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"htmlToMarkdown.js","sourceRoot":"","sources":["../src/htmlToMarkdown.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,SAAS,EAAE,MAAM,UAAU,CAAC;AACrC,OAAO,EAAE,WAAW,EAAE,MAAM,sBAAsB,CAAC;AACnD,OAAO,eAAe,MAAM,UAAU,CAAC;AAEvC,0EAA0E;AAC1E,wEAAwE;AACxE,6EAA6E;AAC7E,6EAA6E;AAC7E,mDAAmD;AACnD,EAAE;AACF,+EAA+E;AAC/E,8EAA8E;AAC9E,4EAA4E;AAC5E,8EAA8E;AAC9E,6EAA6E;AAC7E,
|
|
1
|
+
{"version":3,"file":"htmlToMarkdown.js","sourceRoot":"","sources":["../src/htmlToMarkdown.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,SAAS,EAAE,MAAM,UAAU,CAAC;AACrC,OAAO,EAAE,WAAW,EAAE,MAAM,sBAAsB,CAAC;AACnD,OAAO,eAAe,MAAM,UAAU,CAAC;AAEvC,0EAA0E;AAC1E,wEAAwE;AACxE,6EAA6E;AAC7E,6EAA6E;AAC7E,mDAAmD;AACnD,EAAE;AACF,+EAA+E;AAC/E,8EAA8E;AAC9E,4EAA4E;AAC5E,8EAA8E;AAC9E,8EAA8E;AAC9E,6EAA6E;AAC7E,8EAA8E;AAC9E,+EAA+E;AAC/E,+EAA+E;AAE/E,MAAM,QAAQ,GAAG,IAAI,eAAe,CAAC;IACjC,YAAY,EAAE,KAAK;IACnB,gBAAgB,EAAE,GAAG;IACrB,cAAc,EAAE,QAAQ;CAC3B,CAAC,CAAC;AAEH,MAAM,UAAU,cAAc,CAAC,IAAY;IACvC,IAAI,IAAI,CAAC,IAAI,EAAE,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,SAAS,CAAC;IAE/C,MAAM,EAAE,QAAQ,EAAE,GAAG,SAAS,CAAC,IAAI,CAAC,CAAC;IACrC,yEAAyE;IACzE,IAAI,QAAQ,CAAC,eAAe,KAAK,IAAI;QAAE,OAAO,SAAS,CAAC;IAExD,IAAI,WAAsC,CAAC;IAC3C,IAAI,CAAC;QACD,oEAAoE;QACpE,qDAAqD;QACrD,WAAW,GAAG,IAAI,WAAW,CAAC,QAAQ,CAAC,SAAS,CAAC,IAAI,CAAa,CAAC,CAAC,KAAK,EAAE,EAAE,OAAO,CAAC;IACzF,CAAC;IAAC,MAAM,CAAC;QACL,kEAAkE;QAClE,WAAW,GAAG,SAAS,CAAC;IAC5B,CAAC;IAED,4EAA4E;IAC5E,4EAA4E;IAC5E,4EAA4E;IAC5E,2EAA2E;IAC3E,mEAAmE;IACnE,6EAA6E;IAC7E,iEAAiE;IACjE,IAAI,WAAW,KAAK,IAAI,IAAI,WAAW,KAAK,SAAS,IAAI,WAAW,CAAC,IAAI,EAAE,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QACvF,WAAW,GAAG,YAAY,CAAC,QAAQ,CAAC,CAAC;IACzC,CAAC;IAED,MAAM,QAAQ,GAAG,QAAQ,CAAC,QAAQ,CAAC,WAAW,CAAC,CAAC,IAAI,EAAE,CAAC;IACvD,OAAO,QAAQ,CAAC,MAAM,KAAK,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,QAAQ,CAAC;AACxD,CAAC;AAED,gFAAgF;AAChF,6EAA6E;AAC7E,8EAA8E;AAC9E,wEAAwE;AACxE,SAAS,YAAY,CAAC,QAAkB;IACpC,MAAM,EAAE,GAAG,QAAQ,CAAC,eAAe,CAAC;IACpC,IAAI,EAAE,KAAK,IAAI;QAAE,OAAO,EAAE,CAAC;IAC3B,4EAA4E;IAC5E,6EAA6E;IAC7E,KAAK,MAAM,IAAI,IAAI,EAAE,CAAC,gBAAgB,CAAC,mCAAmC,CAAC;QAAE,IAAI,CAAC,MAAM,EAAE,CAAC;IAC3F,MAAM,QAAQ,GAAG,QAAQ,CAAC,IAAI,EAAE,SAAS,IAAI,EAAE,CAAC;IAChD,OAAO,QAAQ,CAAC,IAAI,EAAE,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,EAAE,CAAC,SAAS,CAAC;AAChE,CAAC"}
|
package/package.json
CHANGED
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
{
|
|
2
2
|
"name": "@plurnk/plurnk-mimetypes-text-html",
|
|
3
|
-
"version": "1.0.
|
|
3
|
+
"version": "1.0.4",
|
|
4
4
|
"description": "text/html and application/xhtml+xml mimetype handler for plurnk-service. Structural extraction via parse5; readable markdown (content channel) via Readability + turndown.",
|
|
5
5
|
"type": "module",
|
|
6
6
|
"license": "MIT",
|
|
@@ -64,6 +64,6 @@
|
|
|
64
64
|
"@types/turndown": "^5.0.6"
|
|
65
65
|
},
|
|
66
66
|
"peerDependencies": {
|
|
67
|
-
"@plurnk/plurnk-mimetypes": "1.0.
|
|
67
|
+
"@plurnk/plurnk-mimetypes": "1.0.4"
|
|
68
68
|
}
|
|
69
69
|
}
|