@mdgate/subtitle 0.1.0

package/README.md

@@ -0,0 +1,26 @@
+# @mdgate/subtitle
+
+Convert SRT and WebVTT subtitle files to Markdown. Outputs GitHub-Flavored Markdown. Works in Node,
+Edge, and browsers. No native addons.
+
+Handles: `.srt`, `.vtt`, `.webvtt`
+
+## Usage
+
+```ts
+import { toMarkdown } from '@mdgate/subtitle';
+
+const markdown = await toMarkdown(bytes);
+```
+
+Compose with other converters:
+
+```ts
+import { create } from '@mdgate/core';
+import { subtitle } from '@mdgate/subtitle';
+
+const convert = create([subtitle()]);
+```
+
+Part of [mdgate converters](https://github.com/mdgate/converters); install
+`@mdgate/converters` for every format at once.

package/dist/index.d.ts

@@ -0,0 +1,2 @@
+export { subtitle } from './subtitle.js';
+export { toMarkdown } from './to-markdown.js';

package/dist/index.js

@@ -0,0 +1,3 @@
+import{ConvertError as d}from"@mdgate/core";import{documentToMarkdown as H}from"@mdgate/document";import{fileExtension as M}from"@mdgate/utils";import{emptyDocument as T,PLAIN as N}from"@mdgate/document";import{cleanText as w,collapseWs as C,decode as c,lines as S,trim as o}from"@mdgate/utils";var U={bold:!1,italic:!0,strike:!1,code:!1},k=/^\s*(?:(\d{1,3}):)?(\d{1,2}):(\d{2})[,.](\d{1,3})\s*-->\s*(?:(\d{1,3}):)?(\d{1,2}):(\d{2})[,.](\d{1,3})\b/,I=/\{[^}]*\}/g,E=/\\an\d+/gi,W=/<[^>]*>/g,v=/&(?:amp|lt|gt|quot|apos|nbsp|lrm|rlm|#\d+|#x[\da-f]+);/gi;function x(n){let e=R(L(n)).replace(/\r\n/g,`
+`).replace(/\r/g,`
+`),r=S(e),i=T(),t=0;if(B(r[0])){t=1;while(t<r.length&&o(r[t]).length>0){if(f(r[t])!==void 0)break;if(t+1<r.length&&f(r[t+1])!==void 0)break;t+=1}}while(t<r.length){let u=r[t];if(o(u).length===0){t+=1;continue}if($(u)){t+=1;while(t<r.length&&o(r[t]).length>0)t+=1;continue}let a=f(u);if(a===void 0&&t+1<r.length){if(a=f(r[t+1]),a!==void 0)t+=1}if(a===void 0){t+=1;while(t<r.length&&o(r[t]).length>0)t+=1;continue}t+=1;let g=[];while(t<r.length&&o(r[t]).length>0){if(f(r[t])!==void 0)break;if(t+1<r.length&&f(r[t+1])!==void 0)break;g.push(r[t]),t+=1}let m=D(g.join(" "));if(m.length===0)continue;i.blocks.push({type:"paragraph",inlines:F(a,m)})}return i}function F(n,e){return[{type:"text",text:`[${n}]`,style:U},{type:"text",text:` ${e}`,style:N}]}function f(n){let e=k.exec(n);if(e===null)return;let r=e[1]!==void 0?Number(e[1]):0,i=Number(e[2]),t=Number(e[3]),u=Number(e[4]);if(!Number.isFinite(r+i+t+u))return;return`${l(r,2)}:${l(i,2)}:${l(t,2)}.${l(u,3)}`}function B(n){if(n===void 0)return!1;return/^WEBVTT(?:$|[\s\uFEFF])/.test(n)}function $(n){let e=o(n);return s(e,"NOTE")||s(e,"STYLE")||s(e,"REGION")}function s(n,e){if(n.length<e.length)return!1;if(n.slice(0,e.length).toUpperCase()!==e)return!1;return n.length===e.length||P(n.charCodeAt(e.length))}function P(n){return n===9||n===10||n===13||n===32}function D(n){let e=n.replace(I,"").replace(E,"").replace(W,"");return o(C(w(G(e))))}function G(n){return n.replace(v,(e)=>{let r=e.toLowerCase();switch(r){case"&amp;":return"&";case"&lt;":return"<";case"&gt;":return">";case"&quot;":return'"';case"&apos;":return"'";case"&nbsp;":return" ";case"&lrm;":case"&rlm;":return"";default:break}if(r.startsWith("&#x")){let i=Number.parseInt(r.slice(3,-1),16);return Number.isFinite(i)?String.fromCodePoint(i):""}if(r.startsWith("&#")){let i=Number.parseInt(r.slice(2,-1),10);return Number.isFinite(i)?String.fromCodePoint(i):""}return""})}function L(n){if(n.length>=2&&n[0]===255&&n[1]===254)return c(n,"utf-16le");if(n.length>=2&&n[0]===254&&n[1]===255)return c(n,"utf-16be");let e=n.length>=3&&n[0]===239&&n[1]===187&&n[2]===191?n.subarray(3):n;try{return new TextDecoder("utf-8",{fatal:!0}).decode(e)}catch{return c(e,"windows-1252")}}function R(n){return n.charCodeAt(0)===65279?n.slice(1):n}function l(n,e){return n.toString().padStart(e,"0")}var O=new Set(["srt","vtt","webvtt"]);function p(){return{id:"subtitle",sniff(n,e){if(_(n))return 2;if(e?.path!==void 0&&O.has(M(e.path)??""))return 1;return 0},convert(n){return q(n),{markdown:H(x(n))}}}}function _(n){let e=h(n);if(!A(n,"WEBVTT",e))return!1;return!K(n[e+6])}function q(n){if(V(n))throw d.unsupported("pdf");if(X(n))throw d.unsupported("ole");if(Y(n))throw d.unsupported("zip")}function V(n){let e=z(n);return A(n,"%PDF-",e)}function X(n){return n.length>=8&&n[0]===208&&n[1]===207&&n[2]===17&&n[3]===224&&n[4]===161&&n[5]===177&&n[6]===26&&n[7]===225}function Y(n){if(n.length<4||n[0]!==80||n[1]!==75)return!1;return n[2]===3&&n[3]===4||n[2]===5&&n[3]===6||n[2]===7&&n[3]===8}function h(n){if(n.length>=3&&n[0]===239&&n[1]===187&&n[2]===191)return 3;return 0}function z(n){let e=h(n);while(e<n.length){let r=n[e];if(r!==9&&r!==10&&r!==13&&r!==32)break;e+=1}return e}function A(n,e,r){if(r+e.length>n.length)return!1;for(let i=0;i<e.length;i+=1)if(n[r+i]!==e.charCodeAt(i))return!1;return!0}function K(n){if(n===void 0)return!1;return n>=48&&n<=57||n>=65&&n<=90||n>=97&&n<=122||n===45||n===95||n===58}import{create as Z}from"@mdgate/core";var b=Z([p()]);function j(n,e){return b(n,e)}export{j as toMarkdown,p as subtitle};

package/dist/subtitle.d.ts

@@ -0,0 +1,2 @@
+import type { Converter } from '@mdgate/core';
+export declare function subtitle(): Converter;

package/dist/to-markdown.d.ts

@@ -0,0 +1,6 @@
+import { type ConvertHint } from '@mdgate/core';
+/**
+ * Convert a single file's bytes to Markdown with the subtitle converter.
+ * `hint.path` is only a sniff hint; it is never read from disk.
+ */
+export declare function toMarkdown(bytes: Uint8Array, hint?: ConvertHint): Promise<string>;

package/package.json

@@ -0,0 +1,43 @@
+{
+  "name": "@mdgate/subtitle",
+  "version": "0.1.0",
+  "description": "mdgate SRT and WebVTT converter",
+  "type": "module",
+  "license": "MIT",
+  "sideEffects": false,
+  "exports": {
+    ".": {
+      "types": "./dist/index.d.ts",
+      "import": "./dist/index.js"
+    }
+  },
+  "main": "./dist/index.js",
+  "types": "./dist/index.d.ts",
+  "files": [
+    "dist/**/*.js",
+    "dist/**/*.d.ts"
+  ],
+  "scripts": {
+    "build": "bun ../../scripts/build-package.ts",
+    "prepublishOnly": "bun run build"
+  },
+  "dependencies": {
+    "@mdgate/core": "0.1.3",
+    "@mdgate/document": "0.1.0",
+    "@mdgate/utils": "0.1.2"
+  },
+  "publishConfig": {
+    "access": "public"
+  },
+  "engines": {
+    "node": ">=20"
+  },
+  "keywords": [
+    "mdgate",
+    "markdown",
+    "srt",
+    "vtt",
+    "webvtt",
+    "subtitle"
+  ]
+}