From ebbb8ddb8b77234a9800ea5ab92bdd3a208ac7bb Mon Sep 17 00:00:00 2001
From: Matt Knight <matthew.knight@gmail.com>
Date: Wed, 13 Mar 2019 06:55:57 -0700
Subject: [PATCH] Support parsing metadata from input

---
 README.md           | 42 ++++++++++++++++++++++++++
 lib/parser.js       | 27 +++++++++++++++--
 test/parser.test.js | 73 +++++++++++++++++++++++++++++++++++++++++++++
 3 files changed, 139 insertions(+), 3 deletions(-)

diff --git a/README.md b/README.md
index 1052c2e..6c645b8 100644
--- a/README.md
+++ b/README.md
@@ -89,6 +89,48 @@ For the above example we'd get:
 }
 ```
 
+### Metadata
+
+Some WebVTT strings may also contain lines of metadata after the initial `WEBVTT` line, for example:
+
+```
+WEBVTT
+Kind: captions
+Language: en
+
+00:00:00.000 --> 00:00:01.000
+Hello world!
+```
+
+By passing `{ meta: true }` to the `parse` method, these metadata will be returned as an object called `meta`. For example, parsing the above example:
+
+```js
+parse(webvtt, { meta: true });
+```
+
+would return the following:
+
+```json
+{
+   "valid":true,
+   "meta":{
+      "Kind": "captions",
+      "Language": "en"
+   },
+   "cues":[
+      {
+         "identifier":"",
+         "start":0,
+         "end":1,
+         "text":"Hello world!",
+         "styles":""
+      }
+   ]
+}
+```
+
+If no metadata is available, `meta` will be set to `null` in the result if the option is specified.
+
 ### Compiling
 
 Compiles JSON from the above format back into a WebVTT string.
diff --git a/lib/parser.js b/lib/parser.js
index 4370094..defa214 100644
--- a/lib/parser.js
+++ b/lib/parser.js
@@ -13,7 +13,12 @@ ParserError.prototype = Object.create(Error.prototype);
 
 const TIMESTAMP_REGEXP = /([0-9]{1,2})?:?([0-9]{2}):([0-9]{2}\.[0-9]{3})/;
 
-function parse (input) {
+function parse (input, options) {
+  if (!options) options = {};
+
+  options = {
+    meta: options.meta === true
+  };
 
   if (typeof input !== 'string') {
     throw new ParserError('Input must be a string');
@@ -46,13 +51,29 @@ function parse (input) {
     return { valid: true };
   }
 
-  if (headerParts.length > 1 && headerParts[1] !== '') {
+  if (!options.meta && headerParts.length > 1 && headerParts[1] !== '') {
     throw new ParserError('Missing blank line after signature');
   }
 
   const cues = parseCues(parts);
+  const meta = options.meta ? parseMeta(headerParts) : null;
+
+  const result = { valid: true, cues };
+
+  if (options.meta) {
+    result.meta = meta;
+  }
+
+  return result;
+}
 
-  return { valid: true, cues };
+function parseMeta (headerParts) {
+  const meta = {};
+  headerParts.slice(1).forEach(header => {
+    const [key, value] = header.split(':').map(t => t.trim());
+    meta[key] = value;
+  });
+  return Object.keys(meta).length > 0 ? meta : null;
 }
 
 function parseCues (cues) {
diff --git a/test/parser.test.js b/test/parser.test.js
index da81bcf..ad5d307 100644
--- a/test/parser.test.js
+++ b/test/parser.test.js
@@ -220,4 +220,77 @@ Chapter 17`;
 
     parse(input).cues.should.have.length(3);
   });
+
+  it('should not return meta by default', () => {
+    const input = `WEBVTT
+
+1
+00:00.000 --> 00:00.001`;
+
+    parse(input).should.have.property('valid').be.true;
+    parse(input).should.not.have.property('meta');
+  });
+
+  it('should accept an options object', () => {
+    const input = `WEBVTT
+
+1
+00:00.000 --> 00:00.001`;
+    const options = { meta: true };
+
+    parse(input, options).cues[0].start.should.equal(0);
+    parse(input, options).cues[0].end.should.equal(0.001);
+  });
+
+  it('should fail if metadata exists but the meta option is not set', () => {
+    const input = `WEBVTT
+Kind: captions
+Language: en
+
+1
+00:00.000 --> 00:00.001`;
+    const options = { };
+
+    (() => { parse(input, options); })
+      .should.throw(parserError, /Missing blank line after signature/);
+  });
+
+  it('should fail if metadata exists but the meta option is false', () => {
+    const input = `WEBVTT
+Kind: captions
+Language: en
+
+1
+00:00.000 --> 00:00.001`;
+    const options = { meta: false };
+
+    (() => { parse(input, options); })
+      .should.throw(parserError, /Missing blank line after signature/);
+  });
+
+  it('should return meta if meta option is true', () => {
+    const input = `WEBVTT
+Kind: captions
+Language: en
+
+1
+00:00.000 --> 00:00.001`;
+    const options = { meta: true };
+
+    parse(input, options).should.have.property('valid').be.true;
+    parse(input, options).should.have.property('meta').be.deep.equal(
+      { Kind: 'captions', Language: 'en' }
+    );
+  });
+
+  it('should return null if meta option is true but no meta', () => {
+    const input = `WEBVTT
+
+1
+00:00.000 --> 00:00.001`;
+    const options = { meta: true };
+
+    parse(input, options).should.have.property('valid').be.true;
+    parse(input, options).should.have.property('meta').be.equal(null);
+  });
 });