transcript-parser

transcript-parser

• Description
• Usage
• Config
• Documentation
  • .parseStream()
  • .parseOneSync()
  • .parseOne()
  • .resolveAliasesSync()
  • .resolveAliases()
• Example

Description

Parses plaintext speech/debate/radio transcripts into JavaScript objects. It is still in early development. Pull requests are welcome.

Tests can be run with npm test and a benchmark can be run with npm run benchmark. For a full coverage report using Istanbul, run npm run travis-test.

Tested for Node.js >= v4.4.6

Usage

npm install transcript-parser

'use strict';
const fs = require('fs');
const TranscriptParser = require('transcript-parser');
const tp = new TranscriptParser();

//Synchronous example
const parsed = tp.parseOneSync(fs.readFileSync('transcript.txt', 'utf8'));
console.log(parsed);

//Asynchronous example
fs.readFile('transcript.txt', (err, data) => {
  if(err) return console.error('Error:', err);
  tp.parseOne(data, (err, parsed) => {
    if(err) return console.error('Error:', err);
    console.log(parsed);
  }));
});

//Stream example
const stream = fs.createReadStream('transcript.txt', 'utf8');
tp.parseStream(stream, (err, parsed) => {
  if(err) return console.error('Error:', err);
  console.log(parsed);
});

Config

The constructor for TranscriptParser accepts a settings object.

• removeActions
  • default: true
  • Specifies if the parser should remove actions (e.g. (APPLAUSE)).
• removeAnnotations
  • default: true
  • Specifies if the parser should remove annotations (surrounded by []).
• removeTimestamps
  • default: true
  • True if removeAnnotations is true
  • Specifies if the parser should remove timestamps (in the [##:##:##] format).
• removeUnknownSpeakers
  • default: false
  • Specifies if the parser should remove lines that have no associated speaker.
  • If true, lines that have no associated speaker will be stored under the key none.
• blacklist
  • default: []
  • A list of speakers (as strings) that the parser should ignore.
• aliases
  • default: {}
  • A object with the real name as the key and an Array of the aliases' regular expressions as the value.
  • Example: { "Mr. Robot": [ /[A-Z\ ]*SLATER[A-Z\ ]*/ ] }
    • Renames all speakers who match the regex to "Mr. Robot".

Settings can be changed after object creation by changing the corresponding properties of tp.settings, where tp is an instance of TranscriptParser.

Documentation

.parseStream()

The parseStream() method parses a Stream and returns an object representing it.

This is the preferred method for parsing streams asynchronously as it doesn't have to load the entire transcript into memory (unlike parseOne()).

Syntax

tp.parseOneSync(stream, callback)

Parameters

• stream
  • The Readable stream to read.
• callback(err, data)
  • A callback to be executed on function completion or error.

.parseOneSync()

The parseOneSync() method parses a string and returns an object representing it.

Syntax

tp.parseOneSync(transcript)

Parameters

• transcript
  • The transcript, as a string.

.parseOne()

The parseOne() method parses a string and returns an object representing it.

Syntax

tp.parseOne(transcript, callback)

Parameters

• transcript
  • The transcript, as a string.
• callback(err, data)
  • A callback to be exectuted on function completion or error.

.resolveAliasesSync()

The resolveAliasesSync() method resolves all aliases specified in the configuration passed to the TranscriptParser's constructor (see above).

Renames the names in the order list to match the new names in the transcript. Note that there is a signifigant performance penalty, so don't use this method unless you need it.

Syntax

tp.resolveAliasesSync(data)

Parameters

• data
  • The transcript object after being parsed.

.resolveAliases()

The resolveAliases() method resolves all aliases specified in the configuration passed to the TranscriptParser's constructor (see above).

Renames the names in the order list to match the new names in the transcript. Note that there is a signifigant performance penalty, so don't use this method unless you need it.

Syntax

tp.resolveAliases(data, callback)

Parameters

• data
  • The transcript object after being parsed.
• callback(err, resolved)
  • A callback to be executed on function completion or error.

Example

Input

A: I like Node.js.
A: I also like C#.
B: I like Node.js too!
A: I especially like the Node Package Manager.

Output

{
  speaker: {
    A: [
      'I like Node.js.',
      'I also like C#.',
      'I especially like the Node Package Manager.'
    ],
    B: ['I like Node.js too!']
  },
  order: ['A', 'A', 'B', 'A']
}