rdelaage
/
castopod
mirror of https://code.castopod.org/adaures/castopod.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

refactor: replace transcript parser by mantas-done/subtitles library

This commit is contained in:
Yassine Doghri 2024-03-06 18:45:13 +00:00
parent cf4808ab63
commit ef0e641f69
7 changed files with 1040 additions and 611 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

1
composer.json
View File

@ -20,6 +20,7 @@
"james-heinrich/getid3": "^2.0.0-beta5",
"league/commonmark": "^2.4.2",
"league/html-to-markdown": "5.1.1",
"mantas-done/subtitles": "^1.0.21",
"melbahja/seo": "^v2.1.1",
"michalsn/codeigniter4-uuid": "v1.0.2",
"mpratt/embera": "^2.0.36",

1137
composer.lock generated
View File

File diff suppressed because it is too large Load Diff

73
modules/Media/CastopodSubtitles/CastopodSubtitles.php Normal file
View File

@ -0,0 +1,73 @@
<?php
/**
* @copyright 2024 Ad Aures
* @license https://www.gnu.org/licenses/agpl-3.0.en.html AGPL3
* @link https://castopod.org/
*/
namespace Modules\Media\CastopodSubtitles;
use Done\Subtitles\Code\Converters\ConverterContract;
use Done\Subtitles\Code\Converters\SrtConverter;
use Done\Subtitles\Code\Converters\VttConverter;
use Done\Subtitles\Subtitles;
use Exception;
class CastopodSubtitles extends Subtitles
{
/**
* @var array<array{extension:string,format:string,name:string,class:class-string<ConverterContract>}>
*/
public static $formats = [
[
'extension' => 'vtt',
'format' => 'vtt',
'name' => 'WebVTT',
'class' => VttConverter::class,
],
[
'extension' => 'srt',
'format' => 'srt',
'name' => 'SubRip',
'class' => SrtConverter::class,
],
[
'extension' => 'json',
'format' => 'json',
'name' => 'JSON',
'class' => JSONConverter::class,
],
];
/**
* @param string $format
* @param array<mixed> $options
* @return string
*/
public function content($format, $options = [])
{
/** @var ConverterContract $converter */
$converter = $this->getConverterByFormat($format);
return $converter->internalFormatToFileContent($this->internal_format, $options);
}
/**
* @param string $format
* @return ConverterContract
*/
private function getConverterByFormat($format)
{
foreach (self::$formats as $row) {
if ($row['format'] === $format) {
$full_class_name = $row['class'];
/** @var ConverterContract $converter */
$converter = new $full_class_name();
return $converter;
}
}
throw new Exception("Can't find suitable converter, for format: {$format}");
}
}

67
modules/Media/CastopodSubtitles/JSONConverter.php Normal file
View File

@ -0,0 +1,67 @@
<?php
/**
* @copyright 2024 Ad Aures
* @license https://www.gnu.org/licenses/agpl-3.0.en.html AGPL3
* @link https://castopod.org/
*/
namespace Modules\Media\CastopodSubtitles;
use Done\Subtitles\Code\Converters\ConverterContract;
class JSONConverter implements ConverterContract
{
/**
* @param string $file_content
* @return bool
*/
public function canParseFileContent($file_content)
{
json_decode($file_content);
return json_last_error() === JSON_ERROR_NONE;
}
/**
* @param string $file_content
* @param string $original_file_content
* @return array<array{start:int,end:int,lines:array<string>}>
*/
public function fileContentToInternalFormat($file_content, $original_file_content)
{
/** @var array<array{startTime:int,endTime:int,text:string}> $jsonTranscriptArray */
$jsonTranscriptArray = json_decode($file_content, true);
$internalFormat = [];
foreach ($jsonTranscriptArray as $segment) {
$internalFormat[] = [
'start' => $segment['startTime'],
'end' => $segment['endTime'],
'lines' => explode(PHP_EOL, $segment['text']),
];
}
return $internalFormat;
}
/**
* @param array<array{start:int,end:int,lines:array<string>}> $internal_format
* @param array<mixed> $output_settings
*/
public function internalFormatToFileContent(array $internal_format, array $output_settings): string
{
/** @var array<array{number:int,startTime:int,endTime:int,text:string}> $jsonTranscriptArray */
$jsonTranscriptArray = [];
foreach ($internal_format as $key => $value) {
$jsonTranscriptArray[] = [
'number' => $key,
'startTime' => $value['start'],
'endTime' => $value['end'],
'text' => implode(PHP_EOL, $value['lines']),
];
}
return (string) json_encode($jsonTranscriptArray);
}
}

91
modules/Media/Entities/Transcript.php
View File

@ -10,98 +10,7 @@ declare(strict_types=1);
namespace Modules\Media\Entities;
use CodeIgniter\Files\File;
use Exception;
use Modules\Media\TranscriptParser;
class Transcript extends BaseMedia
{
public ?string $json_key = null;
public ?string $json_url = null;
protected string $type = 'transcript';
public function initFileProperties(): void
{
parent::initFileProperties();
if ($this->file_metadata !== null && array_key_exists('json_key', $this->file_metadata)) {
helper('media');
$this->json_key = $this->file_metadata['json_key'];
$this->json_url = service('file_manager')
->getUrl($this->json_key);
}
}
public function setFile(File $file): self
{
parent::setFile($file);
$metadata = lstat((string) $file) ?? [];
helper('filesystem');
// set metadata (generated json file path)
$this->json_key = change_file_path($this->file_key, '', 'json');
$metadata['json_key'] = $this->json_key;
$this->attributes['file_metadata'] = json_encode($metadata, JSON_INVALID_UTF8_IGNORE);
$this->file = $file;
return $this;
}
public function saveFile(): void
{
$this->saveJsonTranscript();
parent::saveFile();
}
public function deleteFile(): bool
{
if (! parent::deleteFile()) {
return false;
}
if ($this->json_key) {
return service('file_manager')->delete($this->json_key);
}
return true;
}
private function saveJsonTranscript(): void
{
$transcriptContent = file_get_contents($this->file->getRealPath());
$transcriptParser = new TranscriptParser();
if ($transcriptContent === false) {
throw new Exception('Could not read transcript file at ' . $this->file->getRealPath());
}
$transcript_format = $this->file->getExtension();
switch ($transcript_format) {
case 'vtt':
$transcriptJson = $transcriptParser->loadString($transcriptContent)
->parseVtt();
break;
case 'srt':
default:
$transcriptJson = $transcriptParser->loadString($transcriptContent)
->parseSrt();
}
$tempFilePath = WRITEPATH . 'uploads/' . $this->file->getRandomName();
file_put_contents($tempFilePath, $transcriptJson);
$newTranscriptJson = new File($tempFilePath, true);
service('file_manager')
->save($newTranscriptJson, $this->json_key);
}
}

220
modules/Media/TranscriptParser.php
View File

@ -1,220 +0,0 @@
<?php
declare(strict_types=1);
/**
* Converts a SRT or VTT file to JSON
*
* @copyright 2022 Ad Aures
* @license https://www.gnu.org/licenses/agpl-3.0.en.html AGPL3
* @link https://castopod.org/
*/
namespace Modules\Media;
use Exception;
use stdClass;
class TranscriptParser
{
protected string $transcriptContent;
public function loadString(string $content): self
{
$this->transcriptContent = $content;
return $this;
}
/**
* Adapted from: https://stackoverflow.com/a/11659306
*
* @return string Returns the json encoded string
*/
public function parseSrt(): string
{
if (! defined('SRT_STATE_SUBNUMBER')) {
define('SRT_STATE_SUBNUMBER', 0);
}
if (! defined('SRT_STATE_TIME')) {
define('SRT_STATE_TIME', 1);
}
if (! defined('SRT_STATE_TEXT')) {
define('SRT_STATE_TEXT', 2);
}
if (! defined('SRT_STATE_BLANK')) {
define('SRT_STATE_BLANK', 3);
}
$subs = [];
$state = SRT_STATE_SUBNUMBER;
$subNum = 0;
$subText = '';
$subTime = '';
$lines = explode(PHP_EOL, $this->transcriptContent);
foreach ($lines as $line) {
switch ($state) {
case SRT_STATE_SUBNUMBER:
$subNum = trim($line);
$state = SRT_STATE_TIME;
break;
case SRT_STATE_TIME:
$subTime = trim($line);
$state = SRT_STATE_TEXT;
break;
case SRT_STATE_TEXT:
if (trim($line) === '') {
$sub = new stdClass();
$sub->number = (int) $subNum;
[$startTime, $endTime] = explode(' --> ', $subTime);
$sub->startTime = $this->getSecondsFromTimeString($startTime);
$sub->endTime = $this->getSecondsFromTimeString($endTime);
$sub->text = trim($subText);
$subText = '';
$state = SRT_STATE_SUBNUMBER;
$subs[] = $sub;
} elseif ($subText !== '') {
$subText .= PHP_EOL . $line;
} else {
$subText .= $line;
}
break;
}
}
if ($state === SRT_STATE_TEXT) {
// if file was missing the trailing newlines, we'll be in this
// state here. Append the last read text and add the last sub.
// @phpstan-ignore-next-line
$sub->text = $subText;
// @phpstan-ignore-next-line
$subs[] = $sub;
}
$jsonString = json_encode($subs, JSON_PRETTY_PRINT);
if (! $jsonString) {
throw new Exception('Failed to parse SRT to JSON.');
}
return $jsonString;
}
public function parseVtt(): string
{
if (! defined('VTT_STATE_HEADER')) {
define('VTT_STATE_HEADER', 0);
}
if (! defined('VTT_STATE_BLANK')) {
define('VTT_STATE_BLANK', 1);
}
if (! defined('VTT_STATE_TIME')) {
define('VTT_STATE_TIME', 2);
}
if (! defined('VTT_STATE_TEXT')) {
define('VTT_STATE_TEXT', 3);
}
$subs = [];
$state = VTT_STATE_HEADER;
$subNum = 0;
$subText = '';
$subTime = '';
$lines = explode(PHP_EOL, $this->transcriptContent);
// add a newline as last item, if it isn't already a newline
if ($lines[array_key_last($lines)] !== '') {
$lines[] = PHP_EOL;
}
foreach ($lines as $line) {
switch ($state) {
case VTT_STATE_HEADER:
$state = VTT_STATE_BLANK;
break;
case VTT_STATE_BLANK:
$state = VTT_STATE_TIME;
break;
case VTT_STATE_TIME:
$subTime = trim($line);
$state = VTT_STATE_TEXT;
break;
case VTT_STATE_TEXT:
if (trim($line) === '') {
$sub = new stdClass();
$sub->number = $subNum;
[$startTime, $endTime] = explode(' --> ', $subTime);
$sub->startTime = $this->getSecondsFromVTTTimeString($startTime);
$sub->endTime = $this->getSecondsFromVTTTimeString($endTime);
$sub->text = trim($subText);
if ($subSpeaker !== '') {
$sub->speaker = trim((string) $subSpeaker);
}
$subText = '';
$state = VTT_STATE_TIME;
$subs[] = $sub;
++$subNum;
} elseif ($subText !== '') {
$subText .= PHP_EOL . $line;
} else {
/** VTT includes a lot of information on the spoken line
* An example may look like this:
* <v.loud.top John>So this is it
* We need to break this down into it's components, namely:
* 1. The actual words for the caption
* 2. Who is speaking
* 3. Any styling cues encoded in the VTT (which we dump)
* More information: https://www.w3.org/TR/webvtt1/
*/
$vtt_speaker_pattern = '/^<.*>/';
$removethese = ['<', '>'];
preg_match($vtt_speaker_pattern, $line, $matches);
if (isset($matches[0])) {
$subVoiceCue = explode(' ', str_replace($removethese, '', $matches[0]));
$subSpeaker = $subVoiceCue[1];
} else {
$subSpeaker = '';
}
$subText .= preg_replace($vtt_speaker_pattern, '', $line);
}
break;
}
}
$jsonString = json_encode($subs, JSON_PRETTY_PRINT);
if (! $jsonString) {
throw new Exception('Failed to parse VTT to JSON.');
}
return $jsonString;
}
private function getSecondsFromTimeString(string $timeString): float
{
$timeString = explode(',', $timeString);
return (strtotime($timeString[0]) - strtotime('TODAY')) + (float) "0.{$timeString[1]}";
}
private function getSecondsFromVTTTimeString(string $timeString): float
{
$timeString = explode('.', $timeString);
return (strtotime($timeString[0]) - strtotime('TODAY')) + (float) "0.{$timeString[1]}";
}
}

62
modules/MediaClipper/VideoClipper.php
View File

@ -13,6 +13,7 @@ namespace Modules\MediaClipper;
use App\Entities\Episode;
use Exception;
use GdImage;
use Modules\Media\CastopodSubtitles\CastopodSubtitles;
use Modules\Media\Entities\Transcript;
use Modules\Media\FileManagers\FileManagerInterface;
use Modules\MediaClipper\Config\MediaClipper;
@ -122,66 +123,17 @@ class VideoClipper
throw new Exception('Episode does not have a transcript!');
}
if ($this->episode->transcript->json_url) {
$this->generateSubtitlesClipFromJson($this->episode->transcript->json_key);
} else {
$subtitlesInput = $this->episode->transcript->file_url;
$subtitleClipCmd = "ffmpeg -y -i {$subtitlesInput} -ss {$this->start} -t {$this->duration} {$this->subtitlesClipOutput}";
exec($subtitleClipCmd);
}
}
$subtitleString = file_get_contents($this->episode->transcript->file_url);
public function generateSubtitlesClipFromJson(string $jsonFileKey): void
{
/** @var FileManagerInterface $fileManager */
$fileManager = service('file_manager');
$jsonTranscriptString = (string) $fileManager->getFileContents($jsonFileKey);
if ($jsonTranscriptString === '') {
throw new Exception('Cannot get transcript json contents.');
if (! $subtitleString) {
throw new Exception('Could not load transcript: ' . $this->episode->transcript->file_url);
}
$jsonTranscript = json_decode($jsonTranscriptString, true);
if ($jsonTranscript === null) {
throw new Exception('Transcript json is invalid.');
}
$subtitles = CastopodSubtitles::loadFromString($subtitleString);
$srtClip = '';
$segmentIndex = 1;
foreach ($jsonTranscript as $segment) {
$startTime = null;
$endTime = null;
$subtitles->trim($this->start, $this->end);
if ($segment['startTime'] < $this->end && $segment['endTime'] > $this->start) {
$startTime = $segment['startTime'] - $this->start;
$endTime = $segment['endTime'] - $this->start;
}
if ($segment['startTime'] < $this->start && $this->start < $segment['endTime']) {
$startTime = 0;
}
if ($segment['startTime'] < $this->end && $segment['endTime'] >= $this->end) {
$endTime = $this->duration;
}
if ($startTime !== null && $endTime !== null) {
$formattedStartTime = $this->formatSeconds($startTime);
$formattedEndTime = $this->formatSeconds($endTime);
$srtClip .= <<<CODE_SAMPLE
{$segmentIndex}
{$formattedStartTime} --> {$formattedEndTime}
{$segment['text']}
CODE_SAMPLE;
++$segmentIndex;
}
}
// create srt clip file
file_put_contents($this->subtitlesClipOutput, $srtClip);
$subtitles->save($this->subtitlesClipOutput);
}
public function formatSeconds(float $seconds): string