<?php

/**
 * Bart Nagel <bjn@ecs.soton.ac.uk>
 *
 * AFRepo top-level class to slot in with the Greenstone extension framework
 *
 * The audio is expected to be in subdirectories of the 'audio-files' directory, and 
 * for example purposes can be named in such a way that the example 
 * PathClassifier classifier can get meanining from it.
 */

class AFRepo extends AFRepoBase {
	private $allfiles;

	public function AFRepo() 
	{
	  $this->audio_files_dir = "audio-files";
	  $this->af_prefix = getcwd() . "/" . $this->audio_files_dir;
	}

	public function getName() {
		return "Salami 4Store audio repository";
	}

	public function getURIPrefix() {
		#return "@afrepo-http-prefix@/afrepo/";
		return "@public-facing-afrepo-http-prefix@";
	}

	/* Done this way, can talk to the 4store server directly
           rather than relying on proxying through the PHP-based 'afrepo' server */
	public function getTrippleStoreLocalURIPrefix() {
	  //return "@afrepo-http-prefix@/4store/";
		return "@4store-http-prefix@/";
	}

	public function getTrippleStoreURIPrefix() {
		return "@public-facing-4store-http-prefix@";
	}

	public function getSparqlEndpoint() {
		return $this->getTrippleStoreURIPrefix() . "sparql/";
	}

	public function getDataEndpoint() {
		return $this->getTrippleStoreLocalURIPrefix() . "data/";
	}

	/**
	 * getAudioPath
	 * Return the path to the audio links in this repository (without a trailing 
	 * slash)
	 */
	public function getAudioPath() {
		return realpath("audio-ids");
	}


	public function remove_audio_files_prefix($full_filepath) {
#	  echo "full filepath = " . $full_filepath . "\n";
#	  echo "af prefix = " . $this->af_prefix . "\n";

	  $af_prefix_len = strlen($this->af_prefix);
	  $filepath = $full_filepath; 

	  if (substr($filepath, 0, $af_prefix_len) == $this->af_prefix) {
	    $filepath = substr($filepath, $af_prefix_len);
	  } 

#	  echo "** filepath = " . $filepath . "\n";

	  return $filepath;
	}


	/**
	 * fileInRepo
	 * Return true if the audiofile with the given path (canonical or symlink) 
	 * is in the repository or false if not
	 */
	public function fileInRepo($full_filepath) {
	  $realpath = realpath($full_filepath);
	  if ($realpath === false) {
	    trigger_error("file '$filepath' does not exist on disk or is a broken symlink", E_USER_WARNING);
	    return false;
	  }

	  return array_key_exists($full_filepath, $this->getAllFiles());
	}


	/**
	 * idToCanonicalPath
	 * Return the path to the canonical file with the given ID
	 */
	public function idToCanonicalPath($id) {
	    return readlink($this->idToLinkPath($id));
	}


	/**
	 * filePathToId
	 * Return the ID of the audiofile with the given path (which can be 
	 * canonical or a symlink)
	 */
	public function filePathToId($full_filepath) {
	  if (!$this->fileInRepo($full_filepath)) {
	    throw new Exception("file with path '$filepath' is not in the repository");
	  }

	  $hash_filepath = "salami:/" . $this->remove_audio_files_prefix($full_filepath);

##	  echo "Hashing on: " . $hash_filepath . "\n";

	  $id = md5($hash_filepath);
	  return $id;
	}


	// recursive method to delete things from the links directory according to the 
	// options
	function getAllCollectionFilesRec($collection, $path) {

	  echo "Processing directory: " . $path . "\n";

	  $dir = dir($path);

	  while (($entry = $dir->read()) !== false) {

	    if ($entry == "." || $entry == "..") {
	      continue;
	    }

	    $fullpath = $path . "/" . $entry;

	    $isdir = false;
	    if (is_link($fullpath)) {

	      $realpath = readlink($fullpath);

	      // follow potential chain of sym-links
	      while (is_link($realpath)) {
		$realpath = readlink($realpath);
	      }

	      $isdir = is_dir($realpath);
	    }
	    else {
	      $isdir = is_dir($fullpath);
	    }


	    if ($isdir) {

	      $this->getAllCollectionFilesRec($collection,$fullpath);
	    } 
	    else {
	      // assume it is a file
##	      echo "Adding file: " . $fullpath . "\n";
	      $this->allfiles[$fullpath] = true;
	    }
	  }

	  $dir->close();
	  
	}
	
	
	public function getAllFiles() 
	{
	  if (!is_null($this->allfiles)) {
	    return $this->allfiles;
	  }
	  
	  $this->allfiles = array();

##	  $path = realpath($this->audio_files_dir);	  
	  $path = $this->af_prefix;
	  $dir  = dir($path);
	  
	  while (($file = $dir->read()) !== false) {

	    if ($file[0] == ".") {
	      // skip all dot files and dirs
	      continue;
	    }

	    $fullpath = $path . "/" . $file;

	    $isdir = false;
	    if (is_link($fullpath)) {

	      $realpath = readlink($fullpath);

	      // follow potential chain of sym-links
	      while (is_link($realpath)) {
		$realpath = readlink($realpath);
	      }

	      $isdir = is_dir($realpath);
	    }
	    else {
	      $isdir = is_dir($fullpath);
	    }

	    if ($isdir) {

	      // recursively work through each collection directory
	      $this->getAllCollectionFilesRec($file, $fullpath);
	    }
	  }

	  $dir->close();
	  
	  return $this->allfiles;
	}


	public function getAllFilesOLD() {
		if (!is_null($this->allfiles))
			return $this->allfiles;

		$this->allfiles = array();
		$path = realpath("audio-files");
		$dir = dir($path);
		while (($file = $dir->read()) !== false) {
			if ($file[0] != "." && is_dir($path . "/" . $file)) {
				$subdir = dir($path . "/" . $file);
				while (($subfile = $subdir->read()) !== false) {
					if ($subfile[0] != "." && is_file($path . "/" . $file . "/" . $subfile))
						$this->allfiles[$path . "/" . $file . "/" . $subfile] = true;
				}
				$subdir->close();
			}
		}
		$dir->close();

		return $this->allfiles;
	}

	public function getSongFilesOLD($id) {
		$filepath = $this->idToLinkPath($id);
		$origfilepath = realpath($filepath);

		if ($origfilepath === false)
			return array();

		// is it a clip?
		if (preg_match('%\.clip\..{1,4}$%', $origfilepath)) {
			// does full version exist?
			$fullpath = preg_replace('%\.clip%', "", $origfilepath);
			if (file_exists($fullpath))
				return array($fullpath, $origfilepath);
			return array($origfilepath);
		}

		// it's a full song. does clip exist?
		$clippath = preg_replace('%(\..{1,4})$%', '.clip\1', $origfilepath);
		if (file_exists($clippath))
			return array($origfilepath, $clippath);
		return array($origfilepath);
	}

	public function haveMetadataPermission() {
		return true;
	}

	public function haveAudioPermission() {
		return ipInRange($_SERVER["REMOTE_ADDR"], "127.0.0.0/8");
	}

	public function getMBID($id) {
		$classifiers = array(
			new TagClassifier(),
			new EchoprintClassifier(),
			new EchonestClassifier(),
			new PathClassifier(),
		);

		foreach ($classifiers as $classifier) 
			if ($classifier->available() && $classifier->hasMBID($id))
				return $classifier->getMBID($id);
		return null;
	}


	/**
	 * extraTriples
	 * Return an array of any extra implementation-specific triples should be 
	 * added to the RDF given the ID of an audiofile
	 *
	 * This should be an array of triples in Turtle format without the trailing 
	 * dot. (See generateRDF)
	 * It should return triples about all audiofiles of the same song rather 
	 * than just the one with the ID given.
	 */
	protected function extraTriples($id) {

		$extra_triples = array();


		// analyze the file, get some metadata
		$structured_metadata = $this->getStructuredMetadata($id);

		if ($structured_metadata !== false) {

		}

		// loop through the audiofiles for this tune
		/* foreach ($ids as $key => $fileid) { */

		/* 	// some identifiers */
		/* 	$audiofile = "repo:$fileid"; */
		/* 	$digitalsignal = "repo:$fileid#DigitalSignal"; */

		/* 	// triples about the document */
		/* 	$triples[] = "{$audiofile}_ a foaf:Document; foaf:primaryTopic $audiofile"; */

		/* 	// this is a mo:AudioFile, which is a mo:MusicalItem */
		/* 	$triples[] = "$audiofile a mo:AudioFile"; */

		/* 	// this encodes a corresponding digital signal */
		/* 	$triples[] = "$audiofile mo:encodes $digitalsignal"; */

		/* 	// our digital signal is a mo:DigitalSignal (which is a subclass of  */
		/* 	// mo:Signal, which is a subclass of mo:MusicalExpression) */
		/* 	$triples[] = "$digitalsignal a mo:DigitalSignal"; */

		/* 	// different logic depending whether this audiofile is the preferred  */
		/* 	// one or not */
		/* 	if ($key == 0) { */
		/* 		// preferred -- if we have an MBID this Signal is derived from  */
		/* 		// the original (that at Musicbrainz) */
		/* 		if (!is_null($mbid)) */
		/* 			$triples[] = "$digitalsignal mo:derived_from " . mbidToSignalURI($mbid); */
		/* 		// otherwise we don't assert that it derives from anything */
		/* 	} else { */
		/* 		// non-preferred -- we assert that it is derived from our  */
		/* 		// preferred audiofile's Signal */
		/* 		$triples[] = "$digitalsignal mo:derived_from repo:" . $ids[0] . "#DigitalSignal"; */
		/* 	} */

		/* 	// analyze the file, get some metadata */
		/* 	$filemetadata = $this->getFileMetadata($fileid); */

		/* 	// mo:AudioFile metadata */
		/* 	if (isset($filemetadata["dataformat"])) */
		/* 		$triples[] = "$audiofile mo:encoding \"" . $filemetadata["dataformat"] */
		/* 				. (isset($filemetadata["bitrate"]) ? " @ " . $filemetadata["bitrate"] . "bps" : "") */
		/* 				. (isset($filemetadata["bitrate_mode"]) ? " " . $filemetadata["bitrate_mode"] : "") */
		/* 				. "\""; */

		/* 	// mo:DigitalSignal metadata */
		/* 	if (isset($filemetadata["playtime_seconds"])) */
		/* 		$triples[] = "$digitalsignal mo:time [ " */
		/* 				. "a time:Interval; " */
		/* 				. "time:seconds \"" . $filemetadata["playtime_seconds"] . "\"^^xsd:float " */
		/* 				. "]"; */
		/* 	if (isset($filemetadata["channels"])) */
		/* 		$triples[] = "$digitalsignal mo:channels \"" . $filemetadata["channels"] . "\"^^xsd:int"; */
		/* 	if (isset($filemetadata["sample_rate"])) */
		/* 		$triples[] = "$digitalsignal mo:sample_rate \"" . $filemetadata["sample_rate"] . "\"^^xsd:float"; */
		/* } */

		return $extra_triples;


	}


	/**
	 * getgetStructuredMetadata
	 * Return as a hashmap (keyed by algorithm the structured audio data
	 * from the "big crunch"
	 */
	public function getStructuredMetadata($id) {

	  $classifier = new StructuredAudioClassifier();

	  if ($classifier->available() && $classifier->hasMetadata($id)) {
	    return $classifier->loadMetadata($id);
	  }

	  return false;

	}


	public function attributionToHtmlByAlgorithm() {

	  $classifier = new StructuredAudioClassifier();

	  $full_dataset_path = getcwd() . "/" . "datasets";


	  $ametadata = $classifier->getAttribution();

	  $style = "<style>\n";
	  $style .= "  tr:nth-child(even) {background: #DDD}\n";
	  $style .= "  tr:nth-child(odd)  {background: #EEE}\n";
	  $style .= "</style>\n";

	  $html_str = "$style\n<table>";

	  foreach ($ametadata as $key => $value) {

	    $html_str .= "<tr>";

	    $entry = $value;
	    $title      = $entry["title"];
	    $year       = $entry["year"];
	    $articleUrl = $entry["articleUrl"];
	    $authorList = $entry["authorList"];
	    $authors    = $entry["authors"];
  
	    $opt_linked_title = $title;

	    if ($articleUrl != false) {
	      $opt_linked_title = "<a href=\"$articleUrl\">$title</a>";
	    }

	    $desc_entry = "<i>$opt_linked_title</i> ($year)<br />\n";

	    foreach ($authorList as $author) {
	      $author_details = $authors[$author];
	      $organizations  = $author_details["organization"];
	      $email          = $author_details["email"];

	      $desc_entry .= $author;
	      if ($email != false) {
		$email = str_replace("@"," at ",$email);
		$desc_entry .= " (<tt>$email</tt>)";
	      }

	      $desc_entry .= "<br />\n";
	    }

	    $tarfilesize = filesize("$full_dataset_path/$key.tar.gz");

	    $download_key = "<a href=\"datasets/$key.tar.gz\">Download $key</a> ($tarfilesize bytes)"; 
	    $html_str .= "  <td>$download_key</td><td>$desc_entry</td>\n";

	    $html_str .= "</tr>\n";

	  }
	  $html_str .= "</table>";

	  return $html_str;

	}


	public function attributionToHtmlByCollection() {

##	  $classifier = new StructuredAudioClassifier();

	  $full_dataset_path = getcwd() . "/" . "datasets";


	  $cmetadata = array (
			      "american"  => "American",
			      "classical" => "Classical",
			      "coversong" => "Cover Songs",
			      "dram"      => "Dram",
			      "ehmann"    => "Ehmann",
			      "jamendo"   => "Jameno",
			      "latin"     => "Latin",
			      "magnatune" => "Magnatune",
			      "mazurkas"  => "Mazurkas",
			      "uscrap"    => "US Bargin Bin",
			      "uspop"     => "US Pop"
			      );

	  $style = "<style>\n";
	  $style .= "  tr:nth-child(even) {background: #DDD}\n";
	  $style .= "  tr:nth-child(odd)  {background: #EEE}\n";
	  $style .= "</style>\n";

	  $html_str = "$style\n<table>";

	  foreach ($cmetadata as $key => $value) {

	    $html_str .= "<tr>";

	    $tarfilesize = filesize("$full_dataset_path/$key.tar.gz");

	    $download_key = "<a href=\"datasets/$key.tar.gz\">Download $value</a> ($tarfilesize bytes)"; 
	    $html_str .= "  <td>$download_key</td>\n";

	    $html_str .= "</tr>\n";

	  }
	  $html_str .= "</table>";

	  return $html_str;

	}


}

?>