univ-pau-ics/build/service/Tesseract.php

<?php

	namespace service;

	class Tesseract{

		/* [1] Attributes
		=========================================================*/
		private $fname    = null;


		/* (1) Constructs and initialise a readed file
		*
		* @fname<String>						Path of the image
		*
		* @return instance<Tesseract>			New Tesseract
		*
		---------------------------------------------------------*/
		public function __construct($fname=null){

			/* [1] Check argument
			=========================================================*/ {

				/* (1) Check type */
				if( !is_string($fname) )
					throw new \Exception("Tesseract.__construct(<String>) expected but Tesseract.__construct(<".gettype($fname).">) received");

				/* (2) Check file validity */
				if( !file_exists($fname) )
					throw new \Exception("Tesseract.__construct(<PATH>) but <PATH> is not valid");

			}


			/* [2] Store as attribute
			=========================================================*/
		    $this->fname = $fname;

		}



		/* (2) Read the image file
		*
		* @return read<String>					The read content
		*
		---------------------------------------------------------*/
		public function read(){

			/* [1] Record the text from the image
			=========================================================*/ {

				/* (1) Process tesseract */
				$read = shell_exec("tesseract ".$this->fname." stdout -l fra --user-words ".__ROOT__."/config/edt.user-words -c language_model_penalty_non_freq_dict_word=0.3 -c language_model_penalty_non_dict_word=.25 -c tessedit_char_whitelist=abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ0123456789 2>/dev/null");

				/* (2) If empty */
				if( is_null($read) || !preg_match('@\n@m', $read) )
					throw new \Exception("Nothing read");

				/* (3) Split by lines */
				$by_line = explode("\n", $read);
				$lines = [];

				/* (4) Remove empty lines */
				for( $i = 0 ; $i < count($by_line) ; $i++ ){

					if( !empty( trim($by_line[$i]) ))
						$lines[] = $by_line[$i];

				}

				/* (5) Manage if empty */
				if( count($lines) < 2 )
					throw new \Exception("Nothing read");

			}


			/* [2] Extract data
			=========================================================*/ {

				/* (1) Get first non-empty line (title) */
				$title = $lines[0];

				/* (2) Get last non-empty line */
				for( $i = count($lines)-1 ; $i > 0 ; $i-- ){

					// Amphi ...  //
					if( preg_match('@^a[nm]([bp][hln])?[ir] ?(.+)$@i', $lines[$i], $m) ) // 'amphi A', 'amphi 600 droit'
						return [ $title, "Amphi ${m[1]}" ];


					// S... OR 5... //
					if( preg_match('@^[S|5] ?(\d+)@i', $lines[$i], $m) )   // 'S10', 'S22'
						return [ $title, "S. ${m[1]}" ];

				}

			}


			return [ $title, null ];

		}


	}