From 960f490bebe8cc0e764a2161bfe4bd39dd2a75a0 Mon Sep 17 00:00:00 2001
From: xdrm-brackets <doowap31@gmail.com>
Date: Thu, 14 Sep 2017 16:57:42 +0200
Subject: [PATCH] Improved location read + tesseract

---
 build/service/CalendarExtractor.php | 11 ++---
 build/service/Tesseract.php         | 65 +++++++++++++++++++----------
 2 files changed, 50 insertions(+), 26 deletions(-)

diff --git a/build/service/CalendarExtractor.php b/build/service/CalendarExtractor.php
index 44c9fe2..4d8c7ed 100644
--- a/build/service/CalendarExtractor.php
+++ b/build/service/CalendarExtractor.php
@@ -266,7 +266,7 @@
 
 				/* (2) Manage copy error */
 				if( !$copied )
-					throw new \Exception("Cannot clip image");
+					return [ null, null ];
 
 				/* (3) Save to jpeg */
 				\imagesavealpha($clip, true);
@@ -292,7 +292,7 @@
 				/* (2) Manage error */
 				}catch(\Exception $e){
 
-					$read = [ 'unkown', 'unknown' ];
+					$read = [ null, null ];
 
 				}
 
@@ -430,9 +430,10 @@
 					$RAW .= "DTSTART:${start_t}\n";
 					$RAW .= "DTEND:${data[0]}\n";
 					$RAW .= "UID:$start_t-univ-pau-ics\n";     // required
-					$RAW .= "SUMMARY:${data[1]}\n";
-					$RAW .= "LOCATION:${data[2]}\n";
-					// $RAW .= "ATTACH;ENCODING=BASE64;VALUE=BINARY;FILENAME=att.jpg:${data[1]}\n";
+					if( !is_null($data[1]) )
+						$RAW .= "SUMMARY:${data[1]}\n";
+					if( !is_null($data[2]) )
+						$RAW .= "LOCATION:${data[2]}\n";
 					$RAW .= "CATEGORIES: UPPA Calendar\n";
 					$RAW .= "END:VEVENT\n";
 				}
diff --git a/build/service/Tesseract.php b/build/service/Tesseract.php
index d38d815..d96b643 100755
--- a/build/service/Tesseract.php
+++ b/build/service/Tesseract.php
@@ -48,37 +48,60 @@
 		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=0.1 -c language_model_penalty_non_dict_word=.15 -c tessedit_char_whitelist=abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ0123456789 2>/dev/null");
+			=========================================================*/ {
 
-			/* (2) If empty */
-			if( is_null($read) || !preg_match('@\n@g', $read) )
-				throw new \Exception("Nothing read");
+				/* (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=0.1 -c language_model_penalty_non_dict_word=.15 -c tessedit_char_whitelist=abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ0123456789 2>/dev/null");
 
-			/* (3) Split by lines */
-			$by_line = explode("\n", $read);
+				// var_dump($read);
 
-			/* (4) Get first line (title) */
-			$title = $by_line[0];
+				/* (2) If empty */
+				if( is_null($read) || !preg_match('@\n@m', $read) )
+					throw new \Exception("Nothing read");
 
-			/* (5) Get last non-empty line */
-			for( $i = count($by_line)-1 ; $i > 0 ; $i-- ){
+				/* (3) Split by lines */
+				$by_line = explode("\n", $read);
+				$lines = [];
 
-				// {1} Check not empty //
-				if( empty($by_line[$i]) )
-					continue;
+				/* (4) Remove empty lines */
+				for( $i = 0 ; $i < count($by_line) ; $i++ ){
 
-				// {2} Matches //
-				if( preg_match('@^amphi@i', $by_line[$i]) || // 'amphi A', 'amphi 600 droit'
-					preg_match('@^S\d+@i', $by_line[$i])     // 'S10', 'S22'
-				)
-					return [ $title, $by_line[$i] ];
+					if( !empty( trim($by_line[$i]) ))
+						$lines[] = $by_line[$i];
+
+				}
+
+				/* (5) Manage if empty */
+				if( count($lines) < 2 )
+					throw new \Exception("Nothing read");
 
 			}
 
 
-			return [ $title, 'unknown' ];
+			/* [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 ; $i-- ){
+
+					// Amphi ...  //
+					if( preg_match('@^a[nm][bp][hl]i ?(.+)$@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 ];
 
 		}