Limit scaling of 1-bit images for more legible text!

[bookreader.git] / BookReaderIA / datanode / BookReaderImages.php

diff --git a/BookReaderIA/datanode/BookReaderImages.php b/BookReaderIA/datanode/BookReaderImages.php
index 9099ba3..35a67e0 100644 (file)
--- a/BookReaderIA/datanode/BookReaderImages.php
+++ b/BookReaderIA/datanode/BookReaderImages.php
@@ -5,6 +5,10 @@ Copyright(c)2008 Internet Archive. Software license AGPL version 3.
 
 This file is part of BookReader.
 
+The canonical short name of an image type is the same as in the MIME type.
+For example both .jpeg and .jpg are considered to have type "jpeg" since
+the MIME type is "image/jpeg".
+
     BookReader is free software: you can redistribute it and/or modify
     it under the terms of the GNU Affero General Public License as published by
     the Free Software Foundation, either version 3 of the License, or
@@ -19,8 +23,21 @@ This file is part of BookReader.
     along with BookReader.  If not, see <http://www.gnu.org/licenses/>.
 */
 
-$MIMES = array('jpg' => 'image/jpeg',
-               'png' => 'image/png');
+$MIMES = array('gif' => 'image/gif',
+               'jp2' => 'image/jp2',
+               'jpg' => 'image/jpeg',
+               'jpeg' => 'image/jpeg',
+               'png' => 'image/png',
+               'tif' => 'image/tiff',
+               'tiff' => 'image/tiff');
+               
+$EXTENSIONS = array('gif' => 'gif',
+                    'jp2' => 'jp2',
+                    'jpeg' => 'jpeg',
+                    'jpg' => 'jpeg',
+                    'png' => 'png',
+                    'tif' => 'tiff',
+                    'tiff' => 'tiff');
                
 $exiftool = '/petabox/sw/books/exiftool/exiftool';
 
@@ -76,7 +93,21 @@ function getUnarchiveCommand($archivePath, $file)
     BRfatal('Bad image stack path or archive format');
     
 }
- 
+
+/*
+ * Returns the image type associated with the file extension.
+ */
+function imageExtensionToType($extension)
+{
+    global $EXTENSIONS;
+    
+    if (array_key_exists($extension, $EXTENSIONS)) {
+        return $EXTENSIONS[$extension];
+    } else {
+        BRfatal('Unknown image extension');
+    }            
+}
+
 /*
  * Get the image width, height and depth from a jp2 file in zip.
  */
@@ -84,17 +115,51 @@ function getImageInfo($zipPath, $file)
 {
     global $exiftool;
     
-    // $$$ will exiftool work for *all* of our images?
-    // BitsPerComponent present in jp2. Not present in jpeg.
+    $fileExt = strtolower(pathinfo($file, PATHINFO_EXTENSION));
+    $type = imageExtensionToType($fileExt);
+     
+    // We look for all the possible tags of interest then act on the
+    // ones presumed present based on the file type
+    $tagsToGet = ' -ImageWidth -ImageHeight -FileType'        // all formats
+                 . ' -BitsPerComponent -ColorSpace'          // jp2
+                 . ' -BitDepth'                              // png
+                 . ' -BitsPerSample';                        // tiff
+                        
     $cmd = getUnarchiveCommand($zipPath, $file)
-        . ' | '. $exiftool . ' -s -s -s -ImageWidth -ImageHeight -BitsPerComponent -';
+        . ' | '. $exiftool . ' -S -fast' . $tagsToGet . ' -';
     exec($cmd, $output);
     
-    preg_match('/^(\d+)/', $output[2], $groups);
-    $bits = intval($groups[1]);
+    $tags = Array();
+    foreach ($output as $line) {
+        $keyValue = explode(": ", $line);
+        $tags[$keyValue[0]] = $keyValue[1];
+    }
+    
+    $width = intval($tags["ImageWidth"]);
+    $height = intval($tags["ImageHeight"]);
+    $type = strtolower($tags["FileType"]);
     
-    $retval = Array('width' => intval($output[0]), 'height' => intval($output[1]),
-        'bits' => $bits);
+    switch ($type) {
+        case "jp2":
+            $bits = intval($tags["BitsPerComponent"]);
+            break;
+        case "tiff":
+            $bits = intval($tags["BitsPerSample"]);
+            break;
+        case "jpeg":
+            $bits = 8;
+            break;
+        case "png":
+            $bits = intval($tags["BitDepth"]);
+            break;
+        default:
+            BRfatal("Unsupported image type");
+            break;
+    }
+   
+   
+    $retval = Array('width' => $width, 'height' => $height,
+        'bits' => $bits, 'type' => $type);
     
     return $retval;
 }
@@ -117,6 +182,7 @@ $imageInfo = getImageInfo($zipPath, $file);
 
 // Output json if requested
 if ('json' == $ext) {
+    // $$$ we should determine the output size first based on requested scale
     outputJSON($imageInfo, $callback);
     exit;
 }
@@ -149,7 +215,7 @@ $jpegOptions = '-quality 75';
 
 // The pbmreduce reduction factor produces an image with dimension 1/n
 // The kakadu reduction factor produceds an image with dimension 1/(2^n)
-
+// $$$ handle continuous values for scale
 if (isset($_REQUEST['height'])) {
     $ratio = floatval($_REQUEST['origHeight']) / floatval($_REQUEST['height']);
     if ($ratio <= 2) {
@@ -186,6 +252,22 @@ if (isset($_REQUEST['height'])) {
     }
 }
 
+// Override depending on source image format
+// $$$ consider doing a 302 here instead, to make better use of the browser cache
+// Limit scaling for 1-bit images.  See https://bugs.edge.launchpad.net/bookreader/+bug/486011
+if (1 == $imageInfo['bits']) {
+    if ($scale > 1) {
+        $scale /= 2;
+        $powReduce -= 1;
+        
+        // Hard limit so there are some black pixels to use!
+        if ($scale > 4) {
+            $scale = 4;
+            $powReduce = 2;
+        }
+    }
+}
+
 if (!file_exists($stdoutLink)) 
 {  
   system('ln -s /dev/stdout ' . $stdoutLink);  
@@ -196,6 +278,7 @@ putenv('LD_LIBRARY_PATH=/petabox/sw/lib/kakadu');
 
 $unzipCmd  = getUnarchiveCommand($zipPath, $file);
         
+// XXX look at normalized type in imageinfo
 if ('jp2' == $fileExt) {
     $decompressCmd = 
         " | /petabox/sw/bin/kdu_expand -no_seek -quiet -reduce $powReduce -rotate $rotate -i /dev/stdin -o " . $stdoutLink;
@@ -210,13 +293,15 @@ if ('jp2' == $fileExt) {
     // get cleaned up.
     $tempFile = tempnam("/tmp", "BookReaderTiff");
 
-    $pbmReduce = reduceCommand($scale);
-    
+    // $$$ look at bit depth when reducing
     $decompressCmd = 
-        ' > ' . $tempFile . ' ; tifftopnm ' . $tempFile . ' 2>/dev/null' . $pbmReduce;
+        ' > ' . $tempFile . ' ; tifftopnm ' . $tempFile . ' 2>/dev/null' . reduceCommand($scale);
         
 } else if ('jpg' == $fileExt) {
     $decompressCmd = ' | jpegtopnm ' . reduceCommand($scale);
+
+} else if ('png' == $fileExt) {
+    $decompressCmd = ' | pngtopnm ' . reduceCommand($scale);
     
 } else {
     BRfatal('Unknown source file extension: ' . $fileExt);
@@ -243,6 +328,7 @@ if (($ext == $fileExt) && ($scale == 1) && ($rotate === "0")) {
 # print $cmd;
 
 
+// $$$ investigate how to flush cache when this file is changed
 header('Content-type: ' . $MIMES[$ext]);
 header('Cache-Control: max-age=15552000');
 passthru ($cmd); # cmd returns image data