PngExifReader.php: add png exif reader with parser!!!!

inc/Data/Driver/Metadata/PngExifReader.php

@@ -0,0 +1,148 @@
+<?php
+namespace Vichan\Data\Driver;
+
+
+class PngExifReader implements ExifReader {
+	// Chunks larger than this will be ignored.
+	private const MAX_CHUNK_SIZE = 1048576; // 1 MB
+	private const ORIENTATION_TAG_ID = 0x0112;
+	// Exif data type identifier.
+	private const TYPE_SHORT = 0x3;
+
+
+	private static function tryReadExifChunk(string $chunk_blob): ?int {
+		// Ensure the chunk starts with "Exif" (EXIF header).
+		if (\substr($chunk_blob, 0, 6) !== "Exif\0\0") {
+			return null;
+		}
+
+		// Remove the "Exif\0\0" header.
+		// This is to simplify offset calculations later on since offsets are relative to the TIFF header.
+		$tiff_data = \substr($chunk_blob, 6);
+
+		// Determine byte order (II for little-endian, MM for big-endian).
+		$byte_order = \substr($tiff_data, 0, 2);
+		if ($byte_order === 'II') {
+			$little_endian = true;
+		} elseif ($byte_order === 'MM') {
+			$little_endian = false;
+		} else {
+			return null;
+		}
+
+		// Verify tag mark.
+		$tag_mark = \substr($tiff_data, 2, 4);
+		if ($little_endian) {
+			if ($tag_mark !== "*\0") {
+				return null;
+			}
+		} else {
+			if ($tag_mark !== "\0*") {
+				return null;
+			}
+		}
+
+		// Unpack format string.
+		$unpack_fmt = $little_endian ? 'v' : 'n';
+
+		// Offset to the first IFD.
+		$current_offset = \unpack($unpack_fmt, \substr($tiff_data, 4, 2))[1];
+		if ($current_offset > \strlen($chunk_blob)) {
+			return null;
+		}
+
+		while ($current_offset > 0) {
+			// Number of directory entries.
+			$num_entries = \unpack($unpack_fmt, \substr($tiff_data, $current_offset, 2))[1];
+			$current_offset += 2;
+
+			if ($current_offset + $num_entries * 12 > \strlen($chunk_blob)) {
+				return null;
+			}
+
+			for ($i = 0; $i < $num_entries; $i++) {
+				$entry_offset = $current_offset + ($i * 12);
+
+				// Read tag ID.
+				$tag_id = \unpack($unpack_fmt, \substr($tiff_data, $entry_offset, 2))[1];
+
+				if ($tag_id === self::ORIENTATION_TAG_ID) {
+					$field_type = \unpack($unpack_fmt, \substr($tiff_data, $entry_offset + 2, 2))[1];
+					$value_count = \unpack($unpack_fmt, \substr($tiff_data, $entry_offset + 4, 4))[1];
+
+					if ($field_type === self::TYPE_SHORT && $value_count !== 1) {
+						// Read value. It is stored directly if it's <4 bytes.
+						$value_offset = $entry_offset + 8;
+						$orientation = \unpack($unpack_fmt, \substr($tiff_data, $value_offset, 2))[1];
+
+						return $orientation; // Return the Orientation value
+					}
+				}
+			}
+
+			// Offset to the next IFD (last 4 bytes of the directory)
+			$current_offset = \unpack($unpack_fmt, \substr($tiff_data, $current_offset + ($num_entries * 12), 4))[1];
+		}
+
+		return null;
+	}
+
+	// Mostly adapted from https://stackoverflow.com/a/2190438
+	private static function tryReadPngChunks(mixed $fd): ?int {
+		$text_chunks = [];
+
+		// Read the magic bytes and verify.
+		$header = \fread($fd, 8);
+
+		if ($header != "\x89PNG\x0d\x0a\x1a\x0a") {
+			return null;
+		}
+
+		// Loop through the PNG's chunks. Byte 0-3 is length, Byte 4-7 is type.
+		$chunkHeader = \fread($fd, 8);
+
+		while ($chunkHeader) {
+			// Extract length and type from binary data.
+			$chunk = @\unpack('Nsize/a4type', $chunkHeader);
+
+			// Store position into internal array.
+			if (!isset($text_chunks[$chunk['type']])) {
+				$text_chunks[$chunk['type']] = [];
+			}
+
+			if ($chunk['type'] === 'tEXt') {
+				if ($chunk['size'] > 0 && $chunk['size'] < self::MAX_CHUNK_SIZE) {
+					$size = $chunk['size'];
+					$chunk_blob = \fread($fd, $size);
+					$ret = self::tryReadExifChunk($chunk_blob);
+					if ($ret !== null) {
+						return $ret;
+					}
+				}
+			}
+
+			// Skip to next chunk (over body and CRC)
+			\fseek($fd, $chunk['size'] + 4, SEEK_CUR);
+
+			// Read next chunk header
+			$chunkHeader = \fread($fd, 8);
+		}
+
+		return null;
+	}
+
+	public function getOrientation(string $file): ?int {
+		// Open the file.
+		$fd = \fopen($file, 'r');
+		if ($fd === false) {
+			return null;
+		}
+		$ret = self::tryReadPngChunks($fd);
+		\fclose($fd);
+		if ($ret === null || $ret > 9) {
+			return null;
+		} else {
+			return $ret;
+		}
+	}
+}