From 2d7c703c52cae50034fc9618c72552365f7cc741 Mon Sep 17 00:00:00 2001 From: jvoisin Date: Sun, 1 Apr 2018 00:43:36 +0200 Subject: Add support for .tiff files --- src/images_pixbuf.py | 49 +++++++++++++++++++++++++++++++++++++++++++++++++ src/jpg.py | 30 ------------------------------ tests/data/dirty.tiff | Bin 0 -> 936524 bytes tests/test_libmat2.py | 30 ++++++++++++++++++++++++++---- 4 files changed, 75 insertions(+), 34 deletions(-) create mode 100644 src/images_pixbuf.py delete mode 100644 src/jpg.py create mode 100644 tests/data/dirty.tiff diff --git a/src/images_pixbuf.py b/src/images_pixbuf.py new file mode 100644 index 0000000..8eeffbe --- /dev/null +++ b/src/images_pixbuf.py @@ -0,0 +1,49 @@ +import subprocess +import json +import os + +import gi +gi.require_version('GdkPixbuf', '2.0') +from gi.repository import GdkPixbuf + +from . import abstract + +class GdkPixbufAbstractParser(abstract.AbstractParser): + def get_meta(self): + out = subprocess.check_output(['exiftool', '-json', self.filename]) + meta = json.loads(out.decode('utf-8'))[0] + for key in self.meta_whitelist: + meta.pop(key, None) + return meta + + def remove_all(self): + _, extension = os.path.splitext(self.filename) + pixbuf = GdkPixbuf.Pixbuf.new_from_file(self.filename) + if extension == '.jpg': + extension = '.jpeg' + pixbuf.savev(self.output_filename, extension[1:], ["quality"], ["100"]) + return True + + +class JPGParser(GdkPixbufAbstractParser): + mimetypes = {'image/jpg'} + meta_whitelist = {'SourceFile', 'ExifToolVersion', 'FileName', + 'Directory', 'FileSize', 'FileModifyDate', 'FileAccessDate', + "FileInodeChangeDate", 'FilePermissions', 'FileType', + 'FileTypeExtension', 'MIMEType', 'ImageWidth', + 'ImageSize', 'BitsPerSample', 'ColorComponents', 'EncodingProcess', + 'JFIFVersion', 'ResolutionUnit', 'XResolution', 'YCbCrSubSampling', + 'YResolution', 'Megapixels', 'ImageHeight'} + + +class TiffParser(GdkPixbufAbstractParser): + mimetypes = {'image/tiff'} + meta_whitelist = {'Compression', 'ExifByteOrder', 'ExtraSamples', + 'FillOrder', 'PhotometricInterpretation', 'PlanarConfiguration', + 'RowsPerStrip', 'SamplesPerPixel', 'StripByteCounts', + 'StripOffsets', 'BitsPerSample', 'Directory', 'ExifToolVersion', + 'FileAccessDate', 'FileInodeChangeDate', 'FileModifyDate', + 'FileName', 'FilePermissions', 'FileSize', 'FileType', + 'FileTypeExtension', 'ImageHeight', 'ImageSize', 'ImageWidth', + 'MIMEType', 'Megapixels', 'SourceFile'} + diff --git a/src/jpg.py b/src/jpg.py deleted file mode 100644 index 34fc04c..0000000 --- a/src/jpg.py +++ /dev/null @@ -1,30 +0,0 @@ -import subprocess -import json - -import gi -gi.require_version('GdkPixbuf', '2.0') -from gi.repository import GdkPixbuf - -from . import abstract - -class JPGParser(abstract.AbstractParser): - mimetypes = {'image/jpg', } - meta_whitelist = {'SourceFile', 'ExifToolVersion', 'FileName', - 'Directory', 'FileSize', 'FileModifyDate', 'FileAccessDate', - "FileInodeChangeDate", 'FilePermissions', 'FileType', - 'FileTypeExtension', 'MIMEType', 'ImageWidth', - 'ImageSize', 'BitsPerSample', 'ColorComponents', 'EncodingProcess', - 'JFIFVersion', 'ResolutionUnit', 'XResolution', 'YCbCrSubSampling', - 'YResolution', 'Megapixels', 'ImageHeight'} - - def get_meta(self): - out = subprocess.check_output(['exiftool', '-json', self.filename]) - meta = json.loads(out.decode('utf-8'))[0] - for key in self.meta_whitelist: - meta.pop(key, None) - return meta - - def remove_all(self): - pixbuf = GdkPixbuf.Pixbuf.new_from_file(self.filename) - pixbuf.savev(self.output_filename, "jpeg", ["quality"], ["100"]) - return True diff --git a/tests/data/dirty.tiff b/tests/data/dirty.tiff new file mode 100644 index 0000000..05b97ad Binary files /dev/null and b/tests/data/dirty.tiff differ diff --git a/tests/test_libmat2.py b/tests/test_libmat2.py index 888c782..89e690e 100644 --- a/tests/test_libmat2.py +++ b/tests/test_libmat2.py @@ -6,7 +6,7 @@ import os import zipfile import tempfile -from src import pdf, png, jpg, audio, office, libreoffice, parser_factory +from src import pdf, png, images_pixbuf, audio, office, libreoffice, parser_factory class TestGetMeta(unittest.TestCase): def test_pdf(self): @@ -22,10 +22,17 @@ class TestGetMeta(unittest.TestCase): self.assertEqual(meta['ModifyDate'], "2018:03:20 21:59:25") def test_jpg(self): - p = jpg.JPGParser('./tests/data/dirty.jpg') + p = images_pixbuf.JPGParser('./tests/data/dirty.jpg') meta = p.get_meta() self.assertEqual(meta['Comment'], 'Created with GIMP') + def test_tiff(self): + p = images_pixbuf.JPGParser('./tests/data/dirty.tiff') + meta = p.get_meta() + self.assertEqual(meta['Make'], 'OLYMPUS IMAGING CORP.') + self.assertEqual(meta['Model'], 'C7070WZ') + self.assertEqual(meta['ModifyDate'], '2005:12:26 17:09:35') + def test_mp3(self): p = audio.MP3Parser('./tests/data/dirty.mp3') meta = p.get_meta() @@ -152,7 +159,7 @@ class TestCleaning(unittest.TestCase): def test_jpg(self): shutil.copy('./tests/data/dirty.jpg', './tests/data/clean.jpg') - p = jpg.JPGParser('./tests/data/clean.jpg') + p = images_pixbuf.JPGParser('./tests/data/clean.jpg') meta = p.get_meta() self.assertEqual(meta['Comment'], 'Created with GIMP') @@ -160,7 +167,7 @@ class TestCleaning(unittest.TestCase): ret = p.remove_all() self.assertTrue(ret) - p = jpg.JPGParser('./tests/data/clean.jpg.cleaned') + p = images_pixbuf.JPGParser('./tests/data/clean.jpg.cleaned') self.assertEqual(p.get_meta(), {}) os.remove('./tests/data/clean.jpg') @@ -240,3 +247,18 @@ class TestCleaning(unittest.TestCase): self.assertEqual(p.get_meta(), {}) os.remove('./tests/data/clean.odt') + + def test_tiff(self): + shutil.copy('./tests/data/dirty.tiff', './tests/data/clean.tiff') + p = images_pixbuf.TiffParser('./tests/data/clean.tiff') + + meta = p.get_meta() + self.assertEqual(meta['Model'], 'C7070WZ') + + ret = p.remove_all() + self.assertTrue(ret) + + p = images_pixbuf.TiffParser('./tests/data/clean.tiff.cleaned') + self.assertEqual(p.get_meta(), {}) + + os.remove('./tests/data/clean.tiff') -- cgit v1.3