def test_no_pageid_without_mets(self):
f = OcrdFile(None)
with self.assertRaisesRegex(Exception,
".*has no member 'mets' pointing.*"):
print(f.pageId)
with self.assertRaisesRegex(Exception,
".*has no member 'mets' pointing.*"):
f.pageId = 'foo'
def test_loctype(self):
f = OcrdFile(None)
self.assertEqual(f.loctype, 'OTHER')
self.assertEqual(f.otherloctype, 'FILE')
f.otherloctype = 'foo'
self.assertEqual(f.otherloctype, 'foo')
f.loctype = 'URN'
self.assertEqual(f.loctype, 'URN')
self.assertEqual(f.otherloctype, None)
f.otherloctype = 'foo'
self.assertEqual(f.loctype, 'OTHER')
def test_ocrd_file_eq(self):
mets = OcrdMets.empty_mets()
f1 = mets.add_file('FOO', ID='FOO_1', mimetype='image/tiff')
self.assertEqual(f1 == f1, True)
self.assertEqual(f1 != f1, False)
f2 = mets.add_file('FOO', ID='FOO_2', mimetype='image/tiff')
self.assertEqual(f1 == f2, False)
f3 = OcrdFile(None, ID='TEMP_1', mimetype='image/tiff')
f4 = OcrdFile(None, ID='TEMP_1', mimetype='image/tif')
# be tolerant of different equivalent mimetypes
self.assertEqual(f3 == f4, True)
f5 = mets.add_file('TEMP', ID='TEMP_1', mimetype='image/tiff')
self.assertEqual(f3 == f5, True)
def test_file_group_wo_parent_new_version():
"""Test for new error message
"""
with pytest.raises(
ValueError,
match=r"Must provide mets:file element this OcrdFile represent"):
OcrdFile(None)
def _resolve_image_as_pil(self, image_url, coords=None):
"""
Resolve an image URL to a PIL image.
Args:
- coords (list) : Coordinates of the bounding box to cut from the image
Returns:
Image or region in image as PIL.Image
"""
log = getLogger('ocrd.workspace._resolve_image_as_pil')
files = self.mets.find_files(url=image_url)
f = files[0] if files else OcrdFile(None, url=image_url)
image_filename = self.download_file(f).local_filename
with pushd_popd(self.directory):
pil_image = Image.open(image_filename)
pil_image.load() # alloc and give up the FD
if coords is None:
return pil_image
log.debug("Converting PIL to OpenCV: %s", image_url)
color_conversion = cv2.COLOR_GRAY2BGR if pil_image.mode in (
'1', 'L') else cv2.COLOR_RGB2BGR
pil_as_np_array = np.array(pil_image).astype(
'uint8') if pil_image.mode == '1' else np.array(pil_image)
cv2_image = cv2.cvtColor(pil_as_np_array, color_conversion)
poly = np.array(coords, np.int32)
log.debug("Cutting region %s from %s", coords, image_url)
region_cut = cv2_image[np.min(poly[:, 1]):np.max(poly[:, 1]),
np.min(poly[:, 0]):np.max(poly[:, 0])]
return Image.fromarray(region_cut)
def test_page_from_file(self):
f = OcrdFile(None,
mimetype='image/tiff',
local_filename=SAMPLE_IMG,
ID='file1')
self.assertEqual(f.mimetype, 'image/tiff')
p = page_from_file(f)
self.assertEqual(p.pcGtsId, f.ID)
self.assertEqual(p.get_Page().imageWidth, 1457)
def download_url(self, url, **kwargs):
"""
Download a URL to the workspace.
Args:
url (string): URL to download to directory
**kwargs : See :py:mod:`ocrd_models.ocrd_file.OcrdFile`
Returns:
The local filename of the downloaded file
"""
f = OcrdFile(None, url=url, **kwargs)
f = self.download_file(f)
return f.local_filename
def resolve_image_exif(self, image_url):
"""
Get the EXIF metadata about an image URL as :class:`OcrdExif`
Args:
image_url (string) : URL of image
Return
:class:`OcrdExif`
"""
f = next(self.mets.find_files(url=image_url), OcrdFile(None, url=image_url))
image_filename = self.download_file(f).local_filename
with Image.open(image_filename) as pil_img:
ocrd_exif = OcrdExif(pil_img)
return ocrd_exif
def resolve_image_exif(self, image_url):
"""
Get the EXIF metadata about an image URL as :class:`OcrdExif`
Args:
image_url (string) : URL of image
Return
:class:`OcrdExif`
"""
files = self.mets.find_files(url=image_url)
f = files[0] if files else OcrdFile(None, url=image_url)
image_filename = self.download_file(f).local_filename
if image_url not in self.image_cache['exif']:
# FIXME must be in the right directory
self.image_cache['exif'][image_url] = OcrdExif(
Image.open(image_filename))
return self.image_cache['exif'][image_url]
def test_page_from_file_unsupported_mimetype(self):
with self.assertRaisesRegex(ValueError, "Unsupported mimetype"):
page_from_file(
OcrdFile(None, local_filename=__file__, mimetype='foo/bar'))
def test_page_from_file_no_existe(self):
with self.assertRaisesRegex(FileNotFoundError,
"File not found: 'no-existe'"):
page_from_file(
OcrdFile(None, local_filename='no-existe', mimetype='foo/bar'))
def test_page_from_file_no_local_filename(self):
with self.assertRaisesRegex(
ValueError, "input_file must have 'local_filename' property"):
page_from_file(OcrdFile(None, mimetype='image/tiff'))
def test_page_from_file_page(self):
f = OcrdFile(None, mimetype=MIMETYPE_PAGE, local_filename=SAMPLE_PAGE)
p = page_from_file(f)
self.assertEqual(p.get_Page().imageWidth, 1457)
def test_fileGrp_wo_parent(self):
f = OcrdFile(None)
self.assertEqual(f.fileGrp, 'TEMP')
def test_set_url(self):
f = OcrdFile(None)
f.url = None
f.url = 'http://foo'
f.url = 'http://bar'
self.assertEqual(f.url, 'http://bar')
def test_file_group_wo_parent():
with pytest.raises(ValueError) as val_err:
OcrdFile(None)
assert "not related to METS" in str(val_err.value)
def test_extension(self):
f = OcrdFile(None, local_filename='/tmp/foo/bar/foo.bar')
self.assertEqual(f.extension, '.bar')
def test_basename(self):
f = OcrdFile(None, local_filename='/tmp/foo/bar/foo.bar')
self.assertEqual(f.basename, 'foo.bar')
def test_set_id_none(self):
f = OcrdFile(None)
f.ID = 'foo12'
self.assertEqual(f.ID, 'foo12')
f.ID = None
self.assertEqual(f.ID, 'foo12')
def test_constructor_url(self):
f = OcrdFile(None, url="foo/bar")
self.assertEqual(f.url, 'foo/bar')
self.assertEqual(f.local_filename, 'foo/bar')
def _resolve_image_as_pil(self, image_url, coords=None):
"""
Resolve an image URL to a PIL image.
Args:
- coords (list) : Coordinates of the bounding box to cut from the image
Returns:
Image or region in image as PIL.Image
"""
log = getLogger('ocrd.workspace._resolve_image_as_pil')
f = next(self.mets.find_files(url=image_url),
OcrdFile(None, url=image_url))
image_filename = self.download_file(f).local_filename
with pushd_popd(self.directory):
pil_image = Image.open(image_filename)
pil_image.load() # alloc and give up the FD
# Pillow does not properly support higher color depths
# (e.g. 16-bit or 32-bit or floating point grayscale),
# clipping its dynamic range to the lower 8-bit in
# many operations (including paste, putalpha, ImageStat...),
# even including conversion.
# Cf. Pillow#3011 Pillow#3159 Pillow#3838 (still open in 8.0)
# So to be on the safe side, we must re-quantize these
# to 8-bit via numpy (conversion to/from which fortunately
# seems to work reliably):
if (pil_image.mode.startswith('I') or pil_image.mode.startswith('F')):
arr_image = np.array(pil_image)
if arr_image.dtype.kind == 'i':
# signed integer is *not* trustworthy in this context
# (usually a mistake in the array interface)
log.debug('Casting image "%s" from signed to unsigned',
image_url)
arr_image.dtype = np.dtype('u' + arr_image.dtype.name)
if arr_image.dtype.kind == 'u':
# integer needs to be scaled linearly to 8 bit
# of course, an image might actually have some lower range
# (e.g. 10-bit in I;16 or 20-bit in I or 4-bit in L),
# but that would be guessing anyway, so here don't
# make assumptions on _scale_, just reduce _precision_
log.debug('Reducing image "%s" from depth %d bit to 8 bit',
image_url, arr_image.dtype.itemsize * 8)
arr_image = arr_image >> 8 * (arr_image.dtype.itemsize - 1)
arr_image = arr_image.astype(np.uint8)
elif arr_image.dtype.kind == 'f':
# float needs to be scaled from [0,1.0] to [0,255]
log.debug('Reducing image "%s" from floating point to 8 bit',
image_url)
arr_image *= 255
arr_image = arr_image.astype(np.uint8)
pil_image = Image.fromarray(arr_image)
if coords is None:
return pil_image
# FIXME: remove or replace this by (image_from_polygon+) crop_image ...
log.debug("Converting PIL to OpenCV: %s", image_url)
color_conversion = cv2.COLOR_GRAY2BGR if pil_image.mode in (
'1', 'L') else cv2.COLOR_RGB2BGR
pil_as_np_array = np.array(pil_image).astype(
'uint8') if pil_image.mode == '1' else np.array(pil_image)
cv2_image = cv2.cvtColor(pil_as_np_array, color_conversion)
poly = np.array(coords, np.int32)
log.debug("Cutting region %s from %s", coords, image_url)
region_cut = cv2_image[np.min(poly[:, 1]):np.max(poly[:, 1]),
np.min(poly[:, 0]):np.max(poly[:, 0])]
return Image.fromarray(region_cut)
def test_basename_without_extension_tar(self):
f = OcrdFile(None, local_filename='/tmp/foo/bar/foo.tar.gz')
self.assertEqual(f.basename_without_extension, 'foo')
def test_basename_from_url(self):
f = OcrdFile(None, url="http://foo.bar/quux")
self.assertEqual(f.basename, 'quux')
def test_page_from_file_unsupported_mimetype(self):
with self.assertRaisesRegex(Exception, "Unsupported mimetype"):
page_from_file(OcrdFile(None, mimetype='foo/bar'))