def _convert_image(self, file_path: str, preview_dims: ImgDims) -> Image:
"""
refer: https://legacy.imagemagick.org/Usage/thumbnails/
like cmd: convert -layers merge -background white -thumbnail widthxheight \
-auto-orient -quality 85 -interlace plane input.jpeg output.jpeg
"""
img = Image(filename=file_path)
resize_dim = compute_resize_dims(dims_in=ImgDims(width=img.width,
height=img.height),
dims_out=preview_dims)
img.auto_orient()
img.iterator_reset()
img.background_color = Color("white")
img.merge_layers("merge")
if self.progressive:
img.interlace_scheme = "plane"
img.compression_quality = self.quality
img.thumbnail(resize_dim.width, resize_dim.height)
return img
def get_thumbnail_from_pdf(self, file):
try:
filename = file.name
img = None
# Convert PDF files
imgs_pdf = Image(file=file)
imgs = imgs_pdf.convert('jpeg')
if imgs:
img = Image(image=imgs.sequence[0])
img.background_color = Color('white')
img.alpha_channel = 'remove'
# resized and save the converted file
img.transform(crop='', resize=THUMBNAIL_SIZE)
img.thumbnail()
temp = NamedTemporaryFile(delete=False)
temp.flush()
temp0 = File(temp)
with temp0.open('wb') as f:
img.save(file=f)
return temp0.open('rb')
except Exception as e:
print(repr(e))
return None
def convex(img: Image):
resize(img)
img.virtual_pixel = "transparent"
img.background_color = Color("white")
img.distort(method="barrel", arguments=[1, 0, 0, 1])
return img
def _run_convert(filename, page, res=120):
idx = page + 1
temp_time = time.time() * 1000
# 由于每次转换的时候都需要重新将整个PDF载入内存，所以这里使用内存缓存
pdfile = getPdfReader(filename)
pageObj = pdfile.getPage(page)
dst_pdf = PdfFileWriter()
dst_pdf.addPage(pageObj)
pdf_bytes = io.BytesIO()
dst_pdf.write(pdf_bytes)
pdf_bytes.seek(0)
img = Image(file=pdf_bytes, resolution=res)
img.format = 'png'
img.compression_quality = 90
img.background_color = Color("white")
# 保存图片
img_path = '%s%d.png' % (filename[:filename.rindex('.')], idx)
img.save(filename=img_path)
img.destroy()
img = None
pdf_bytes = None
dst_pdf = None
print('convert page %d cost time %d' % (idx, (time.time() * 1000 - temp_time)))
def _generate_thumbnail_image_content_file(document):
content = None
if document.file_on_server:
content = document.unique_file.file_field.read()
else:
with requests.request('get', document.external_url,
stream=True) as response:
content = response.content
temp_pdf_path = os.path.join(settings.MEDIA_ROOT, 'document_thumbnails',
'temp.pdf')
with open(temp_pdf_path, 'w+') as f:
f.write(content)
reader = PdfReader(temp_pdf_path)
if len(reader.pages) > 1:
page = reader.pages[0]
writer = PdfWriter()
writer.addpage(page)
writer.write(temp_pdf_path)
images = Image(filename=temp_pdf_path, resolution=38)
images.background_color = Color('white')
images.alpha_channel = 'flatten'
os.remove(temp_pdf_path)
return ContentFile(images.make_blob('jpg'))
def pdf2image(self,dest_width, dest_height):
RESOLUTION = 300
for attachment in self:
img = Image(blob=attachment.datas.decode('base64'),resolution=(RESOLUTION,RESOLUTION))
img.background_color = Color('white')
#img.resize(dest_width,dest_height)
attachment.image = img.make_blob(format='jpg').encode('base64')
def createAndStoreThumbnail(**kwargs):
# fileDirectory = "/home/akshat/Desktop/"
# inFileName="test_doc_2.pdf"
# outFileName="myOutputfile.jpg"
imageFromPdf = Image(filename=kwargs["fileLocation"] + kwargs["fileName"])
# pages = len(imageFromPdf.sequence)
# print(pages)
# creates an empty Image.
image = Image(width=imageFromPdf.width, height=imageFromPdf.height)
# resize the empty image
image.sample(470, 330)
#superimpose on the empty image the argument given, at the position specified
image.composite(imageFromPdf.sequence[0], top=0, left=0)
image.background_color = Color("white")
image.alpha_channel = 'remove'
# for i in range(pages):
# image.composite(
# imageFromPdf.sequence[i],
# top=imageFromPdf.height * i,
# left=0
# )
image.format = "jpg"
image.save(filename=kwargs["imageLocation"] + kwargs["imageName"] + ".jpg")
return kwargs["imageName"] + ".jpg"
# display(image)
def _run_convert_all(filename, res=120):
'''把pdf所有页面转化为图片'''
# 由于每次转换的时候都需要重新将整个PDF载入内存，所以这里使用内存缓存
pdfile, f = get_pdf_reader(filename), open(filename, "rb")
# pdfile,f = get_pdf_reader2(filename)
for i in range(0, pdfile.getNumPages()):
temp_time = time.time() * 1000
pageobj = pdfile.getPage(i)
dst_pdf = PdfFileWriter()
dst_pdf.addPage(pageobj)
pdf_bytes = io.BytesIO()
dst_pdf.write(pdf_bytes)
pdf_bytes.seek(0)
img = Image(file=pdf_bytes, resolution=res)
img.format = 'png'
img.compression_quality = 90
img.background_color = Color("white")
# 保存图片
img_path = 'dest/%s_pg%d.png' % (filename[:filename.rindex('.')],
i + 1)
img.save(filename=img_path)
img.destroy()
img, pdf_bytes, dst_pdf = None, None, None
print('convert page %d cost time %dms' %
(i + 1, (time.time() * 1000 - temp_time)))
f.close()
def _run_convert(filename, page, res=120):
'''把pdf指定页码转化为图片'''
pdfile, f = get_pdf_reader(filename), open(filename, "rb")
# pdfile,f = get_pdf_reader2(filename)
if page <= pdfile.getNumPages():
idx = page + 1
temp_time = time.time() * 1000
# 由于每次转换的时候都需要重新将整个PDF载入内存，所以这里使用内存缓存
pageobj = pdfile.getPage(page)
dst_pdf = PdfFileWriter()
dst_pdf.addPage(pageobj)
pdf_bytes = io.BytesIO()
dst_pdf.write(pdf_bytes)
pdf_bytes.seek(0)
img = Image(file=pdf_bytes, resolution=res)
img.format = 'png'
img.compression_quality = 90
img.background_color = Color("white")
# 保存图片
img_path = 'dest/%s_pg%d.png' % (filename[:filename.rindex('.')], idx)
img.save(filename=img_path)
img.destroy()
img, pdf_bytes, dst_pdf = None, None, None
print('convert page %d cost time %dms' %
(idx, (time.time() * 1000 - temp_time)))
else:
print("pg%r list index out of range" % page)
f.close()
def _pdf_thumbnail(filename):
img = WandImage(filename=filename + '[0]')
img.background_color = Color('white')
tw, th = get_thumbnail_size(img.height, img.width, 50, 50)
img.resize(tw, th)
rawData = img.make_blob('jpeg')
return base64.b64encode(rawData)
def _run_convert(filename, page, res=120):
idx = page + 1
temp_time = time.time() * 1000
# 由于每次转换的时候都需要重新将整个PDF载入内存，所以这里使用内存缓存
pdfile = getPdfReader(filename)
pageObj = pdfile.getPage(page)
dst_pdf = PdfFileWriter()
dst_pdf.addPage(pageObj)
pdf_bytes = io.BytesIO()
dst_pdf.write(pdf_bytes)
pdf_bytes.seek(0)
img = Image(file=pdf_bytes, resolution=res)
img.format = 'png'
img.compression_quality = 90
img.background_color = Color('white')
img.alpha_channel = 'remove'
#img.channels = 'rgb_channels'
# 保存图片
#filename[:filename.rindex('.')]
img_path = './image/%s.png' % (filename[filename.rindex('/pdf/') +
5:filename.rindex('.pdf')])
#)
img.save(filename=img_path)
img.destroy()
img = None
pdf_bytes = None
dst_pdf = None
def image_labels_to_countable_npy():
_user = User.objects.filter(username='******')[0]
_labeler = Labeler.objects.filter(user=_user)[0]
labels = ImageLabel.objects.filter(labeler=_labeler)
foldername = 'npy'
for label in labels:
parent_image = label.parentImage
filename = '%s' % parent_image.name.replace('.JPG', '')
outputFilenameNpy = (settings.STATIC_ROOT +
settings.LABEL_FOLDER_NAME + foldername + '/' +
filename + '.npy')
categorylabels = label.categorylabel_set.all()
height = parent_image.height
width = parent_image.width
total_paths = 254
masks_ndarray = np.zeros((total_paths, height, width))
ctr = 0
for cat_id, categorylabel in enumerate(categorylabels):
svg = categorylabel.labelShapes
paths = []
poly = []
paths = re.findall(SVGRegex.rePath, svg)
poly = re.findall(SVGRegex.rePolygon, svg)
shapes = paths + poly
if len(paths) + len(poly) > 0:
for idx, path in enumerate(shapes):
print(ctr, cat_id, idx, path)
img = WandImage(blob=image_string_to_SVG_string_file(
image_label_string_to_SVG_string(path, height, width)))
img.resize(width, height)
img.background_color = WandColor('white')
img.alpha_channel = 'remove'
img.negate()
img.threshold(0)
img.format = 'png'
if not os.path.exists(settings.STATIC_ROOT +
settings.LABEL_FOLDER_NAME +
foldername):
os.makedirs(settings.STATIC_ROOT +
settings.LABEL_FOLDER_NAME + foldername)
outputFilename = (settings.STATIC_ROOT +
settings.LABEL_FOLDER_NAME + foldername +
'/' + filename + '_' + str(idx) + '_' +
str(ctr) + IMAGE_FILE_EXTENSION)
img.save(filename=outputFilename)
im = imageio.imread(outputFilename)
masks = np.array(im)
category_id = categorylabel.categoryType_id
cat_mask = np.where(masks == 255, category_id, masks)
masks_ndarray[ctr, :, :] = cat_mask
ctr = ctr + 1
else:
print(ctr, cat_id, 0, 'EMPTY')
masks_ndarray.resize(ctr, height, width)
print(masks_ndarray.shape)
np.save(outputFilenameNpy, masks_ndarray)
def pdf2image(self,dest_width, dest_height):
RESOLUTION = 300
#blob = self.datas.decode('base64')
#raise Warning(self.base64_decode(self.datas))
#str = self.datas + '=' *(-len(self.datas)%4)
#img = Image(blob=self[0].datas.decode('base64'))
#img.resize(dest_width,dest_height)
#~ self[0].image = img.make_blob(format='jpg').encode('base64')
img = Image(blob=self[0].datas.decode('base64'),resolution=(RESOLUTION,RESOLUTION))
img.background_color = Color('white')
self[0].image = img.make_blob(format='jpg').encode('base64')
def pagify_pdf(fname):
"""Create a PNG image for each page of a PDF document"""
all_pages = Image(filename=fname+'.pdf', resolution=250)
count = 1
for page in all_pages.sequence:
img = Image(page)
img.format = 'png'
img.background_color = Color('white')
img.alpha_channel = 'remove'
img.save(filename=fname+str(count)+'.png')
count += 1
return count - 1
def makeKmzFile(self,filename,title):
'''Produce the MapRun map, as a georeferenced JPG image in a KMZ file.'''
if self.aspect == "landscape":
ss = "style=streeto|paper=0.297,0.210|scale=10000|centre="
else:
ss = "style=streeto|paper=0.210,0.297|scale=10000|centre="
mp = self.getMidpoint().getEPSG3857()
ss += str(int(mp[1])) + "," + str(int(mp[0])) + "|title="+title+"|club=|mapid=|start="
ss += str(int(self.startFinish.getEPSG3857()[1])) + "," + str(int(self.startFinish.getEPSG3857()[0]))
ss += "|crosses=|controls="
#TODO Remove controls from request for actual MapRun map
ss += ",".join([str(control.getNumber()) + ",45," + str(int(control.getEPSG3857()[1])) + "," + str(
int(control.getEPSG3857()[0])) for control in self.getControls()])
r = requests.get(PDF_SERVER+"/pdf/?"+ss)
pdfFile = tempfile.NamedTemporaryFile(delete=False)
for chunk in r.iter_content(chunk_size=128):
pdfFile.write(chunk)
pdfFile.close()
with Image(filename=pdfFile.name, resolution=180) as img_pdf:
img = Image(image=img_pdf.sequence[0])
img.background_color = Color('white')
img.alpha_channel = 'remove'
img.format = "jpeg"
img.save(filename=pdfFile.name+".jpg")
kml = Element("kml", {"xmlns": "http://www.opengis.net/kml/2.2"})
document = SubElement(kml, "Document")
folder = SubElement(document, "Folder")
groundOverlay = SubElement(folder, "GroundOverlay")
name=SubElement(groundOverlay,"name")
name.text="tile.jpg"
drawOrder=SubElement(groundOverlay,"drawOrder")
drawOrder.text="75"
icon=SubElement(groundOverlay,"Icon")
href=SubElement(icon,"href")
href.text="files/tile.jpg"
latLonBox=SubElement(groundOverlay,"LatLonBox")
north=SubElement(latLonBox,"north")
north.text=str(self.topLeft.getLat())
south=SubElement(latLonBox,"south")
south.text=str(self.bottomRight.getLat())
east=SubElement(latLonBox,"east")
east.text=str(self.bottomRight.getLon())
west=SubElement(latLonBox,"west")
west.text=str(self.topLeft.getLon())
rotation=SubElement(latLonBox,"rotation")
rotation.text="0.0"
docKmlFile = tempfile.NamedTemporaryFile(delete=False)
docKmlFile.write(tostring(kml))
docKmlFile.close()
with zipfile.ZipFile(filename,"w") as kmzFile:
kmzFile.write(docKmlFile.name,"doc.kml")
kmzFile.write(pdfFile.name+".jpg","files/tile.jpg")
def convert_pdf_to_png(filenameVal):
pdf = Image(filename=filenameVal, resolution=200)
pages = len(pdf.sequence)
image = Image(width=pdf.width, height=pdf.height * pages)
for i in xrange(pages):
image.composite(pdf.sequence[i], top=pdf.height * i, left=0)
image.background_color = Color("white")
image.alpha_channel = 'remove'
image.save(filename="out.png")
return remove_points_from_image(filenameVal)
def floor(img: Image):
resize(img, 128)
img.alpha_channel = False
img.background_color = Color("light-blue")
img.virtual_pixel = "tile"
img.distort(
method="perspective",
arguments=[0, 0, 20, 61, 90, 0, 70, 63, 0, 90, 0, 83, 90, 90, 85, 88])
resize(img, 512)
return img
def _floor(img: Image):
x, y = img.size
img.alpha_channel = False
img.background_color = Color("#81cfe0")
img.virtual_pixel = "tile"
img.distort(
method="perspective",
arguments=[0, 0, 20, 61, 90, 0, 70, 63, 0, 90, 0, 83, 90, 90, 85, 88])
img.resize(x, y)
return img
def toPNG(filePath, xResScale=220, yResScale=220):
# converts a pdf to GIF
# for each page, say og file is Goldenrod_Full.pdf
# output will be Goldenrod_Full-0.png, Goldenrod_Full-1.png
img = Image(filename=filePath, resolution=400)
if len(img.sequence) == 1:
# whole file as 1 page for now
img.format = 'png'
img.background_color = Color('white')
img.alpha_channel = 'remove'
# image resolution scale adjustment for display
img.resample(x_res=xResScale, y_res=yResScale)
# saving
img.save(filename=filePath[:-4] + '.png')
else: # multipage image
for i in range(len(img.sequence)):
indiv = Image(image=img.sequence[i], resolution=400)
indiv.format = 'png'
indiv.background_color = Color('white')
indiv.alpha_channel = 'remove'
# adjust resolution for display
indiv.resample(x_res=xResScale, y_res=yResScale)
# save it
indiv.save(filename=filePath[:-4] + '-' + str(i) + '.png')
# returns number of pages converted
return len(img.sequence)
def toGIF(filePath, xResScale=220, yResScale=220):
# converts a pdf to PNG
# beware when converting to multiple pages, it might become an
# animated GIF
img = Image(filename=filePath, resolution=400)
if len(img.sequence) == 1:
# whole file as 1 page for now
img.format = 'gif'
img.background_color = Color('white')
img.alpha_channel = 'remove'
# image resolution scale adjustment for display
img.resample(x_res=xResScale, y_res=yResScale)
# saving
img.save(filename=filePath[:-4] + '.gif')
else: # multipage image
for i in range(len(img.sequence)):
indiv = Image(image=img.sequence[i], resolution=400)
indiv.format = 'gif'
indiv.background_color = Color('white')
indiv.alpha_channel = 'remove'
# adjust resolution for display
indiv.resample(x_res=xResScale, y_res=yResScale)
# save it
indiv.save(filename=filePath[:-4] + '-' + str(i) + '.gif')
# returns number of pages converted
return len(img.sequence)
def resize_image(path, width, height):
with WAND_LOCK.reader_lock, Image(width=width, height=height) as bg:
# Resize our input image.
with Image(filename=path, resolution=300) as s:
d = Image(s.sequence[0])
d.background_color = Color("white")
d.alpha_channel = 'remove'
d.transform(resize='%ix%i>' % (width, height))
# Offset input image on top of background.
left = (bg.width - d.width) // 2
top = (bg.height - d.height) // 2
bg.composite(d, left, top, operator='over')
with tempfile.NamedTemporaryFile(delete=False, suffix='.gif') as t:
bg.save(filename=t.name)
return t.name
def _create_preview(self):
"""
create the PNG file for this pub.
This is called internally after PDF generation, because there is a dependency.
"""
pdf_filename = opj(
self.working_dir,
"pub2/{0}.pdf".format(self.file.preamble["identifier"]))
preview_filename = opj(
self.working_dir,
"pub2/{0}.png".format(self.file.preamble["identifier"]))
img = Image(filename='{0}[0]'.format(pdf_filename))
img.background_color = Color("white")
img.merge_layers('flatten')
img.format = 'png'
img.save(filename=preview_filename)
def ocr(invoice_path):
option = {}
try:
img = Image(blob=invoice_path.read_bytes(), resolution=300)
img.format = 'png'
img.background_color = Color("white")
img_bytes = img.make_blob()
response = ocrClient.request_ocr_service_base64(
req_uri, img_bytes, option)
if response.status_code != 200:
print("Status code:" + str(response.status_code) + "\ncontent:" +
response.text)
return response.json()["result"]
except ValueError as e:
print(e)
def writeImageNoWarp(pdf, out, resolution=pdfDense, trim=True):
with Image(filename=pdf, resolution=resolution) as pdf:
pages = len(pdf.sequence)
height = 0
width = 0
for j in range(pages):
## Go through the sequence and trim them individually
with pdf.sequence.index_context(j):
if trim:
pdf.trim()
height += pdf.height + 10
if (width < pdf.width):
width = pdf.width
else:
width = pdf.width
height = pdf.height
i = Image(width=width, height=height)
height = [r.height for r in pdf.sequence]
height = height[::-1]
height.append(0)
height = height[::-1]
width = [math.floor((width - r.width) / 2) for r in pdf.sequence]
for j in range(pages):
i.composite(pdf.sequence[j], top=height[j] + j * 10, left=width[j])
i.background_color = Color('white') # Set white background.
i.alpha_channel = 'remove'
if trim: ## This is necessary because of stacked charts (above)
i.trim()
i.sharpen(radius=5.0, sigma=5.0)
i.normalize()
i.quantize(16, dither=False)
i.compression_quality = 00
i.save(filename=out.replace(".png", ".png8"))
tmp = out.replace(".png", "") + "*"
#out = glob.glob(tmp)
if (os.system("rename .png8 .png %s*" % tmp)):
print("Error in saving file.")
if (os.system("optipng -quiet %s*" % tmp)):
print("Error in optipng.")
commstr = 'cwebp -quiet -lossless -z 9 -metadata exif %s -o %s' % (
out, out.replace(".png", ".webp"))
if (os.system(commstr)):
print("Error in cwebp")
def convert(filename, page=0, res=120):
reader = PdfFileReader(filename, strict=False)
page_obj = reader.getPage(page)
dst_pdf = PdfFileWriter()
dst_pdf.addPage(page_obj)
pdf_bytes = io.BytesIO()
dst_pdf.write(pdf_bytes)
pdf_bytes.seek(0)
img = Image(file=pdf_bytes, resolution=res)
img.format = 'png'
img.compression_quality = 90
img.background_color = Color('white')
img_path = filename.replace('pdf', 'png')
img.save(filename=img_path)
img.destroy()
def read_image(image_path : str, dpi=300):
assert os.path.exists(image_path)
image_list = []
if image_path.endswith(".pdf"):
with WandImage(filename=image_path, resolution=dpi) as img:
for page_img_seq in img.sequence:
page_img = WandImage(page_img_seq)
page_img.background_color = Color('white')
page_img.alpha_channel = 'remove'
img_buffer = np.asarray(bytearray(page_img.make_blob(format='jpeg')), dtype=np.uint8)
if img_buffer is not None:
image_list.append(imdecode(img_buffer, IMREAD_COLOR))
elif image_path.endswith(".png") or image_path.endswith(".jpg") or image_path.endswith(".jpeg"):
image_list.append(imread(image_path, IMREAD_COLOR))
else:
sys.exit("Unknown input file format. Accepted inputs: .pdf, .jpg, .jpeg, .png")
return image_list
def convert(filename, res=120):
pdf_file = PdfFileReader(open(filename, "rb"))
pageObj = pdf_file.getPage(0)
dst_pdf = PdfFileWriter()
dst_pdf.addPage(pageObj)
pdf_bytes = io.BytesIO()
dst_pdf.write(pdf_bytes)
pdf_bytes.seek(0)
img_filename = filename.replace('old', 'img').replace('pdf', 'png')
img = Image(file=pdf_bytes, resolution=res)
img.format = 'png'
img.compression_quality = 120
img.background_color = Color("white")
img.save(filename=img_filename)
img.destroy()
def _convert_pdf2img(self, file, store_path):
pdf_file = PdfFileReader(file, strict=False)
pageObj = pdf_file.getPage(0)
dst_pdf = PdfFileWriter()
dst_pdf.addPage(pageObj)
pdf_bytes = io.BytesIO()
dst_pdf.write(pdf_bytes)
pdf_bytes.seek(0)
img = Image(file=pdf_bytes, resolution=200)
img.format = 'jpg'
img.compression_quality = 90
img.background_color = Color("white")
img_path = store_path + '%s.jpg' % (uuid.uuid3(uuid.uuid1(), file))
print(img_path)
img.save(filename=img_path)
img.destroy()
return img_path
def _run_convert(filename, page, res=120):
idx = page + 1
pdfile = getPdfReader(filename)
pageObj = pdfile.getPage(page)
dst_pdf = PdfFileWriter()
dst_pdf.addPage(pageObj)
pdf_bytes = io.BytesIO()
dst_pdf.write(pdf_bytes)
pdf_bytes.seek(0)
img = Image(file=pdf_bytes, resolution=res)
img.format = 'jpg'
img.compression_quality = 90
img.background_color = Color("white")
img_path = '%s%d.jpg' % (filename[:filename.rindex('.')], idx)
img.save(filename=img_path)
img.destroy()
def _run_convert(pdfile, savedfilename, page_index, index, res=120):
pageObj = pdfile.getPage(page_index) #获取pdf的第page_index页
dst_pdf = PdfFileWriter()
dst_pdf.addPage(pageObj)
pdf_bytes = io.BytesIO()
dst_pdf.write(pdf_bytes)
pdf_bytes.seek(0)
img = Image(file=pdf_bytes, resolution=res)
img.format = 'png'
img.compression_quality = 90
img.background_color = Color("white")
img_path = '%s%04d.jpg' % (savedfilename, index)
img.save(filename=img_path)
print(img_path)
img.destroy()
pages = 1
image = Image(
width = imageFromPdf.width,
height = imageFromPdf.height*pages
)
for i in range(pages):
image.composite(
imageFromPdf.sequence[i],
top = imageFromPdf.height * i,
left = 0
)
image.resize(250,250)
image.alpha_channel = False
image.format = 'png'
print(image.size)
image.background_color = Color('pink')
image.type = 'grayscale'
image.caption = file.split('.')[0]
image.save(filename = fileDirectory+file.split('.')[0]+".png")
image.clear()
image.close()
#display(image)
