def gather_elements(self, client, node, style):
try:
for entry in node["entries"]:
client.pending_targets.append(docutils.nodes.make_id(entry[2]))
except IndexError:
if node["entries"]:
log.error("Can't process index entry: %s [%s]", node["entries"], nodeid(node))
return []
def gather_elements(self, client, node, style):
try:
client.pending_targets.append(node['entries'][0][2])
except IndexError:
if node['entries']:
log.error("Can't process index entry: %s [%s]",
node['entries'], nodeid(node))
return []
def gather_elements(self, client, node, style):
try:
for entry in node['entries']:
client.pending_targets.append(docutils.nodes.make_id(entry[2]))
except IndexError:
if node['entries']:
log.error("Can't process index entry: %s [%s]",
node['entries'], nodeid(node))
return []
def parseRaw(data, node):
"""Parse and process a simple DSL to handle creation of flowables.
Supported (can add others on request):
* PageBreak
* EvenPageBreak
* OddPageBreak
* FrameBreak
* Spacer width, height
* Transition
* SetPageCounter
* TextAnnotation "text of annotation" x_begin=-1 y_begin=-1 x_end=-1 y_end=-1
"""
elements = []
lines = data.splitlines()
for line in lines:
lexer = shlex.shlex(line)
lexer.whitespace += ','
tokens = list(lexer)
if not tokens:
continue # Empty line
command = tokens[0]
if command == 'PageBreak':
if len(tokens) == 1:
elements.append(MyPageBreak())
else:
elements.append(MyPageBreak(tokens[1]))
elif command == 'EvenPageBreak':
if len(tokens) == 1:
elements.append(MyPageBreak(breakTo='even'))
else:
elements.append(MyPageBreak(tokens[1], breakTo='even'))
elif command == 'OddPageBreak':
if len(tokens) == 1:
elements.append(MyPageBreak(breakTo='odd'))
else:
elements.append(MyPageBreak(tokens[1], breakTo='odd'))
elif command == 'FrameBreak':
if len(tokens) == 1:
elements.append(CondPageBreak(99999))
else:
elements.append(CondPageBreak(float(tokens[1])))
elif command == 'Spacer':
elements.append(
MySpacer(adjustUnits(tokens[1]), adjustUnits(tokens[2])))
elif command == 'Transition':
elements.append(Transition(*tokens[1:]))
elif command == 'SetPageCounter':
elements.append(flowables.PageCounter(*tokens[1:]))
elif command == 'TextAnnotation':
elements.append(TextAnnotation(*tokens[1:]))
else:
log.error('Unknown command %s in raw pdf directive [%s]' %
(command, nodeid(node)))
return elements
def log_unknown(self, node, during):
if not hasattr(self, "unkn_node"):
self.unkn_node = set()
cln = self.getclassname(node)
if not cln in self.unkn_node:
self.unkn_node.add(cln)
log.warning("Unkn. node (self.%s): %s [%s]", during, cln, nodeid(node))
try:
log.debug(node)
except (UnicodeDecodeError, UnicodeEncodeError):
log.debug(repr(node))
def log_unknown(self, node, during):
if not hasattr(self, 'unkn_node'):
self.unkn_node = set()
cln = self.getclassname(node)
if not cln in self.unkn_node:
self.unkn_node.add(cln)
log.warning("Unkn. node (self.%s): %s [%s]", during, cln,
nodeid(node))
try:
log.debug(node)
except (UnicodeDecodeError, UnicodeEncodeError):
log.debug(repr(node))
def getstyle(self, client, node, style):
try:
if node["classes"] and node["classes"][0]:
# FIXME: Supports only one class, sorry ;-)
if client.styles.StyleSheet.has_key(node["classes"][0]):
style = client.styles[node["classes"][0]]
else:
log.info("Unknown class %s, ignoring. [%s]", node["classes"][0], nodeid(node))
except TypeError: # Happens when a docutils.node.Text reaches here
pass
if style is None or style == client.styles["bodytext"]:
style = client.styles.styleForNode(node)
return style
def parseRaw(data, node):
"""Parse and process a simple DSL to handle creation of flowables.
Supported (can add others on request):
* PageBreak
* Spacer width, height
"""
elements = []
lines = data.splitlines()
for line in lines:
lexer = shlex.shlex(line)
lexer.whitespace += ','
tokens = list(lexer)
if not tokens:
continue # Empty line
command = tokens[0]
if command == 'PageBreak':
if len(tokens) == 1:
elements.append(MyPageBreak())
else:
elements.append(MyPageBreak(tokens[1]))
elif command == 'EvenPageBreak':
if len(tokens) == 1:
elements.append(MyPageBreak(breakTo='even'))
else:
elements.append(MyPageBreak(tokens[1],breakTo='even'))
elif command == 'OddPageBreak':
if len(tokens) == 1:
elements.append(MyPageBreak(breakTo='odd'))
else:
elements.append(MyPageBreak(tokens[1],breakTo='odd'))
elif command == 'FrameBreak':
if len(tokens) == 1:
elements.append(CondPageBreak(99999))
else:
elements.append(CondPageBreak(float(tokens[1])))
elif command == 'Spacer':
elements.append(MySpacer(adjustUnits(tokens[1]),
adjustUnits(tokens[2])))
elif command == 'Transition':
elements.append(Transition(*tokens[1:]))
elif command == 'SetPageCounter':
elements.append(flowables.PageCounter(*tokens[1:]))
else:
log.error('Unknown command %s in raw pdf directive [%s]'%(command,nodeid(node)))
return elements
def getstyle(self, client, node, style):
try:
if node['classes'] and node['classes'][0]:
# FIXME: Supports only one class, sorry ;-)
if client.styles.StyleSheet.has_key(node['classes'][0]):
style = client.styles[node['classes'][0]]
else:
log.info("Unknown class %s, ignoring. [%s]",
node['classes'][0], nodeid(node))
except TypeError: # Happens when a docutils.node.Text reaches here
pass
if style is None or style == client.styles['bodytext']:
style = client.styles.styleForNode(node)
return style
def findsubclass(self, node, during):
handlerinfo = '%s.%s' % (self.getclassname(self), during)
log.debug("%s: %s", handlerinfo, self.getclassname(node))
log.debug("%s: [%s]", handlerinfo, nodeid(node))
try:
log.debug("%s: %s", handlerinfo, node)
except (UnicodeDecodeError, UnicodeEncodeError):
log.debug("%s: %r", handlerninfo, node)
log.debug("")
# Dispatch to the first matching class in the MRO
dispatchdict = self.dispatchdict
for baseclass in inspect.getmro(node.__class__):
result = dispatchdict.get(baseclass)
if result is not None:
break
else:
self.log_unknown(node, during)
result = self
return result
def findsubclass(self, node, during):
handlerinfo = '%s.%s' % (self.getclassname(self), during)
log.debug("%s: %s", handlerinfo, self.getclassname(node))
log.debug("%s: [%s]", handlerinfo, nodeid(node))
try:
log.debug("%s: %s", handlerinfo, node)
except (UnicodeDecodeError, UnicodeEncodeError):
log.debug("%s: %r", handlerninfo, node)
log.debug("")
# Dispatch to the first matching class in the MRO
dispatchdict = self.dispatchdict
for baseclass in inspect.getmro(node.__class__):
result = dispatchdict.get(baseclass)
if result is not None:
break
else:
self.log_unknown(node, during)
result = self
return result
def size_for_node(self, node, client):
'''Given a docutils image node, returns the size the image should have
in the PDF document, and what 'kind' of size that is.
That involves lots of guesswork'''
uri = os.path.join(client.basedir, str(node.get("uri")))
srcinfo = client, uri
# Extract all the information from the URI
imgname, extension, options = self.split_uri(uri)
if not os.path.isfile(imgname):
imgname = missing
scale = float(node.get('scale', 100)) / 100
size_known = False
# Figuring out the size to display of an image is ... annoying.
# If the user provides a size with a unit, it's simple, adjustUnits
# will return it in points and we're done.
# However, often the unit wil be "%" (specially if it's meant for
# HTML originally. In which case, we will use a percentage of
# the containing frame.
# Find the image size in pixels:
kind = 'direct'
xdpi, ydpi = client.styles.def_dpi, client.styles.def_dpi
extension = imgname.split('.')[-1].lower()
if extension in ['svg', 'svgz'] and SVGImage.available():
iw, ih = SVGImage(imgname, srcinfo=srcinfo).wrap(0, 0)
# These are in pt, so convert to px
iw = iw * xdpi / 72
ih = ih * ydpi / 72
elif extension in [
"ai", "ccx", "cdr", "cgm", "cmx", "sk1", "sk", "xml", "wmf",
"fig"
] and VectorImage.available():
iw, ih = VectorImage(imgname, srcinfo=srcinfo).wrap(0, 0)
# These are in pt, so convert to px
iw = iw * xdpi / 72
ih = ih * ydpi / 72
elif extension == 'pdf':
if VectorPdf is not None:
box = VectorPdf.load_xobj(srcinfo).BBox
else:
pdf = LazyImports.pdfinfo
if pdf is None:
log.warning(
'PDF images are not supported without pyPdf or pdfrw [%s]',
nodeid(node))
return 0, 0, 'direct'
reader = pdf.PdfFileReader(open(imgname, 'rb'))
box = [float(x) for x in reader.getPage(0)['/MediaBox']]
x1, y1, x2, y2 = box
# These are in pt, so convert to px
iw = float((x2 - x1) * xdpi / 72)
ih = float((y2 - y1) * ydpi / 72)
size_known = True # Assume size from original PDF is OK
else:
keeptrying = True
if LazyImports.PILImage:
try:
img = LazyImports.PILImage.open(imgname)
img.load()
iw, ih = img.size
xdpi, ydpi = img.info.get('dpi', (xdpi, ydpi))
keeptrying = False
except IOError: # PIL throws this when it's a broken/unknown image
pass
if keeptrying and LazyImports.PMImage:
img = LazyImports.PMImage(imgname)
iw = img.size().width()
ih = img.size().height()
density = img.density()
# The density is in pixelspercentimeter (!?)
xdpi = density.width() * 2.54
ydpi = density.height() * 2.54
keeptrying = False
if keeptrying:
log.error(
"The image (%s, %s) is broken or in an unknown format",
imgname, nodeid(node))
raise ValueError
# Try to get the print resolution from the image itself via PIL.
# If it fails, assume a DPI of 300, which is pretty much made up,
# and then a 100% size would be iw*inch/300, so we pass
# that as the second parameter to adjustUnits
#
# Some say the default DPI should be 72. That would mean
# the largest printable image in A4 paper would be something
# like 480x640. That would be awful.
#
w = node.get('width')
h = node.get('height')
if h is None and w is None: # Nothing specified
# Guess from iw, ih
log.warning(
"Using image %s without specifying size."
"Calculating based on image size at %ddpi [%s]", imgname, xdpi,
nodeid(node))
w = iw * inch / xdpi
h = ih * inch / ydpi
elif w is not None:
# Node specifies only w
# In this particular case, we want the default unit
# to be pixels so we work like rst2html
if w[-1] == '%':
kind = 'percentage_of_container'
w = int(w[:-1])
else:
# This uses default DPI setting because we
# are not using the image's "natural size"
# this is what LaTeX does, according to the
# docutils mailing list discussion
w = client.styles.adjustUnits(w,
client.styles.tw,
default_unit='px')
if h is None:
# h is set from w with right aspect ratio
h = w * ih / iw
else:
h = client.styles.adjustUnits(h,
ih * inch / ydpi,
default_unit='px')
elif h is not None and w is None:
if h[-1] != '%':
h = client.styles.adjustUnits(h,
ih * inch / ydpi,
default_unit='px')
# w is set from h with right aspect ratio
w = h * iw / ih
else:
log.error('Setting height as a percentage does **not** work. '\
'ignoring height parameter [%s]', nodeid(node))
# Set both from image data
w = iw * inch / xdpi
h = ih * inch / ydpi
# Apply scale factor
w = w * scale
h = h * scale
# And now we have this probably completely bogus size!
log.info("Image %s size calculated: %fcm by %fcm [%s]", imgname,
w / cm, h / cm, nodeid(node))
return w, h, kind
def size_for_node(self, node, client):
'''Given a docutils image node, returns the size the image should have
in the PDF document, and what 'kind' of size that is.
That involves lots of guesswork'''
uri = str(node.get("uri"))
if uri.split("://")[0].lower() not in ('http','ftp','https'):
uri = os.path.join(client.basedir,uri)
else:
uri, _ = urllib.urlretrieve(uri)
client.to_unlink.append(uri)
srcinfo = client, uri
# Extract all the information from the URI
imgname, extension, options = self.split_uri(uri)
if not os.path.isfile(imgname):
imgname = missing
scale = float(node.get('scale', 100))/100
size_known = False
# Figuring out the size to display of an image is ... annoying.
# If the user provides a size with a unit, it's simple, adjustUnits
# will return it in points and we're done.
# However, often the unit wil be "%" (specially if it's meant for
# HTML originally. In which case, we will use a percentage of
# the containing frame.
# Find the image size in pixels:
kind = 'direct'
xdpi, ydpi = client.styles.def_dpi, client.styles.def_dpi
extension = imgname.split('.')[-1].lower()
if extension in ['svg','svgz'] and SVGImage.available():
iw, ih = SVGImage(imgname, srcinfo=srcinfo).wrap(0, 0)
# These are in pt, so convert to px
iw = iw * xdpi / 72
ih = ih * ydpi / 72
elif extension in [
"ai", "ccx", "cdr", "cgm", "cmx",
"sk1", "sk", "xml", "wmf", "fig"] and VectorImage.available():
iw, ih = VectorImage(imgname, srcinfo=srcinfo).wrap(0, 0)
# These are in pt, so convert to px
iw = iw * xdpi / 72
ih = ih * ydpi / 72
elif extension == 'pdf':
if VectorPdf is not None:
box = VectorPdf.load_xobj(srcinfo).BBox
else:
pdf = LazyImports.pdfinfo
if pdf is None:
log.warning('PDF images are not supported without pyPdf or pdfrw [%s]', nodeid(node))
return 0, 0, 'direct'
reader = pdf.PdfFileReader(open(imgname, 'rb'))
box = [float(x) for x in reader.getPage(0)['/MediaBox']]
x1, y1, x2, y2 = box
# These are in pt, so convert to px
iw = float((x2-x1) * xdpi / 72)
ih = float((y2-y1) * ydpi / 72)
size_known = True # Assume size from original PDF is OK
else:
keeptrying = True
if sys.platform[0:4] == 'java':
try:
from javax.imageio import ImageIO
from java.io import File
iis = ImageIO.createImageInputStream(File(imgname))
readers = ImageIO.getImageReaders(iis)
if readers.hasNext():
reader = readers.next()
reader.setInput(iis, True)
metadata = reader.getImageMetadata(0)
iw = reader.getWidth(0)
ih = reader.getHeight(0)
xdpi = None
ydpi = None
if metadata.getNativeMetadataFormatName() == 'javax_imageio_png_1.0':
png_metadata = metadata.getAsTree(metadata.getNativeMetadataFormatName())
if png_metadata is not None:
phys_metadata = png_metadata.getElementsByTagName('pHYs')
if phys_metadata.getLength() == 1:
xdpi = phys_metadata.item(0).getAttribute('pixelsPerUnitXAxis')
if xdpi is not None:
xdpi = float(xdpi)
ydpi = phys_metadata.item(0).getAttribute('pixelsPerUnitYAxis')
if ydpi is not None:
ydpi = float(ydpi)
elif metadata.getNativeMetadataFormatName() == 'javax_imageio_jpeg_image_1.0':
jpg_metadata = metadata.getAsTree(metadata.getNativeMetadataFormatName())
if jpg_metadata is not None:
jfif_metadata = jpg_metadata.getElementsByTagName('app0JFIF')
if jfif_metadata.getLength() == 1:
resUnits = jfif_metadata.item(0).getAttribute('resUnits')
# 0 = plain aspect ratio, 1 = dots per inch, 2 = dots per cm
if resUnits is not None and (resUnits == '1' or resUnits == '2'):
xdpi = jfif_metadata.item(0).getAttribute('pixelsPerUnitXAxis')
if xdpi is not None:
xdpi = float(xdpi)
if resUnits == '2':
xdpi = xdpi * 2.54
ydpi = jfif_metadata.item(0).getAttribute('pixelsPerUnitYAxis')
if ydpi is not None:
ydpi = float(ydpi)
if resUnits == '2':
ydpi = ydpi * 2.54
if xdpi is None:
xdpi = 300
if ydpi is None:
ydpi = 300
keeptrying = False
except Exception, err:
log.error('Error %s', str(err))
pass
if LazyImports.PILImage:
try:
img = LazyImports.PILImage.open(imgname)
img.load()
iw, ih = img.size
xdpi, ydpi = img.info.get('dpi', (xdpi, ydpi))
keeptrying = False
except IOError: # PIL throws this when it's a broken/unknown image
pass
if keeptrying and LazyImports.PMImage:
img = LazyImports.PMImage(imgname)
iw = img.size().width()
ih = img.size().height()
density=img.density()
# The density is in pixelspercentimeter (!?)
xdpi=density.width()*2.54
ydpi=density.height()*2.54
keeptrying = False
if keeptrying:
if extension not in ['jpg', 'jpeg']:
log.error("The image (%s, %s) is broken or in an unknown format"
, imgname, nodeid(node))
raise ValueError
else:
# Can be handled by reportlab
log.warning("Can't figure out size of the image (%s, %s). Install PIL for better results."
, imgname, nodeid(node))
iw = 1000
ih = 1000
# If it fails, assume a DPI of 300, which is pretty much made up,
# and then a 100% size would be iw*inch/300, so we pass
# that as the second parameter to adjustUnits
#
# Some say the default DPI should be 72. That would mean
# the largest printable image in A4 paper would be something
# like 480x640. That would be awful.
#
w = node.get('width')
h = node.get('height')
if h is None and w is None: # Nothing specified
# Guess from iw, ih
log.debug("Using image %s without specifying size."
"Calculating based on image size at %ddpi [%s]",
imgname, xdpi, nodeid(node))
w = iw*inch/xdpi
h = ih*inch/ydpi
elif w is not None:
# Node specifies only w
# In this particular case, we want the default unit
# to be pixels so we work like rst2html
if w[-1] == '%':
kind = 'percentage_of_container'
w=int(w[:-1])
else:
# This uses default DPI setting because we
# are not using the image's "natural size"
# this is what LaTeX does, according to the
# docutils mailing list discussion
w = client.styles.adjustUnits(w, client.styles.tw,
def size_for_node(self, node, client):
'''Given a docutils image node, returns the size the image should have
in the PDF document, and what 'kind' of size that is.
That involves lots of guesswork'''
uri = str(node.get("uri"))
if uri.split("://")[0].lower() not in ('http','ftp','https'):
uri = os.path.join(client.basedir,uri)
else:
uri, _ = urllib.urlretrieve(uri)
client.to_unlink.append(uri)
srcinfo = client, uri
# Extract all the information from the URI
imgname, extension, options = self.split_uri(uri)
if not os.path.isfile(imgname):
imgname = missing
scale = float(node.get('scale', 100))/100
size_known = False
# Figuring out the size to display of an image is ... annoying.
# If the user provides a size with a unit, it's simple, adjustUnits
# will return it in points and we're done.
# However, often the unit wil be "%" (specially if it's meant for
# HTML originally. In which case, we will use a percentage of
# the containing frame.
# Find the image size in pixels:
kind = 'direct'
xdpi, ydpi = client.styles.def_dpi, client.styles.def_dpi
extension = imgname.split('.')[-1].lower()
if extension in ['svg','svgz'] and SVGImage.available():
iw, ih = SVGImage(imgname, srcinfo=srcinfo).wrap(0, 0)
# These are in pt, so convert to px
iw = iw * xdpi / 72
ih = ih * ydpi / 72
elif extension in [
"ai", "ccx", "cdr", "cgm", "cmx",
"sk1", "sk", "xml", "wmf", "fig"] and VectorImage.available():
iw, ih = VectorImage(imgname, srcinfo=srcinfo).wrap(0, 0)
# These are in pt, so convert to px
iw = iw * xdpi / 72
ih = ih * ydpi / 72
elif extension == 'pdf':
if VectorPdf is not None:
box = VectorPdf.load_xobj(srcinfo).BBox
else:
pdf = LazyImports.pdfinfo
if pdf is None:
log.warning('PDF images are not supported without pyPdf or pdfrw [%s]', nodeid(node))
return 0, 0, 'direct'
reader = pdf.PdfFileReader(open(imgname, 'rb'))
box = [float(x) for x in reader.getPage(0)['/MediaBox']]
x1, y1, x2, y2 = box
# These are in pt, so convert to px
iw = float((x2-x1) * xdpi / 72)
ih = float((y2-y1) * ydpi / 72)
size_known = True # Assume size from original PDF is OK
else:
keeptrying = True
if LazyImports.PILImage:
try:
img = LazyImports.PILImage.open(imgname)
img.load()
iw, ih = img.size
xdpi, ydpi = img.info.get('dpi', (xdpi, ydpi))
keeptrying = False
except IOError: # PIL throws this when it's a broken/unknown image
pass
if keeptrying and LazyImports.PMImage:
img = LazyImports.PMImage(imgname)
iw = img.size().width()
ih = img.size().height()
density=img.density()
# The density is in pixelspercentimeter (!?)
xdpi=density.width()*2.54
ydpi=density.height()*2.54
keeptrying = False
if keeptrying:
if extension not in ['jpg', 'jpeg']:
log.error("The image (%s, %s) is broken or in an unknown format"
, imgname, nodeid(node))
raise ValueError
else:
# Can be handled by reportlab
log.warning("Can't figure out size of the image (%s, %s). Install PIL for better results."
, imgname, nodeid(node))
iw = 1000
ih = 1000
# Try to get the print resolution from the image itself via PIL.
# If it fails, assume a DPI of 300, which is pretty much made up,
# and then a 100% size would be iw*inch/300, so we pass
# that as the second parameter to adjustUnits
#
# Some say the default DPI should be 72. That would mean
# the largest printable image in A4 paper would be something
# like 480x640. That would be awful.
#
w = node.get('width')
h = node.get('height')
if h is None and w is None: # Nothing specified
# Guess from iw, ih
log.debug("Using image %s without specifying size."
"Calculating based on image size at %ddpi [%s]",
imgname, xdpi, nodeid(node))
w = iw*inch/xdpi
h = ih*inch/ydpi
elif w is not None:
# Node specifies only w
# In this particular case, we want the default unit
# to be pixels so we work like rst2html
if w[-1] == '%':
kind = 'percentage_of_container'
w=int(w[:-1])
else:
# This uses default DPI setting because we
# are not using the image's "natural size"
# this is what LaTeX does, according to the
# docutils mailing list discussion
w = client.styles.adjustUnits(w, client.styles.tw,
default_unit='px')
if h is None:
# h is set from w with right aspect ratio
h = w*ih/iw
else:
h = client.styles.adjustUnits(h, ih*inch/ydpi, default_unit='px')
elif h is not None and w is None:
if h[-1] != '%':
h = client.styles.adjustUnits(h, ih*inch/ydpi, default_unit='px')
# w is set from h with right aspect ratio
w = h*iw/ih
else:
log.error('Setting height as a percentage does **not** work. '\
'ignoring height parameter [%s]', nodeid(node))
# Set both from image data
w = iw*inch/xdpi
h = ih*inch/ydpi
# Apply scale factor
w = w*scale
h = h*scale
# And now we have this probably completely bogus size!
log.info("Image %s size calculated: %fcm by %fcm [%s]",
imgname, w/cm, h/cm, nodeid(node))
return w, h, kind