def extract_images(doc):
"""If the given document (proposal.models.Document) has been copied to
the local filesystem, extract its images to a subdirectory of the
document's directory (docs/<doc id>/images). Extracts the text
contents to docs/<doc id>/text.txt.
:param doc: proposal.models.Document object with a corresponding PDF
file that has been copied to the local filesystem
:returns: A list of proposal.model.Image objects
"""
# TODO: Break this into smaller subtasks
docfile = doc.document
if not docfile:
logger.error("Document has not been copied to the local filesystem.")
return []
path = docfile.path
if not os.path.exists(path):
logger.error("Document %s is not where it says it is: %s", doc.pk,
path)
return []
images_dir = os.path.join(os.path.dirname(path), "images")
os.makedirs(images_dir, exist_ok=True)
images_pattern = os.path.join(images_dir, "image")
logger.info("Extracting images to '%s'", images_dir)
status = subprocess.call(
["pdfimages", "-png", "-tiff", "-j", "-jp2", path, images_pattern])
images = []
if status:
logger.warn("pdfimages failed with exit code %i", status)
else:
# Do stuff with the images in the directory
for image_name in os.listdir(images_dir):
image_path = os.path.join(images_dir, image_name)
if not is_interesting(image_path):
# Delete 'uninteresting' images
os.unlink(image_path)
continue
image = Image(proposal=doc.proposal, document=doc)
image.image = image_path
images.append(image)
try:
image.save()
except IntegrityError:
# This can occur if the image has already been fetched
# and associated with the Proposal.
pass
return images
def extract_images(doc_id):
"""If the given document (proposal.models.Document) has been copied to
the local filesystem, extract its images to a subdirectory of the
document's directory (docs/<doc id>/images).
:param doc: proposal.models.Document object with a corresponding PDF
file that has been copied to the local filesystem
:returns: A list of proposal.model.Image objects
"""
doc = Document.objects.get(pk=doc_id)
docfile = doc.document
if not docfile:
logger.error("Document has not been copied to the local filesystem.")
return []
path = docfile.path
if not os.path.exists(path):
logger.error("Document %s is not where it says it is: %s",
doc.pk, path)
return []
images_dir = os.path.join(os.path.dirname(path), "images")
os.makedirs(images_dir, exist_ok=True)
logger.info("Extracting images to '%s'", images_dir)
image_paths = pdf.extract_images(path, dirname=images_dir)
images = []
# Do stuff with the images in the directory
for image_name in image_paths:
image_path = os.path.join(images_dir, image_name)
if not is_interesting(image_path):
# Delete 'uninteresting' images
os.unlink(image_path)
continue
image = Image(proposal=doc.proposal,
document=doc)
image.image = image_path
images.append(image)
try:
image.save()
except IntegrityError:
# This can occur if the image has already been fetched
# and associated with the Proposal.
pass
logger.info("Extracted %i image(s) from %s.", len(images), path)
return [image.pk for image in images]
def extract_images(doc_id):
"""If the given document (proposal.models.Document) has been copied to
the local filesystem, extract its images to a subdirectory of the
document's directory (docs/<doc id>/images).
:param doc: proposal.models.Document object with a corresponding PDF
file that has been copied to the local filesystem
:returns: A list of proposal.model.Image objects
"""
doc = Document.objects.get(pk=doc_id)
docfile = doc.document
if not docfile:
logger.error("Document has not been copied to the local filesystem.")
return []
path = docfile.path
if not os.path.exists(path):
logger.error("Document %s is not where it says it is: %s", doc.pk,
path)
return []
images_dir = os.path.join(os.path.dirname(path), "images")
os.makedirs(images_dir, exist_ok=True)
logger.info("Extracting images to '%s'", images_dir)
image_paths = pdf.extract_images(path, dirname=images_dir)
images = []
# Do stuff with the images in the directory
for image_name in image_paths:
image_path = os.path.join(images_dir, image_name)
if not is_interesting(image_path):
# Delete 'uninteresting' images
os.unlink(image_path)
continue
image = Image(proposal=doc.proposal, document=doc)
image.image = image_path
images.append(image)
try:
image.save()
except IntegrityError:
# This can occur if the image has already been fetched
# and associated with the Proposal.
pass
logger.info("Extracted %i image(s) from %s.", len(images), path)
return [image.pk for image in images]
def extract_content(doc, encoding="ISO-8859-9"):
"""If the given document (proposal.models.Document) has been copied to
the local filesystem, extract its images to a subdirectory of the
document's directory (docs/<doc id>/images). Extracts the text
content to docs/<doc id>/content.txt.
"""
docfile = doc.document
logger = extract_content.get_logger()
if not docfile:
logger.error("Document has not been copied to the local filesystem.")
return
try:
path = docfile.path
except:
path = docfile.name
if not os.path.exists(path):
logger.error("Document %s is not where it says it is: %s",
doc.pk, path)
return
images_dir = os.path.join(os.path.dirname(path), "images")
os.makedirs(images_dir, exist_ok=True)
images_pattern = os.path.join(images_dir, "image")
logger.info("Extracting images to '%s'", images_dir)
status = subprocess.call(["pdfimages", "-png", "-tiff", "-j", "-jp2",
path, images_pattern])
if status:
logger.warn("pdfimages failed with exit code %i", status)
else:
# Do stuff with the images in the directory
for image_name in os.listdir(images_dir):
image_path = os.path.join(images_dir, image_name)
if not images.is_interesting(image_path):
# Delete 'uninteresting' images
os.unlink(image_path)
continue
image = Image(proposal=doc.proposal,
document=doc)
image.image = image_path
#image.set_image_path(image_path)
try:
image.save()
except IntegrityError:
# This can occur if the image has already been fetched
# and associated with the Proposal.
pass
# Could consider storing the full extracted text of the document in
# the database and indexing it, rather than extracting it to a file.
text_path = os.path.join(os.path.dirname(path), "text.txt")
# TODO: It may be practical to sniff pdfinfo, determine the PDF
# producer used, and make a best guess at encoding based on that
# information. We should be able to get away with using ISO-8859-9
# for now.
status = subprocess.call(["pdftotext", "-enc", encoding, path, text_path])
if status:
logger.error("Failed to extract text from {doc}".\
format(doc=path))
else:
# Do stuff with the contents of the file.
# Possibly perform some rudimentary scraping?
doc.fulltext = text_path
doc.encoding = encoding
doc.save()
def extract_images(doc):
"""If the given document (proposal.models.Document) has been copied to
the local filesystem, extract its images to a subdirectory of the
document's directory (docs/<doc id>/images). Extracts the text
contents to docs/<doc id>/text.txt.
:param doc: proposal.models.Document object with a corresponding PDF
file that has been copied to the local filesystem
:returns: A list of proposal.model.Image objects
"""
# TODO: Break this into smaller subtasks
docfile = doc.document
logger = extract_images.get_logger()
if not docfile:
logger.error("Document has not been copied to the local filesystem.")
return []
path = docfile.path
if not os.path.exists(path):
logger.error("Document %s is not where it says it is: %s",
doc.pk, path)
return []
images_dir = os.path.join(os.path.dirname(path), "images")
os.makedirs(images_dir, exist_ok=True)
images_pattern = os.path.join(images_dir, "image")
logger.info("Extracting images to '%s'", images_dir)
status = subprocess.call(["pdfimages", "-png", "-tiff", "-j", "-jp2",
path, images_pattern])
images = []
if status:
logger.warn("pdfimages failed with exit code %i", status)
else:
# Do stuff with the images in the directory
for image_name in os.listdir(images_dir):
image_path = os.path.join(images_dir, image_name)
if not images.is_interesting(image_path):
# Delete 'uninteresting' images
os.unlink(image_path)
continue
image = Image(proposal=doc.proposal,
document=doc)
image.image = image_path
images.append(image)
try:
image.save()
except IntegrityError:
# This can occur if the image has already been fetched
# and associated with the Proposal.
pass
return images
