def ocr_page_image(doc_path, page_num, lang):
"""
image = jpg, jpeg, png
"""
logger.debug("OCR image (jpeg, jpg, png) document")
page_url = PagePath(
document_path=doc_path,
page_num=page_num,
step=Step(1),
# jpeg, jpg, png are 1 page documents
page_count=1)
# resize and eventually convert (png -> jpg)
resize_img(page_url, media_root=settings.MEDIA_ROOT)
extract_txt(page_url, lang=lang, media_root=settings.MEDIA_ROOT)
# First quickly generate preview images
for step in Steps():
page_url.step = step
resize_img(page_url, media_root=settings.MEDIA_ROOT)
# reset page's step
page_url.step = Step(1)
# Now OCR each image
for step in Steps():
if not step.is_thumbnail:
extract_hocr(page_url, lang=lang, media_root=settings.MEDIA_ROOT)
return page_url
def ocr_page_pdf(doc_path, page_num, lang):
"""
doc_path is an mglib.path.DocumentPath instance
"""
logger.debug("OCR PDF document")
page_count = get_pagecount(default_storage.abspath(doc_path.url()))
if page_num <= page_count:
# first quickly generate preview images
page_url = PagePath(document_path=doc_path,
page_num=page_num,
step=Step(1),
page_count=page_count)
for step in Steps():
page_url.step = step
extract_img(page_url, media_root=settings.MEDIA_ROOT)
if page_num <= page_count:
page_url = PagePath(document_path=doc_path,
page_num=page_num,
step=Step(1),
page_count=page_count)
extract_txt(page_url, lang=lang, media_root=settings.MEDIA_ROOT)
for step in Steps():
page_url.step = step
if not step.is_thumbnail:
extract_hocr(page_url,
lang=lang,
media_root=settings.MEDIA_ROOT)
return page_url
def ocr_page_pdf(doc_path, page_num, lang, **kwargs):
"""
doc_path is an mglib.path.DocumentPath instance
On success returns ``mglib.path.PagePath`` instance.
"""
logger.debug("OCR PDF document")
file_name = kwargs.pop('file_name', None)
if not file_name:
file_name = doc_path.file_name
page_count = get_pagecount(default_storage.abspath(doc_path.url()))
if page_num <= page_count:
# first quickly generate preview images
page_path = PagePath(document_path=doc_path,
page_num=page_num,
step=Step(1),
page_count=page_count)
for step in Steps():
page_path.step = step
extract_img(page_path, media_root=settings.MEDIA_ROOT)
notify_pre_page_ocr(page_path,
page_num=page_num,
lang=lang,
file_name=doc_path.file_name,
**kwargs)
if page_num <= page_count:
page_path = PagePath(document_path=doc_path,
page_num=page_num,
step=Step(1),
page_count=page_count)
extract_txt(page_path, lang=lang, media_root=settings.MEDIA_ROOT)
notify_txt_ready(page_path,
page_num=page_num,
lang=lang,
file_name=file_name,
**kwargs)
for step in Steps():
page_path.step = step
if not step.is_thumbnail:
extract_hocr(page_path,
lang=lang,
media_root=settings.MEDIA_ROOT)
notify_hocr_ready(
page_path,
page_num=page_num,
lang=lang,
# step as integer number
step=step.current,
file_name=file_name,
**kwargs)
return page_path
def ocr_page_image(doc_path, page_num, lang, **kwargs):
"""
image = jpg, jpeg, png
On success returns ``mglib.path.PagePath`` instance.
"""
logger.debug("OCR image (jpeg, jpg, png) document")
page_path = PagePath(
document_path=doc_path,
page_num=page_num,
step=Step(1),
# jpeg, jpg, png are 1 page documents
page_count=1)
notify_pre_page_ocr(page_path,
page_num=page_num,
lang=lang,
file_name=doc_path.file_name,
**kwargs)
# resize and eventually convert (png -> jpg)
resize_img(page_path, media_root=settings.MEDIA_ROOT)
extract_txt(page_path, lang=lang, media_root=settings.MEDIA_ROOT)
notify_txt_ready(page_path,
page_num=page_num,
lang=lang,
file_name=doc_path.file_name,
**kwargs)
# First quickly generate preview images
for step in Steps():
page_path.step = step
resize_img(page_path, media_root=settings.MEDIA_ROOT)
# reset page's step
page_path.step = Step(1)
# Now OCR each image
for step in Steps():
if not step.is_thumbnail:
extract_hocr(page_path, lang=lang, media_root=settings.MEDIA_ROOT)
notify_hocr_ready(
page_path,
page_num=page_num,
lang=lang,
# step as integer number
step=step.current,
file_name=doc_path.file_name,
**kwargs)
return page_path
def paste_pages(self,
dest_doc_path,
data_list,
dest_doc_is_new=False,
after_page_number=False,
before_page_number=False):
"""
Pastes pages in the document pointed by dest_doc_path
from src_doc_path. Both dest and src are instances of
mglib.path.DocumentPath
"""
next_ver_dp = DocumentPath.copy_from(dest_doc_path,
version=dest_doc_path.version + 1)
self.make_sure_path_exists(self.abspath(next_ver_dp))
pdftk.paste_pages(src=self.abspath(dest_doc_path),
dst=self.abspath(next_ver_dp),
data_list=data_list,
dst_doc_is_new=dest_doc_is_new,
after_page_number=after_page_number,
before_page_number=before_page_number)
if not dest_doc_is_new:
# migrate document's own pages from previous
# version (this differs from pasting into newly
# created docs)
pcount = self.get_pagecount(dest_doc_path)
data_list.insert(0, {
'doc_path': dest_doc_path,
'page_nums': list(range(1, pcount + 1))
})
dest_page_num = 1
dest_page_count = sum([len(item['page_nums']) for item in data_list])
for item in data_list:
src_path = item['doc_path']
for page_num in item['page_nums']:
for step in Steps():
src_page_path = PagePath(
document_path=src_path,
page_num=int(page_num),
step=step,
page_count=self.get_pagecount(src_path))
dst_page_path = PagePath(document_path=next_ver_dp,
page_num=dest_page_num,
step=step,
page_count=dest_page_count)
logger.debug(f"src={src_page_path} dst={dst_page_path}")
self.copy_page(src_page_path=src_page_path,
dst_page_path=dst_page_path)
dest_page_num += 1
return dest_doc_path.version + 1
def reorder_pages(self, doc_path, new_order):
"""
Reorders pages in the document pointed by doc_path.
doc_path is an instance of mglib.path.DocumentPath
In case of success returns document's new version.
new_order is a list of following format:
[
{'page_num': 2, page_order: 1},
{'page_num': 1, page_order: 2},
{'page_num': 3, page_order: 3},
{'page_num': 4, page_order: 4},
]
Example above means that in current document of 4 pages,
first page was swapped with second one.
page_num = older page order
page_order = current page order
So in human language, each hash is read:
<page_num> now should be <page_order>
"""
src_doc_path = doc_path
dst_doc_path = DocumentPath.copy_from(src_doc_path,
version=doc_path.version + 1)
self.make_sure_path_exists(self.abspath(dst_doc_path))
stapler.reorder_pages(src=self.abspath(src_doc_path),
dst=self.abspath(dst_doc_path),
new_order=new_order)
page_count = self.get_pagecount(doc_path)
if len(new_order) > page_count:
logger.error(
f"deleted_pages({new_order}) > page_count({page_count})")
return
for item in new_order:
for step in Steps():
src_page_path = PagePath(document_path=src_doc_path,
page_num=int(item['page_num']),
step=step,
page_count=len(new_order))
dst_page_path = PagePath(document_path=dst_doc_path,
page_num=int(item['page_order']),
step=step,
page_count=len(new_order))
self.copy_page(src_page_path=src_page_path,
dst_page_path=dst_page_path)
return doc_path.version + 1
def delete_pages(self, doc_path, page_numbers, skip_migration=False):
"""
Delets pages in the document pointed by doc_path.
doc_path is an instance of mglib.path.DocumentPath
In case of success returns document's new version.
"""
if not isinstance(page_numbers, list):
logger.error("Expecting list argument")
return False
src_doc_path = doc_path
dst_doc_path = DocumentPath.copy_from(src_doc_path,
version=doc_path.version + 1)
self.make_sure_path_exists(self.abspath(dst_doc_path))
stapler.delete_pages(self.abspath(src_doc_path),
self.abspath(dst_doc_path), page_numbers)
if skip_migration:
return doc_path.version + 1
page_count = self.get_pagecount(doc_path)
if len(page_numbers) > page_count:
logger.error(
f"deleted_pages({page_numbers}) > page_count({page_count})")
return
assigns = get_assigns_after_delete(total_pages=page_count,
deleted_pages=page_numbers)
for a in assigns:
for step in Steps():
src_page_path = PagePath(document_path=src_doc_path,
page_num=a[1],
step=step,
page_count=page_count)
dst_page_path = PagePath(document_path=dst_doc_path,
page_num=a[0],
step=step,
page_count=page_count -
len(page_numbers))
self.copy_page(src_page_path=src_page_path,
dst_page_path=dst_page_path)
return doc_path.version + 1
def upload_to(page_path, **kwargs):
"""
Will upload results to remote storage if such storage
is supported.
Default storage class ``mglib.storage.FileSystemStorage``
``download`` and ``upload`` methods are empty.
Method is useful for cloud production (where storage class
will be replaced to support remote storage e.g. S3, SFTP)
"""
default_storage.upload(page_path.txt_url(), **kwargs)
for step in Steps():
if not step.is_thumbnail:
page_path.step = step
default_storage.upload(page_path.hocr_url(), **kwargs)
default_storage.upload(page_path.img_url(), **kwargs)
def ocr_page_pdf(doc_path, page_num, lang):
"""
doc_path is an mglib.path.DocumentPath instance
"""
page_count = get_pagecount(default_storage.abspath(doc_path.url()))
if page_num <= page_count:
page_url = PagePath(document_path=doc_path,
page_num=page_num,
step=Step(1),
page_count=page_count)
extract_img(page_url, media_root=settings.MEDIA_ROOT)
extract_txt(page_url, lang=lang, media_root=settings.MEDIA_ROOT)
for step in Steps():
page_url.step = step
extract_img(page_url, media_root=settings.MEDIA_ROOT)
# tesseract unterhalt-1.jpg page-1 -l deu hocr
if not step.is_thumbnail:
extract_hocr(page_url,
lang=lang,
media_root=settings.MEDIA_ROOT)
return page_url
