def test_document_save():
with tempfile.TemporaryDirectory() as d:
doc = Document(res('sample.pdf'))
fname = f'{d}/temp.pdf'
doc.save(fname)
doc2 = Document(fname)
assert len(doc) == len(doc2)
assert doc2.meta['Title'] == 'Red Stork'
def test_geometry_sane():
for fname in [
'arxiv1901.01387.pdf',
'arxiv1901.02668.pdf',
'arxiv1901.09059.pdf',
'arxiv1901.02066.pdf',
'arxiv1901.08145v1.pdf',
'arxiv1901.10092.pdf',
# 'arxiv1901.02527.pdf',
# 'arxiv1901.08637.pdf',
'arxiv1901.11067.pdf'
]:
doc = Document(res(fname))
for pageidx, page in enumerate(doc):
objcount = 0
for obj in page.flat_iter():
if obj.type == PageObject.OBJ_TYPE_TEXT:
txt_boxes = list(obj.text_geometry_iter())
for _, bx in txt_boxes:
assert inside(page.crop_box,
bx), (fname, pageidx, page.crop_box,
objcount, bx)
# assert all(inside(obj.rect, bx) for _,bx in txt_boxes), (fname, pageidx, obj.rect, objcount)
objcount += 1
def test_text_geometry_iter():
doc = Document(res('sample.pdf'))
page = doc[2]
assert len(page) == 19
text_objs = [
x for x in page.flat_iter() if x.type == PageObject.OBJ_TYPE_TEXT
]
text = text_objs[0]
assert len(text) == 8
items = list(text.text_geometry_iter())
assert items == [
('M', (489.1150207519531, 619.3770141601562, 500.0322828043718,
629.8354499922134)),
('a', (501.0654058456421, 619.0477800783701, 508.73936194344424,
627.1105125918984)),
('n', (509.0418930053711, 619.3770141601562, 516.0293610177469,
627.1105125918984)),
('u', (516.8749179840088, 619.0477800783701, 523.7888337015174,
626.8933581975289)),
('a', (524.8514060974121, 619.0477800783701, 532.5253621952143,
627.1105125918984)),
('l', (532.8278923034668, 619.3770141601562, 535.0099437178578,
629.8354499922134)),
(':', (536.0127487182617, 619.3770141601562, 539.0844326515216,
626.8933581975289)),
]
def test_unicode_map_multichar3():
doc = Document(res('arxiv1901.09059.pdf'))
text = []
for obj in doc[0]:
if obj.type == PageObject.OBJ_TYPE_TEXT:
text.append(obj.text) # should not crash
assert ''.join(text)[-20:] == 'd applicability. '
def test_unicode_map_multichar2():
doc = Document(res('arxiv1901.08145v1.pdf'))
text = []
for obj in doc[0]:
if obj.type == PageObject.OBJ_TYPE_TEXT:
text.append(obj.text) # should not crash
assert ''.join(text)[-20:] == 'ing-centerpositionX,'
def test_page_smoke():
doc = Document(res('sample.pdf'))
page = doc[0]
assert page.media_box == (0., 0., 612., 792.)
assert page.crop_box == (0., 0., 612., 792.)
def test_document():
doc = Document(res('sample.pdf'))
assert len(doc) == 15
assert doc.meta['Title'] == 'Red Stork'
str(doc) # should not crash
def test_rotated_text():
doc = Document(res('arxiv1901.02066.pdf'))
page = doc[0]
objs = list(page.flat_iter())
obj = objs[296]
rect = obj.rect
text_geom = list(obj.text_geometry_iter())
assert all(inside(rect, bx) for _, bx in text_geom)
def test_page_object_smoke():
doc = Document(res('sample.pdf'))
page = doc[2]
assert len(page) == 19
text_objs = [x for x in page if x.type == PageObject.OBJ_TYPE_TEXT]
path_objs = [x for x in page if x.type == PageObject.OBJ_TYPE_PATH]
assert len(text_objs) == 16
assert len(path_objs) == 3
def test_flatiter():
'''used to crash'''
doc = Document(res('arxiv1901.10092.pdf'))
text = []
for obj in doc[0].flat_iter():
if obj.type == PageObject.OBJ_TYPE_TEXT:
text.append(obj.text) # should not crash
assert ''.join(text)[-20:] == 're.com/naturephysics'
def test_document_edit_meta():
with tempfile.TemporaryDirectory() as d:
doc = Document(res('sample.pdf'))
doc.meta['Title'] = 'Best PDF Parsing tool in this world!'
fname = f'{d}/temp.pdf'
doc.save(fname)
doc2 = Document(fname)
assert len(doc) == len(doc2)
assert doc2.meta['Title'] == 'Best PDF Parsing tool in this world!'
def test_page_object_chars():
doc = Document(res('sample.pdf'))
page = doc[2]
assert len(page) == 19
text_objs = [x for x in page if x.type == PageObject.OBJ_TYPE_TEXT]
text = text_objs[0]
assert len(text) == 8
items = list(text)
assert items == [
(77, 0.0, 0.0),
(97, 11.950385093688965, 0.0),
(110, 19.92687225341797, 0.0),
(-1, 10.0, 0.0),
(117, 27.759897232055664, 0.0),
(97, 35.736385345458984, 0.0),
(108, 43.71287155151367, 0.0),
(58, 46.897727966308594, 0.0),
]
assert text.text == 'Manual:'
def test_page_label():
doc = Document(res('sample.pdf'))
assert doc[2].label == 'i'
assert doc[-3].label == '9'
from redstork import Document, PageObject, Glyph
from redstork.test import res
doc = Document(res('sample.pdf'))
print('Number of pages:', len(doc))
print('MediaBox of the first page is:', doc[0].media_box)
print('Rotation of the first page is:', doc[0].rotation)
print('Document title:', doc.meta['Title'])
print('First page has', len(doc[0]), 'objects')
doc[0].render('page-0.ppm', scale=2) # render page #1 as image
page = doc[0]
for o in page:
if o.type == PageObject.OBJ_TYPE_TEXT:
for code, _, _ in o:
print(o.font[code], end='')
print()
for fid, font in doc.fonts.items():
print(font.simple_name, fid)
# lets generate an SVG file of the first letter on page 1
text_object = [o for o in page
if o.type == PageObject.OBJ_TYPE_TEXT][0] # first text object
charcode, _, _ = text_object[0] # first character of the first text object