def test_add_underline_annotation(self):
# create output directory if it does not exist yet
if not self.output_file.parent.exists():
self.output_file.parent.mkdir()
# attempt to read PDF
doc = None
with open(self.input_file, "rb") as in_file_handle:
print("\treading (1) ..")
doc = PDF.loads(in_file_handle)
# add annotation
doc.get_page(0).append_underline_annotation(
stroke_color=HexColor("CBEF43"),
rectangle=Rectangle(Decimal(72.86), Decimal(486.82), Decimal(129),
Decimal(13)),
)
# attempt to store PDF
with open(self.output_file, "wb") as out_file_handle:
print("\twriting ..")
PDF.dumps(out_file_handle, doc)
# attempt to re-open PDF
with open(self.output_file, "rb") as in_file_handle:
print("\treading (2) ..")
doc = PDF.loads(in_file_handle)
return True
def test_add_rubber_stamp_annotation(self):
# create output directory if it does not exist yet
if not self.output_file.parent.exists():
self.output_file.parent.mkdir()
# attempt to read PDF
doc = None
with open(self.input_file, "rb") as in_file_handle:
print("\treading (1) ..")
doc = PDF.loads(in_file_handle)
# add annotation
doc.get_page(0).append_stamp_annotation(
name=RubberStampAnnotationIconType.CONFIDENTIAL,
contents="Approved by Joris Schellekens",
color=X11Color("Red"),
rectangle=Rectangle(Decimal(128), Decimal(56), Decimal(132),
Decimal(58)),
)
# attempt to store PDF
with open(self.output_file, "wb") as out_file_handle:
print("\twriting ..")
PDF.dumps(out_file_handle, doc)
# attempt to re-open PDF
with open(self.output_file, "rb") as in_file_handle:
print("\treading (2) ..")
doc = PDF.loads(in_file_handle)
return True
def test_write_document(self):
# create output directory if it does not exist yet
if not self.output_dir.exists():
self.output_dir.mkdir()
# create document
pdf = Document()
# add page
page = Page()
pdf.append_page(page)
ul = OrderedList()
ul.add(Paragraph(text="Lorem Ipsum Dolor Sit Amet Consectetur Nunc"))
ul.add(Paragraph(text="Ipsum"))
ul.add(Paragraph(text="Dolor"))
ul.add(Paragraph(text="Sit"))
ul.add(Paragraph(text="Amet"))
layout = SingleColumnLayout(page)
layout.add(ul)
# determine output location
out_file = self.output_dir / ("output.pdf")
# attempt to store PDF
with open(out_file, "wb") as in_file_handle:
PDF.dumps(in_file_handle, pdf)
# attempt to re-open PDF
with open(out_file, "rb") as in_file_handle:
PDF.loads(in_file_handle)
def test_document(self):
# create output directory if it does not exist yet
if not self.output_file.parent.exists():
self.output_file.parent.mkdir()
# read document
doc = None
with open(self.input_file, "rb") as pdf_file_handle:
doc = PDF.loads(pdf_file_handle)
# append document
doc.append_embedded_file(
"the_raven.txt",
b"Once upon a midnight dreary, while I pondered weak and weary over many a quaint and curious volume of forgotten lore.",
)
# attempt to store PDF
with open(self.output_file, "wb") as out_file_handle:
print("\twriting ..")
PDF.dumps(out_file_handle, doc)
# attempt to re-open PDF
with open(self.output_file, "rb") as in_file_handle:
print("\treading (2) ..")
doc = PDF.loads(in_file_handle)
embedded_files = doc.get_embedded_files()
assert len(embedded_files) == 1
assert "the_raven.txt" in embedded_files
assert b"Once upon a midnight" in embedded_files["the_raven.txt"]
def test_document(self, file):
# create output directory if it does not exist yet
if not self.output_dir.exists():
self.output_dir.mkdir()
# determine output location
out_file = self.output_dir / (file.stem + "_out.pdf")
# attempt to read PDF
doc = None
with open(file, "rb") as in_file_handle:
print("\treading (1) ..")
doc = PDF.loads(in_file_handle)
# add annotation
doc.get_page(0).append_stamp_annotation(
name="Confidential",
contents="Approved by Joris Schellekens",
color=X11Color("White"),
rectangle=(Decimal(128), Decimal(128), Decimal(32), Decimal(64)),
)
# attempt to store PDF
with open(out_file, "wb") as out_file_handle:
print("\twriting ..")
PDF.dumps(out_file_handle, doc)
# attempt to re-open PDF
with open(out_file, "rb") as in_file_handle:
print("\treading (2) ..")
doc = PDF.loads(in_file_handle)
return True
def test_document(self, file):
# create output directory if it does not exist yet
if not self.output_dir.exists():
self.output_dir.mkdir()
# determine output location
out_file = self.output_dir / (file.stem + "_out.pdf")
# attempt to read PDF
doc = None
with open(file, "rb") as in_file_handle:
print("\treading (1) ..")
doc = PDF.loads(in_file_handle)
# add annotation
doc.get_page(0).append_watermark_annotation(
contents="pText",
rectangle=Rectangle(Decimal(128), Decimal(128), Decimal(64),
Decimal(64)),
)
# attempt to store PDF
with open(out_file, "wb") as out_file_handle:
print("\twriting ..")
PDF.dumps(out_file_handle, doc)
# attempt to re-open PDF
with open(out_file, "rb") as in_file_handle:
print("\treading (2) ..")
doc = PDF.loads(in_file_handle)
def test_document(self, file):
# create output directory if it does not exist yet
if not self.output_dir.exists():
self.output_dir.mkdir()
# determine output location
out_file = self.output_dir / (file.stem + "_out.pdf")
# attempt to read PDF
doc = None
with open(file, "rb") as in_file_handle:
print("\treading (1) ..")
doc = PDF.loads(in_file_handle)
# add annotation
doc.get_page(0).append_circle_annotation(
rectangle=(Decimal(128), Decimal(128), Decimal(64), Decimal(64)),
interior_color=X11Color("Plum"),
color=X11Color("Crimson"),
)
# attempt to store PDF
with open(out_file, "wb") as out_file_handle:
print("\twriting ..")
PDF.dumps(out_file_handle, doc)
# attempt to re-open PDF
with open(out_file, "rb") as in_file_handle:
print("\treading (2) ..")
doc = PDF.loads(in_file_handle)
return True
def test_write_document(self):
# create output directory if it does not exist yet
if not self.output_dir.exists():
self.output_dir.mkdir()
# create document
pdf = Document()
# add page
page = Page()
pdf.append_page(page)
ChunkOfText(
"é",
font_size=Decimal(24),
).layout(
page,
Rectangle(Decimal(100), Decimal(600), Decimal(100), Decimal(100)))
# determine output location
out_file = self.output_dir / "output.pdf"
# attempt to store PDF
with open(out_file, "wb") as in_file_handle:
PDF.dumps(in_file_handle, pdf)
# attempt to re-open PDF
with open(out_file, "rb") as in_file_handle:
PDF.loads(in_file_handle)
def test_document(self, file):
# create output directory if it does not exist yet
if not self.output_dir.exists():
self.output_dir.mkdir()
# determine output location
out_file = self.output_dir / (file.stem + "_out.pdf")
# attempt to read PDF
doc = None
with open(file, "rb") as in_file_handle:
print("\treading (1) ..")
doc = PDF.loads(in_file_handle)
# add annotation
doc.get_page(0).append_polygon_annotation(
points=LineArtFactory.sticky_note(
Rectangle(Decimal(128), Decimal(128), Decimal(64),
Decimal(64))),
stroke_color=X11Color("PowderBlue"),
)
# attempt to store PDF
with open(out_file, "wb") as out_file_handle:
print("\twriting ..")
PDF.dumps(out_file_handle, doc)
# attempt to re-open PDF
with open(out_file, "rb") as in_file_handle:
print("\treading (2) ..")
doc = PDF.loads(in_file_handle)
return True
def test_document(self, file):
# create output directory if it does not exist yet
if not self.output_dir.exists():
self.output_dir.mkdir()
# determine output location
out_file = self.output_dir / (file.stem + "_out.pdf")
# attempt to read PDF
doc = None
with open(file, "rb") as in_file_handle:
print("\treading (1) ..")
doc = PDF.loads(in_file_handle)
# add annotation
doc.get_page(0).append_text_annotation(
contents="The quick brown fox ate the lazy mouse",
rectangle=Rectangle(Decimal(128), Decimal(128), Decimal(64), Decimal(64)),
name_of_icon="Key",
open=True,
color=X11Color("Orange"),
)
# attempt to store PDF
with open(out_file, "wb") as out_file_handle:
print("\twriting ..")
PDF.dumps(out_file_handle, doc)
# attempt to re-open PDF
with open(out_file, "rb") as in_file_handle:
print("\treading (2) ..")
doc = PDF.loads(in_file_handle)
return True
def test_add_circle_annotation(self):
# create output directory if it does not exist yet
if not self.output_file.parent.exists():
self.output_file.parent.mkdir()
# attempt to read PDF
doc = None
with open(self.input_file, "rb") as in_file_handle:
print("\treading (1) ..")
doc = PDF.loads(in_file_handle)
# add annotation
doc.get_page(0).append_circle_annotation(
rectangle=Rectangle(Decimal(128), Decimal(128), Decimal(64), Decimal(64)),
stroke_color=X11Color("Plum"),
fill_color=X11Color("Crimson"),
)
# attempt to store PDF
with open(self.output_file, "wb") as out_file_handle:
print("\twriting ..")
PDF.dumps(out_file_handle, doc)
# attempt to re-open PDF
with open(self.output_file, "rb") as in_file_handle:
print("\treading (2) ..")
doc = PDF.loads(in_file_handle)
return True
def test_document(self, file):
# create output directory if it does not exist yet
if not self.output_dir.exists():
self.output_dir.mkdir()
# determine output location
out_file = self.output_dir / (file.stem + "_out.pdf")
# attempt to read PDF
delta = time.time()
doc = None
with open(file, "rb") as in_file_handle:
print("\treading (1) ..")
doc = PDF.loads(in_file_handle)
print("time elapsed : %d" % (time.time() - delta))
delta = time.time()
# attempt to store PDF
with open(out_file, "wb") as out_file_handle:
print("\twriting ..")
PDF.dumps(out_file_handle, doc)
print("time elapsed : %d" % (time.time() - delta))
delta = time.time()
# attempt to re-open PDF
with open(out_file, "rb") as in_file_handle:
print("\treading (2) ..")
doc = PDF.loads(in_file_handle)
print("time elapsed : %d" % (time.time() - delta))
delta = time.time()
return True
def test_add_link_annotation(self):
# create output directory if it does not exist yet
if not self.output_file.parent.exists():
self.output_file.parent.mkdir()
# attempt to read PDF
doc = None
with open(self.input_file, "rb") as in_file_handle:
print("\treading (1) ..")
doc = PDF.loads(in_file_handle)
# add annotation
doc.get_page(0).append_link_annotation(
page=Decimal(0),
destination_type=DestinationType.FIT,
color=X11Color("Red"),
rectangle=Rectangle(Decimal(128), Decimal(128), Decimal(64), Decimal(64)),
)
# attempt to store PDF
with open(self.output_file, "wb") as out_file_handle:
print("\twriting ..")
PDF.dumps(out_file_handle, doc)
# attempt to re-open PDF
with open(self.output_file, "rb") as in_file_handle:
print("\treading (2) ..")
doc = PDF.loads(in_file_handle)
return True
def test_document(self, file):
# create output directory if it does not exist yet
if not self.output_dir.exists():
self.output_dir.mkdir()
# determine output location
out_file = self.output_dir / (file.stem + "_out.pdf")
# attempt to read PDF
doc = None
with open(file, "rb") as in_file_handle:
print("\treading (1) ..")
doc = PDF.loads(in_file_handle)
number_of_pages = int(doc.get_document_info().get_number_of_pages())
if number_of_pages == 1:
return
# remove page
doc.pop_page(0)
# attempt to store PDF
with open(out_file, "wb") as out_file_handle:
print("\twriting ..")
PDF.dumps(out_file_handle, doc)
# attempt to re-open PDF
with open(out_file, "rb") as in_file_handle:
print("\treading (2) ..")
doc = PDF.loads(in_file_handle)
return True
def test_document(self, file):
# create output directory if it does not exist yet
if not self.output_dir.exists():
self.output_dir.mkdir()
# determine output location
out_file = self.output_dir / (file.stem + "_out.pdf")
# attempt to read PDF
doc = None
with open(file, "rb") as in_file_handle:
print("\treading (1) ..")
doc = PDF.loads(in_file_handle)
# add annotation
doc.get_page(0).append_redact_annotation(
overlay_text="Lorem Ipsum",
repeat_overlay_text=True,
fill_color=X11Color("AliceBlue"),
rectangle=Rectangle(Decimal(72.86), Decimal(486.82), Decimal(129),
Decimal(13)),
)
# attempt to store PDF
with open(out_file, "wb") as out_file_handle:
print("\twriting ..")
PDF.dumps(out_file_handle, doc)
# attempt to re-open PDF
with open(out_file, "rb") as in_file_handle:
print("\treading (2) ..")
doc = PDF.loads(in_file_handle)
return True
def _test_document(self, file):
# create output directory if it does not exist yet
if not self.output_dir.exists():
self.output_dir.mkdir()
# determine output location
out_file = self.output_dir / (file.stem + "_out.pdf")
# attempt to read PDF
doc_a = None
with open(file, "rb") as in_file_handle:
print("\treading (1) ..")
doc_a = PDF.loads(in_file_handle)
# attempt to read PDF
with open(self.input_file_b, "rb") as in_file_handle_b:
print("\treading (2) ..")
doc_b = PDF.loads(in_file_handle_b)
# concat all pages to same document
doc_c = Document()
doc_c.append_document(doc_a)
doc_c.append_document(doc_b)
# attempt to store PDF
with open(out_file, "wb") as out_file_handle:
print("\twrite ..")
PDF.dumps(out_file_handle, doc_c)
return True
def test_add_polygon_annotation(self):
# create output directory if it does not exist yet
if not self.output_file.parent.exists():
self.output_file.parent.mkdir()
# attempt to read PDF
doc = None
with open(self.input_file, "rb") as in_file_handle:
print("\treading (1) ..")
doc = PDF.loads(in_file_handle)
# add annotation
doc.get_page(0).append_polygon_annotation(
points=[
(Decimal(72), Decimal(390)),
(Decimal(242), Decimal(500)),
(Decimal(156), Decimal(390)),
],
stroke_color=X11Color("Crimson"),
)
# attempt to store PDF
with open(self.output_file, "wb") as out_file_handle:
print("\twriting ..")
PDF.dumps(out_file_handle, doc)
# attempt to re-open PDF
with open(self.output_file, "rb") as in_file_handle:
print("\treading (2) ..")
doc = PDF.loads(in_file_handle)
return True
def test_add_polyline_annotation_using_lineart_factory(self):
# create output directory if it does not exist yet
if not self.output_file.parent.exists():
self.output_file.parent.mkdir()
# attempt to read PDF
doc = None
with open(self.input_file, "rb") as in_file_handle:
print("\treading (1) ..")
doc = PDF.loads(in_file_handle)
# add annotation
doc.get_page(0).append_polyline_annotation(
points=LineArtFactory.droplet(
Rectangle(Decimal(100), Decimal(100), Decimal(100),
Decimal(100))),
stroke_color=X11Color("Crimson"),
)
# attempt to store PDF
with open(self.output_file, "wb") as out_file_handle:
print("\twriting ..")
PDF.dumps(out_file_handle, doc)
# attempt to re-open PDF
with open(self.output_file, "rb") as in_file_handle:
print("\treading (2) ..")
doc = PDF.loads(in_file_handle)
return True
def _test_document(self, file):
# create output directory if it does not exist yet
if not self.output_dir.exists():
self.output_dir.mkdir()
# determine output location
out_file = self.output_dir / (file.stem + "_out.pdf")
# attempt to read PDF
doc = None
with open(file, "rb") as in_file_handle:
print("\treading (1) ..")
doc = PDF.loads(in_file_handle)
# determine free space
space_finder = FreeSpaceFinder(doc.get_page(0))
# debug purposes
if self.in_debug:
for i in range(0, len(space_finder.grid)):
for j in range(0, len(space_finder.grid[i])):
if space_finder.grid[i][j]:
continue
w = Decimal(space_finder.grid_resolution)
x = Decimal(i) * w
y = Decimal(j) * w
doc.get_page(0).append_square_annotation(
Rectangle(x, y, w, w), stroke_color=X11Color("Salmon"))
# add annotation
w, h = doc.get_page(0).get_page_info().get_size()
free_rect = space_finder.find_free_space(
Rectangle(
Decimal(w / Decimal(2)),
Decimal(h * Decimal(2)),
Decimal(64),
Decimal(64),
))
if free_rect is not None:
doc.get_page(0).append_square_annotation(
rectangle=free_rect,
stroke_color=HexColor("#F75C03"),
fill_color=HexColor("#04A777"),
)
# attempt to store PDF
with open(out_file, "wb") as out_file_handle:
print("\twriting ..")
PDF.dumps(out_file_handle, doc)
# attempt to re-open PDF
with open(out_file, "rb") as in_file_handle:
print("\treading (2) ..")
doc = PDF.loads(in_file_handle)
return True
def test_write_document(self):
# create output directory if it does not exist yet
if not self.output_dir.exists():
self.output_dir.mkdir()
# create document
pdf = Document()
# add page
page = Page()
pdf.append_page(page)
layout = SingleColumnLayout(page)
# title
layout.add(
Paragraph(
"Lissajours Line Art",
font_size=Decimal(20),
font_color=X11Color("Blue"),
))
# table
N = 7
fill_colors = [
HSVColor(Decimal(x / N), Decimal(1), Decimal(1))
for x in range(0, N)
]
stroke_colors = [HSVColor.darker(x) for x in fill_colors]
fixed_bb = Rectangle(Decimal(0), Decimal(0), Decimal(100),
Decimal(100))
t = Table(number_of_rows=N, number_of_columns=N)
for i in range(0, N):
for j in range(0, N):
t.add(
Shape(
LineArtFactory.lissajours(fixed_bb, i + 1, j + 1),
fill_color=fill_colors[(i + j) % N],
stroke_color=stroke_colors[(i + j) % N],
line_width=Decimal(2),
))
t.set_padding_on_all_cells(Decimal(10), Decimal(10), Decimal(10),
Decimal(10))
layout.add(t)
# determine output location
out_file = self.output_dir / ("output.pdf")
# attempt to store PDF
with open(out_file, "wb") as in_file_handle:
PDF.dumps(in_file_handle, pdf)
# attempt to re-open PDF
with open(out_file, "rb") as in_file_handle:
PDF.loads(in_file_handle)
def test_document(self, file):
# create output directory if it does not exist yet
if not self.output_dir.exists():
self.output_dir.mkdir()
# get text
txt_ground_truth_file = self.input_dir / (file.stem + ".txt")
txt_ground_truth = ""
with open(txt_ground_truth_file, "r") as txt_ground_truth_file_handle:
txt_ground_truth = txt_ground_truth_file_handle.read()
words = [x for x in re.split("[^a-zA-Z]+", txt_ground_truth) if len(x) > 5]
w = words[5] if len(words) > 5 else None
if w is None:
return True
# determine output location
out_file = self.output_dir / (file.stem + "_" + w + "_out.pdf")
# attempt to read PDF
doc = None
l = RegularExpressionTextExtraction(w)
with open(file, "rb") as in_file_handle:
print("\treading (1) ..")
doc = PDF.loads(in_file_handle, [l])
# add annotation
print(
"\tAdding %d annotations"
% len(l.get_matched_text_render_info_events_per_page(0))
)
for e in l.get_matched_text_render_info_events_per_page(0):
baseline = e.get_baseline()
doc.get_page(0).append_square_annotation(
rectangle=Rectangle(
Decimal(baseline.x0),
Decimal(baseline.y0 - 2),
Decimal(baseline.x1 - baseline.x0),
Decimal(12),
),
stroke_color=X11Color("Firebrick"),
)
# attempt to store PDF
with open(out_file, "wb") as out_file_handle:
print("\twriting ..")
PDF.dumps(out_file_handle, doc)
# attempt to re-open PDF
with open(out_file, "rb") as in_file_handle:
print("\treading (2) ..")
doc = PDF.loads(in_file_handle)
return True
def test_write_document(self):
# create output directory if it does not exist yet
if not self.output_dir.exists():
self.output_dir.mkdir()
# create document
pdf = Document()
# add page
page = Page()
pdf.append_page(page)
layout = SingleColumnLayout(page)
# write title
layout.add(
Paragraph("Nonogram",
font_size=Decimal(20),
font_color=X11Color("YellowGreen")))
# write text
layout.add(
Paragraph(
"""
Nonograms, also known as Paint by Numbers, Picross, Griddlers, Pic-a-Pix, and various other names,
are picture logic puzzles in which cells in a grid must be colored or left blank according to numbers
at the side of the grid to reveal a hidden picture.
In this puzzle type, the numbers are a form of discrete tomography that measures how many
unbroken lines of filled-in squares there are in any given row or column.
For example, a clue of "4 8 3" would mean there are sets of four, eight, and three filled squares,
in that order, with at least one blank square between successive sets.
""",
font_color=X11Color("SlateGray"),
font_size=Decimal(8),
))
# write nonogram
ng = Nonogram(
# "https://i.pinimg.com/originals/f8/23/88/f823882e7c5fa42790e78f43ecf7e8bf.jpg"
"https://cdn.shopify.com/s/files/1/2123/8425/products/166422700-LRG_242a4c8b-cad5-476e-afd1-c8b882d48fc2_530x.jpg"
)
layout.add(ng)
# determine output location
out_file = self.output_dir / ("output.pdf")
# attempt to store PDF
with open(out_file, "wb") as in_file_handle:
PDF.dumps(in_file_handle, pdf)
# attempt to re-open PDF
with open(out_file, "rb") as in_file_handle:
PDF.loads(in_file_handle)
def test_document(self, file):
# create output directory if it does not exist yet
if not self.output_dir.exists():
self.output_dir.mkdir()
# determine output location
out_file = self.output_dir / (file.stem + "_out.pdf")
# attempt to read PDF
doc = None
with open(file, "rb") as in_file_handle:
print("\treading (1) ..")
doc = PDF.loads(in_file_handle)
# add annotation
for index, name in enumerate([
"Approved",
"Experimental",
"NotApproved",
"Asis",
"Expired",
"NotForPublicRelease",
"Confidential",
"Final",
"Sold",
"Departmental",
"ForComment",
"TopSecret",
"Draft",
"ForPublicRelease",
]):
doc.get_page(0).append_stamp_annotation(
name=name,
contents="Approved by Joris Schellekens",
color=X11Color("White"),
rectangle=Rectangle(Decimal(128), Decimal(128 + index * 34),
Decimal(64), Decimal(32)),
)
# attempt to store PDF
with open(out_file, "wb") as out_file_handle:
print("\twriting ..")
PDF.dumps(out_file_handle, doc)
# attempt to re-open PDF
with open(out_file, "rb") as in_file_handle:
print("\treading (2) ..")
doc = PDF.loads(in_file_handle)
return True
def test_document(self, file):
# create output directory if it does not exist yet
if not self.output_dir.exists():
self.output_dir.mkdir()
# determine output location
out_file = self.output_dir / (file.stem + "_out.pdf")
doc = None
with open(file, "rb") as pdf_file_handle:
doc = PDF.loads(pdf_file_handle)
# remove first annotation
if "Annots" in doc.get_page(0):
annots = doc.get_page(0)["Annots"]
annots[0]["P"] = None
doc.get_page(0)["Annots"] = annots[1:0]
pass
# attempt to store PDF
with open(out_file, "wb") as out_file_handle:
print("\twriting ..")
PDF.dumps(out_file_handle, doc)
return True
def test_document(self, file):
# create output directory if it does not exist yet
if not self.output_dir.exists():
self.output_dir.mkdir()
with open(file, "rb") as pdf_file_handle:
l = RegularExpressionTextExtraction("[sS]orbitol")
doc = PDF.loads(pdf_file_handle, [l])
# export matches
output_file = self.output_dir / (file.stem + ".json")
with open(output_file, "w") as json_file_handle:
obj = [
{
"text": x.get_text(),
"x0": int(x.get_baseline().x0),
"y0": int(x.get_baseline().y0),
"x1": int(x.get_baseline().x1),
"y1": int(x.get_baseline().y1),
}
for x in l.get_matched_text_render_info_events_per_page(0)
]
json_file_handle.write(json.dumps(obj, indent=4))
return True
def _test_document(self, file):
# create output directory if it does not exist yet
if not self.output_file.parent.exists():
self.output_file.parent.mkdir()
with open(file, "rb") as pdf_file_handle:
l = ColorSpectrumExtraction()
doc = PDF.loads(pdf_file_handle, [l])
colors = []
for t in l.get_colors_per_page(0, limit=16):
colors.append(
{
"red": float(t[0].red),
"green": float(t[0].green),
"blue": float(t[0].blue),
"count": int(t[1]),
}
)
# write output
with open(self.output_file, "w") as json_file_handle:
json_file_handle.write(json.dumps(colors))
return True
def test_document(self, file) -> bool:
doc = None
with open(file, "rb") as pdf_file_handle:
doc = None
with open(file, "rb") as pdf_file_handle:
doc = PDF.loads(pdf_file_handle)
if "XRef" not in doc:
return False
if "Trailer" not in doc["XRef"]:
return False
if "Info" not in doc["XRef"]["Trailer"]:
doc["XRef"]["Trailer"][Name("Info")] = Dictionary()
# change producer
doc["XRef"]["Trailer"]["Info"]["Producer"] = String("pText")
# determine output location
out_file = self.output_dir / (file.stem + "_out.pdf")
with open(out_file, "wb") as pdf_file_handle:
PDF.dumps(out_file, doc)
return True
def test_document(self, file):
with open(file, "rb") as pdf_file_handle:
l = FontExtraction()
doc = PDF.loads(pdf_file_handle, [l])
for fn in l.get_font_names_per_page(0):
print(fn)
return True
def test_document(self, file) -> bool:
# create output directory if it does not exist yet
if not self.output_dir.exists():
self.output_dir.mkdir()
doc = None
with open(file, "rb") as pdf_file_handle:
doc = PDF.loads(pdf_file_handle)
if "XRef" not in doc:
return False
if "Trailer" not in doc["XRef"]:
return False
if "Info" not in doc["XRef"]["Trailer"]:
doc["XRef"]["Trailer"][Name("Info")] = Dictionary()
# change author
doc["XRef"]["Trailer"]["Info"]["Author"] = String("Joris Schellekens")
# determine output location
out_file = self.output_dir / (file.stem + "_out.pdf")
with open(out_file, "wb") as pdf_file_handle:
PDF.dumps(pdf_file_handle, doc)
return True
def test_extract_all_text(self):
l = SimpleTextExtraction()
with open(
Path("/home/joris/Code/pdf-corpus/0600.pdf"), "rb"
) as pdf_file_handle:
doc = PDF.loads(pdf_file_handle, [l])
print(l.get_text(0))
