def __prepare_display(self):
frames = []
lines = []
for rect in self.ltRects:
if (rect.horizontal()
and rect.height() > 8) or (rect.vertical()
and rect.width() > 8):
table = SingleCellTable([])
table.rect = rect
frames.append(table)
else:
lines.append(rect)
orphans = []
while len(lines) > 0:
cluster = pdftable.cluster_rects(lines)
if len(cluster) >= 4:
try:
frames.append(pdftable.Table(cluster))
continue
except:
pass
orphans += cluster
curves = sorted(self.curves +
[pdftable.Curve(o.points()) for o in orphans],
key=cmp_to_key(sort_topdown_ltr))
textLines = sorted(self.textLines, key=cmp_to_key(sort_topdown_ltr))
# explicit tables
tables = []
for table in frames:
orphans = []
bounds = table.bounds()
for i in range(0, len(textLines)):
line = textLines[i]
if bounds.contains(line.bounds()):
# Some pages have their "NOTES" section embedded inside the
# table rectangle. What were you thinking, Intel?
if line.font_name() == "NeoSansIntelMedium" and str(
line).lower().startswith("notes"):
orphans += textLines[i:]
break
table.get_at_pixel(line.rect.xmid(),
line.rect.ymid()).append(line)
else:
orphans.append(line)
textLines = orphans
tables.append(table)
# exception tables
orphans = []
table_data = []
is_table_section = False
expected_format = None
for line in textLines:
if line.font_name() == "NeoSansIntelMedium":
orphans.append(line)
title = str(line).strip().lower()
if title[-10:] == "exceptions":
is_table_section = True
expected_format = exceptions_format__
elif title == "fpu flags affected":
is_table_section = True
expected_format = fpu_flags_format__
if is_table_section and len(table_data) > 0:
tables.append(SingleCellTable(table_data))
table_data = []
continue
if is_table_section:
if line.bounds().x1() > 50:
table_data.append(line)
elif expected_format.search(str(line)) == None:
orphans.append(line)
if len(table_data) > 0:
tables.append(SingleCellTable(table_data))
table_data = []
else:
table_data.append(line)
else:
orphans.append(line)
if len(table_data) > 0:
tables.append(SingleCellTable(table_data))
# tables versus figures (versus useless frames)
all_tables = sorted(tables, key=lambda x: x.bounds().area())
tables = set()
figures = set()
sublevel_figures = set()
i = 0
while i < len(all_tables):
smaller = all_tables[i]
if smaller.rows() != 1 or smaller.columns() != 1:
tables.add(smaller)
else:
j = i + 1
smaller_bounds = smaller.bounds()
while j < len(all_tables):
bigger = all_tables[j]
if bigger.bounds().contains(smaller_bounds):
bigger.get_at_pixel(
smaller_bounds.xmid(),
smaller_bounds.ymid()).append(smaller)
figures.add(bigger)
figures.add(smaller)
sublevel_figures.add(smaller)
break
j += 1
else:
tables.add(smaller)
i += 1
top_figures = [Figure(t) for t in figures - sublevel_figures]
top_tables = list(tables - figures)
orphanCurves = []
for figure in top_figures:
for curve in curves:
if figure.bounds().contains(curve.bounds()):
figure.data.get_at(0, 0).append(curve)
else:
orphanCurves.append(curve)
curves = orphanCurves
orphanCurves = []
i = 0
while i < len(top_tables):
count = top_tables[i].item_count()
if count > 1: i += 1
else:
if count == 1:
orphans += top_tables[i].get_at(0, 0)
top_tables.pop(i)
# lists
textLines = self.__merge_text(orphans)
orphans = []
lists = []
this_list = []
i = 0
while i < len(textLines):
line = textLines[i]
if line.chars[0].get_text() == "•":
if len(line.chars) == 1:
i += 1
line = textLines[i]
else:
for j in range(1, len(line.chars)):
if not line.chars[j].get_text().isspace(): break
line.chars = line.chars[j:]
this_list.append(line)
else:
if len(this_list) > 0:
lists.append(pdftable.List(this_list))
this_list = []
orphans.append(line)
i += 1
displayable = self.__merge_text(orphans) + top_tables + top_figures
displayable.sort(key=cmp_to_key(sort_topdown_ltr))
return displayable
def process_curve(self, curve):
curve = pdftable.Curve([self.__fix_point(p) for p in curve.pts])
self.curves.append(curve)