Rename table_bbox (singular) to table_areas
The object is an index of bounding boxes, in some cases given by users. It's called areas in one section of the code making it systematic.pull/127/head
parent
270c76a3e7
commit
49d3f0f3aa
|
|
@ -207,7 +207,7 @@ class Lattice(BaseParser):
|
|||
t.cells[i][j].text = t.cells[i - 1][j].text
|
||||
return t
|
||||
|
||||
def _generate_table_bbox(self):
|
||||
def _generate_table_areas(self):
|
||||
def scale_areas(areas):
|
||||
scaled_areas = []
|
||||
for area in areas:
|
||||
|
|
@ -258,7 +258,7 @@ class Lattice(BaseParser):
|
|||
)
|
||||
|
||||
contours = find_contours(vertical_mask, horizontal_mask)
|
||||
table_bbox = find_joints(contours, vertical_mask, horizontal_mask)
|
||||
table_areas = find_joints(contours, vertical_mask, horizontal_mask)
|
||||
else:
|
||||
vertical_mask, vertical_segments = find_lines(
|
||||
self.threshold,
|
||||
|
|
@ -274,20 +274,20 @@ class Lattice(BaseParser):
|
|||
)
|
||||
|
||||
areas = scale_areas(self.table_areas)
|
||||
table_bbox = find_joints(areas, vertical_mask, horizontal_mask)
|
||||
table_areas = find_joints(areas, vertical_mask, horizontal_mask)
|
||||
|
||||
self.table_bbox_unscaled = copy.deepcopy(table_bbox)
|
||||
self.table_areas_unscaled = copy.deepcopy(table_areas)
|
||||
|
||||
[
|
||||
self.table_bbox,
|
||||
self.table_areas,
|
||||
self.vertical_segments,
|
||||
self.horizontal_segments
|
||||
] = scale_image(
|
||||
table_bbox, vertical_segments, horizontal_segments, pdf_scalers
|
||||
table_areas, vertical_segments, horizontal_segments, pdf_scalers
|
||||
)
|
||||
|
||||
def _generate_columns_and_rows(self, table_idx, tk):
|
||||
# select elements which lie within table_bbox
|
||||
# select elements which lie within table_areas
|
||||
t_bbox = {}
|
||||
v_s, h_s = segments_in_bbox(
|
||||
tk, self.vertical_segments, self.horizontal_segments
|
||||
|
|
@ -300,7 +300,7 @@ class Lattice(BaseParser):
|
|||
|
||||
self.t_bbox = t_bbox
|
||||
|
||||
cols, rows = zip(*self.table_bbox[tk])
|
||||
cols, rows = zip(*self.table_areas[tk])
|
||||
cols, rows = list(cols), list(rows)
|
||||
cols.extend([tk[0], tk[2]])
|
||||
rows.extend([tk[1], tk[3]])
|
||||
|
|
@ -366,7 +366,7 @@ class Lattice(BaseParser):
|
|||
_text.extend([(t.x0, t.y0, t.x1, t.y1) for t in self.horizontal_text])
|
||||
_text.extend([(t.x0, t.y0, t.x1, t.y1) for t in self.vertical_text])
|
||||
table._text = _text
|
||||
table._image = (self.pdf_image, self.table_bbox_unscaled)
|
||||
table._image = (self.pdf_image, self.table_areas_unscaled)
|
||||
table._segments = (self.vertical_segments, self.horizontal_segments)
|
||||
table._textedges = None
|
||||
|
||||
|
|
@ -391,12 +391,12 @@ class Lattice(BaseParser):
|
|||
return []
|
||||
|
||||
self._generate_image_file()
|
||||
self._generate_table_bbox()
|
||||
self._generate_table_areas()
|
||||
|
||||
_tables = []
|
||||
# sort tables based on y-coord
|
||||
for table_idx, tk in enumerate(
|
||||
sorted(self.table_bbox.keys(), key=lambda x: x[1], reverse=True)
|
||||
sorted(self.table_areas.keys(), key=lambda x: x[1], reverse=True)
|
||||
):
|
||||
cols, rows, v_s, h_s = self._generate_columns_and_rows(
|
||||
table_idx, tk)
|
||||
|
|
|
|||
|
|
@ -299,14 +299,14 @@ class Stream(BaseParser):
|
|||
relevant_textedges = textedges.get_relevant()
|
||||
self.textedges.extend(relevant_textedges)
|
||||
# guess table areas using textlines and relevant edges
|
||||
table_bbox = textedges.get_table_areas(textlines, relevant_textedges)
|
||||
table_areas = textedges.get_table_areas(textlines, relevant_textedges)
|
||||
# treat whole page as table area if no table areas found
|
||||
if not table_bbox:
|
||||
table_bbox = {(0, 0, self.pdf_width, self.pdf_height): None}
|
||||
if not table_areas:
|
||||
table_areas = {(0, 0, self.pdf_width, self.pdf_height): None}
|
||||
|
||||
return table_bbox
|
||||
return table_areas
|
||||
|
||||
def _generate_table_bbox(self):
|
||||
def _generate_table_areas(self):
|
||||
self.textedges = []
|
||||
if self.table_areas is None:
|
||||
all_text_segments = self.horizontal_text + self.vertical_text
|
||||
|
|
@ -325,20 +325,20 @@ class Stream(BaseParser):
|
|||
(x1, y2, x2, y1), all_text_segments)
|
||||
text_segments.extend(region_text)
|
||||
# find tables based on nurminen's detection algorithm
|
||||
table_bbox = self._nurminen_table_detection(text_segments)
|
||||
table_areas = self._nurminen_table_detection(text_segments)
|
||||
else:
|
||||
table_bbox = {}
|
||||
table_areas = {}
|
||||
for area in self.table_areas:
|
||||
x1, y1, x2, y2 = area.split(",")
|
||||
x1 = float(x1)
|
||||
y1 = float(y1)
|
||||
x2 = float(x2)
|
||||
y2 = float(y2)
|
||||
table_bbox[(x1, y2, x2, y1)] = None
|
||||
self.table_bbox = table_bbox
|
||||
table_areas[(x1, y2, x2, y1)] = None
|
||||
self.table_areas = table_areas
|
||||
|
||||
def _generate_columns_and_rows(self, table_idx, tk):
|
||||
# select elements which lie within table_bbox
|
||||
# select elements which lie within table_areas
|
||||
t_bbox = {}
|
||||
t_bbox["horizontal"] = text_in_bbox(tk, self.horizontal_text)
|
||||
t_bbox["vertical"] = text_in_bbox(tk, self.vertical_text)
|
||||
|
|
@ -464,7 +464,7 @@ class Stream(BaseParser):
|
|||
_text.extend([(t.x0, t.y0, t.x1, t.y1) for t in self.vertical_text])
|
||||
table._text = _text
|
||||
self.generate_image()
|
||||
table._image = (self.pdf_image, self.table_bbox)
|
||||
table._image = (self.pdf_image, self.table_areas)
|
||||
table._segments = None
|
||||
table._textedges = self.textedges
|
||||
|
||||
|
|
@ -492,13 +492,13 @@ class Stream(BaseParser):
|
|||
return []
|
||||
|
||||
# Identify plausible areas within the doc where tables lie,
|
||||
# populate table_bbox keys with these areas.
|
||||
self._generate_table_bbox()
|
||||
# populate table_areas keys with these areas.
|
||||
self._generate_table_areas()
|
||||
|
||||
_tables = []
|
||||
# sort tables based on y-coord
|
||||
for table_idx, tk in enumerate(
|
||||
sorted(self.table_bbox.keys(), key=lambda x: x[1], reverse=True)
|
||||
sorted(self.table_areas.keys(), key=lambda x: x[1], reverse=True)
|
||||
):
|
||||
cols, rows = self._generate_columns_and_rows(table_idx, tk)
|
||||
table = self._generate_table(table_idx, cols, rows)
|
||||
|
|
|
|||
|
|
@ -121,7 +121,7 @@ class PlotMethods(object):
|
|||
fig : matplotlib.fig.Figure
|
||||
|
||||
"""
|
||||
img, table_bbox = table._image
|
||||
img, table_areas = table._image
|
||||
_FOR_LATTICE = table.flavor == "lattice"
|
||||
fig = plt.figure()
|
||||
ax = fig.add_subplot(111, aspect="equal")
|
||||
|
|
@ -137,7 +137,7 @@ class PlotMethods(object):
|
|||
)
|
||||
)
|
||||
|
||||
for t in table_bbox.keys():
|
||||
for t in table_areas.keys():
|
||||
ax.add_patch(
|
||||
patches.Rectangle(
|
||||
(t[0], t[1]), t[2] - t[0], t[3] - t[1],
|
||||
|
|
@ -204,13 +204,13 @@ class PlotMethods(object):
|
|||
fig : matplotlib.fig.Figure
|
||||
|
||||
"""
|
||||
img, table_bbox = table._image
|
||||
img, table_areas = table._image
|
||||
fig = plt.figure()
|
||||
ax = fig.add_subplot(111, aspect="equal")
|
||||
x_coord = []
|
||||
y_coord = []
|
||||
for k in table_bbox.keys():
|
||||
for coord in table_bbox[k]:
|
||||
for k in table_areas.keys():
|
||||
for coord in table_areas[k]:
|
||||
x_coord.append(coord[0])
|
||||
y_coord.append(coord[1])
|
||||
ax.plot(x_coord, y_coord, "ro")
|
||||
|
|
|
|||
|
|
@ -192,6 +192,7 @@ def test_stream_vertical_header():
|
|||
|
||||
filename = os.path.join(testdir, "vertical_header.pdf")
|
||||
tables = camelot.read_pdf(filename, flavor="stream")
|
||||
assert len(tables) == 1
|
||||
assert_frame_equal(df, tables[0].df)
|
||||
|
||||
|
||||
|
|
|
|||
Loading…
Reference in New Issue