Add table_area
[MRG] Add table_area
This commit is contained in:
Vinayak Mehta
+30
-24
@@ -40,9 +40,9 @@ options:
|
||||
-W, --wmargin <wmargin> Word margin. Insert blank spaces between chars
|
||||
if distance between words is greater than word
|
||||
margin. [default: 0.1]
|
||||
-S, --save-info Save parsing info for each page to a file.
|
||||
-S, --print-stats List stats on the parsing process.
|
||||
-T, --save-stats Save stats to a file.
|
||||
-X, --plot <dist> Plot distributions. (page,all,rc)
|
||||
-Z, --summary Summarize metrics.
|
||||
|
||||
camelot methods:
|
||||
lattice Looks for lines between data.
|
||||
@@ -55,19 +55,21 @@ lattice_doc = """
|
||||
Lattice method looks for lines between text to form a table.
|
||||
|
||||
usage:
|
||||
camelot lattice [options] [--] <file>
|
||||
camelot lattice [-t <tarea>...] [-F <fill>...] [-j <jtol>...]
|
||||
[-m <mtol>...] [options] [--] <file>
|
||||
|
||||
options:
|
||||
-t, --tarea <tarea> Specific table areas to analyze.
|
||||
-F, --fill <fill> Fill data in horizontal and/or vertical spanning
|
||||
cells. Example: -F h, -F v, -F hv
|
||||
-s, --scale <scale> Scaling factor. Large scaling factor leads to
|
||||
smaller lines being detected. [default: 15]
|
||||
-i, --invert Invert pdf image to make sure that lines are
|
||||
in foreground.
|
||||
-j, --jtol <jtol> Tolerance to account for when comparing joint
|
||||
and line coordinates. [default: 2]
|
||||
-m, --mtol <mtol> Tolerance to account for when merging lines
|
||||
which are very close. [default: 2]
|
||||
-s, --scale <scale> Scaling factor. Large scaling factor leads to
|
||||
smaller lines being detected. [default: 15]
|
||||
-i, --invert Invert pdf image to make sure that lines are
|
||||
in foreground.
|
||||
-d, --debug <debug> Debug by visualizing pdf geometry.
|
||||
(contour,line,joint,table) Example: -d table
|
||||
"""
|
||||
@@ -76,12 +78,14 @@ stream_doc = """
|
||||
Stream method looks for whitespaces between text to form a table.
|
||||
|
||||
usage:
|
||||
camelot stream [options] [--] <file>
|
||||
camelot stream [-t <tarea>...] [-c <columns>...] [-n <ncols>...] [-y <ytol>...]
|
||||
[-m <mtol>...] [options] [--] <file>
|
||||
|
||||
options:
|
||||
-n, --ncols <ncols> Number of columns. [default: 0]
|
||||
-t, --tarea <tarea> Specific table areas to analyze.
|
||||
-c, --columns <columns> Comma-separated list of column x-coordinates.
|
||||
Example: -c 10.1,20.2,30.3
|
||||
-n, --ncols <ncols> Number of columns. [default: -1]
|
||||
-y, --ytol <ytol> Tolerance to account for when grouping rows
|
||||
together. [default: 2]
|
||||
-m, --mtol <mtol> Tolerance to account for when merging columns
|
||||
@@ -166,7 +170,7 @@ def plot_rc_piechart(data, output):
|
||||
plt.savefig(''.join([output, '_rc.png']), dpi=300)
|
||||
|
||||
|
||||
def summary(data, p_time):
|
||||
def print_stats(data, p_time):
|
||||
from operator import itemgetter
|
||||
from itertools import groupby
|
||||
|
||||
@@ -331,17 +335,18 @@ if __name__ == '__main__':
|
||||
else:
|
||||
p.append({'start': int(r), 'end': int(r)})
|
||||
|
||||
margin_tuple = (float(args['--cmargin']), float(args['--lmargin']),
|
||||
margins = (float(args['--cmargin']), float(args['--lmargin']),
|
||||
float(args['--wmargin']))
|
||||
if args['<method>'] == 'lattice':
|
||||
try:
|
||||
manager = Pdf(Lattice(
|
||||
table_area=args['--tarea'],
|
||||
fill=args['--fill'],
|
||||
jtol=[int(j) for j in args['--jtol']],
|
||||
mtol=[int(m) for m in args['--mtol']],
|
||||
scale=int(args['--scale']),
|
||||
invert=args['--invert'],
|
||||
jtol=int(args['--jtol']),
|
||||
mtol=int(args['--mtol']),
|
||||
pdf_margin=margin_tuple,
|
||||
margins=margins,
|
||||
debug=args['--debug']),
|
||||
filename,
|
||||
pagenos=p,
|
||||
@@ -374,10 +379,10 @@ if __name__ == '__main__':
|
||||
if 'rc' in plot_type:
|
||||
plot_rc_piechart(data, pngname)
|
||||
|
||||
if args['--summary']:
|
||||
summary(data, processing_time)
|
||||
if args['--print-stats']:
|
||||
print_stats(data, processing_time)
|
||||
|
||||
if args['--save-info']:
|
||||
if args['--save-stats']:
|
||||
if args['--output']:
|
||||
scorename = os.path.join(args['--output'], os.path.basename(scorename))
|
||||
with open(scorename, 'w') as score_file:
|
||||
@@ -402,11 +407,12 @@ if __name__ == '__main__':
|
||||
elif args['<method>'] == 'stream':
|
||||
try:
|
||||
manager = Pdf(Stream(
|
||||
ncolumns=int(args['--ncols']),
|
||||
table_area=args['--tarea'],
|
||||
columns=args['--columns'],
|
||||
ytol=int(args['--ytol']),
|
||||
mtol=int(args['--mtol']),
|
||||
pdf_margin=margin_tuple,
|
||||
ncolumns=[int(nc) for nc in args['--ncols']],
|
||||
ytol=[int(y) for y in args['--ytol']],
|
||||
mtol=[int(m) for m in args['--mtol']],
|
||||
margins=margins,
|
||||
debug=args['--debug']),
|
||||
filename,
|
||||
pagenos=p,
|
||||
@@ -439,10 +445,10 @@ if __name__ == '__main__':
|
||||
if 'rc' in plot_type:
|
||||
plot_rc_piechart(data, pngname)
|
||||
|
||||
if args['--summary']:
|
||||
summary(data, processing_time)
|
||||
if args['--print-stats']:
|
||||
print_stats(data, processing_time)
|
||||
|
||||
if args['--save-info']:
|
||||
if args['--save-stats']:
|
||||
if args['--output']:
|
||||
scorename = os.path.join(args['--output'], os.path.basename(scorename))
|
||||
with open(scorename, 'w') as score_file:
|
||||
|
||||
Reference in New Issue
Block a user