Question on ability to read checkbox data that is not a widget

[dgrunspan]

Is there a way to identify a checkbox and read its value?
Its not a classic form with widgets
I get empty space instead
Attached an example PDF

150109DSP-Milw-505-90D.pdf

Thx
Danny

[JorjMcKie]

Atypical "Discussions" post -> converting ...

[JorjMcKie]

If these things are no widget, then there are 2 more possibilities:

1. special characters chr(0x2610), chr(0x2611), chr(0x2612), or
2. vector graphics

In your case, these are no characters, so I checked for drawings like follows:

# make a rectangle starting from "Gender:" and ending at right page border:
r=page.search_for("gender:")[0]
r.x1=page.rect.width  # extend to right border
r.y0-=5  # decrease upper bound a bit
r.y1+=5  # increase lower bound somewhat
print(page.get_text(clip=r))  # extract text within to confirm we are ok
Gender:  
 Female   
 Male 

# we can also determine the rectangle for the words:
pprint(page.get_text("words", clip=r))
[(204.47999572753906,
  123.88499450683594,
  237.09597778320312,
  136.25099182128906,
  'Gender:',
  0,
  0,
  0),
 (256.0799865722656,
  123.88499450683594,
  286.1669921875,
  136.25099182128906,
  'Female',
  0,
  1,
  0),
 (307.55999755859375,
  123.88499450683594,
  326.9639892578125,
  136.25099182128906,
  'Male',
  0,
  2,
  0)]

Now extract drawings within that rect and look at them

paths = [p for p in page.get_drawings() if p["rect"] in r]  # vector graphics within my rectangle
len(paths)  # we have 4 different ones!
4
from pprint import pprint
pprint(paths[0])  # first one is a rectangle
{'closePath': False,
 'color': (0.0, 0.0, 0.0),
 'dashes': '[] 0',
 'even_odd': None,
 'fill': None,
 'fill_opacity': None,
 'items': [('re',
            Rect(243.1199951171875, 125.280029296875, 252.36000061035156, 134.52003479003906),
            -1)],
 'layer': '',
 'lineCap': (2, 2, 2),
 'lineJoin': 0.0,
 'rect': Rect(243.1199951171875, 125.280029296875, 252.36000061035156, 134.52003479003906),
 'seqno': 41,
 'stroke_opacity': 1.0,
 'type': 's',
 'width': 0.7200000286102295}
pprint(paths[1])  # second one is another rectangle
{'closePath': False,
 'color': (0.0, 0.0, 0.0),
 'dashes': '[] 0',
 'even_odd': None,
 'fill': None,
 'fill_opacity': None,
 'items': [('re',
            Rect(294.6000061035156, 125.280029296875, 303.8399963378906, 134.52003479003906),
            -1)],
 'layer': '',
 'lineCap': (2, 2, 2),
 'lineJoin': 0.0,
 'rect': Rect(294.6000061035156, 125.280029296875, 303.8399963378906, 134.52003479003906),
 'seqno': 44,
 'stroke_opacity': 1.0,
 'type': 's',
 'width': 0.7200000286102295}
pprint(paths[2])  # 3rd vector graphic is a line
{'closePath': False,
 'color': (0.0, 0.0, 0.0),
 'dashes': '[] 0',
 'even_odd': None,
 'fill': None,
 'fill_opacity': None,
 'items': [('l',
            Point(294.4800109863281, 125.15997314453125),
            Point(303.96002197265625, 134.6399688720703))],
 'layer': '',
 'lineCap': (2, 2, 2),
 'lineJoin': 0.0,
 'rect': Rect(294.4800109863281, 125.15997314453125, 303.96002197265625, 134.6399688720703),
 'seqno': 45,
 'stroke_opacity': 1.0,
 'type': 's',
 'width': 0.47999998927116394}
pprint(paths[3])  # number 4 another line
{'closePath': False,
 'color': (0.0, 0.0, 0.0),
 'dashes': '[] 0',
 'even_odd': None,
 'fill': None,
 'fill_opacity': None,
 'items': [('l',  # line
            Point(303.9599914550781, 125.15997314453125),  # start point
            Point(294.47998046875, 134.6399688720703))],  # end point
 'layer': '',
 'lineCap': (2, 2, 2),
 'lineJoin': 0.0,
 'rect': Rect(294.47998046875, 125.15997314453125, 303.9599914550781, 134.6399688720703),
 'seqno': 46,
 'stroke_opacity': 1.0,
 'type': 's',
 'width': 0.47999998927116394}

Maybe it looks complicated ... blame the PDF creator!
You now would probably look at text rectangles of "Male" and "Female" and compare the positions of the 2 identified rectangles. Then check if one of them contains (or intersects) the 2 lines.

[JorjMcKie]

I have done lots of similar analyses in that area and seen all kinds of alternative solutions instead of widgets. If someone doesn't want to make the PDF updateable via widgets ... ok. But why not using characters '☐', '☑', '☒' instead?
I have also seen combinations: a square drawing with a character "x" placed inside it, little inline images even (I did not bother to mention that above), etc.

[dgrunspan]

Ok, Thx