min_height/width to filter out tables? about camelot

[JosePVB]

First of all, let me say that this is a great library! Parsing table data out of PDFs is very hard and this library makes this much more accessible.

I'd like express my interest in this feature as well. For the last page of the following PDF, http://web2.gov.mb.ca/bills/40-5/billstatus.en.pdf, I tried the suggested table_regions argument that was suggested in atlanhq/camelot#357, but I encountered the IndexError mentioned in that same issue, as camelot identified two tables within the specified table_regions.

Here is the exact call:

>>> columns = ['93,242,305,350,395,468,517,566,629,693']
>>> tables = camelot.read_pdf('camb-40-5.pdf', flavor='stream', edge_tol=200, pages='5', columns=columns, table_regions=[(40,540,730,55)])
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/home/jose/venv/camelot/local/lib/python2.7/site-packages/camelot/io.py", line 106, in read_pdf
    layout_kwargs=layout_kwargs, **kwargs)
  File "/home/jose/venv/camelot/local/lib/python2.7/site-packages/camelot/handlers.py", line 162, in parse
    layout_kwargs=layout_kwargs)
  File "/home/jose/venv/camelot/local/lib/python2.7/site-packages/camelot/parsers/stream.py", line 425, in extract_tables
    cols, rows = self._generate_columns_and_rows(table_idx, tk)
  File "/home/jose/venv/camelot/local/lib/python2.7/site-packages/camelot/parsers/stream.py", line 321, in _generate_columns_and_rows
    if self.columns is not None and self.columns[table_idx] != "":
IndexError: list index out of range

Here is my OS/Python information:

>>> import platform; print(platform.platform())
Linux-4.15.0-70-generic-x86_64-with-Ubuntu-18.04-bionic
>>> import sys; print('Python', sys.version)
('Python', '2.7.17 (default, Nov  7 2019, 10:07:09) \n[GCC 7.4.0]')
>>> import numpy; print('NumPy', numpy.__version__)
('NumPy', '1.16.6')
>>> import cv2; print('OpenCV', cv2.__version__)
('OpenCV', '4.1.2')
>>> import camelot; print('Camelot', camelot.__version__)
('Camelot', '0.7.2')

[sh-ankar]

If I give the "columns=" parameter, I get the exact same error as above.
If I remove "columns=" , I get no errors, but the output has repeated duplicate subset(s) as part of the TableList. Here is what I mean:

>>>table = camelot.read_pdf("file20.pdf", password="xxxx", pages="20",
              flavor='stream', table_regions=['21,761,575,55'],
              columns=['74,345,412,470,520'])
>>> table
<TableList n=3>
>>> table[0]
<Table shape=(25, 7)>
>>> table[1]
<Table shape=(28, 6)>
>>> **table[2]** 
<Table shape=(66, 7)>      
>>> 

Note here that that complete page is correctly extracted into table[2]. table[0] and table[1] are a subset - hence I am getting duplicate data.

Interestingly in another page in same doc, i get a similar error and here I have 2 tables - but in this case, the table[0] has the full page and table[1] has a subset] ... so its not consitent either

The problem seems be that we are parsing the same rows multiple times. Any parameter I am missing?
The PDF source file I am using is a report from the CAMS mutual fund detailed statement, so I cannot share :(

BTW, this works awesome for the rest of the pages ...