Sikuli 1.0.1 - find() identifies wrong text

sorry, currently know. we have to live with it.

do you have a chance, to run the sheet with a larger font?

But there is the package xlrd, that you can use to access an XLS directly from a Sikuli script.

I ran it with zoom set to 150%, but the issue did not go away - still it selects the cell with "11/26/18" most of the time.

I was running the excel test to isolate fonts etc. used in my application. The real test would have to work with the application, so unfortunately xlrd package wouldn't help me in this instance.

Ok, then there is no chance currently, to get that working with Sikuli for now in a robust way.

Another way would be to turn it around and try to use the OCR feature (Region.text()), to find out, what Sikuli reads in specific cell.

For the planned app testing this approach could be used if it works better.

Tried to see what OCR gives in this case, by 'iterating' over the rows using Region.below(x).

cell_region = region.find("<Image of the column header>.png").below(20)

#Loop over the rows, read the text in the cell and print it out
for x in xrange(0, 25):
    print cell_region.text()
    cell_region = cell_region.below(20)
    click(cell_region)

The output is given below, which has a few interesting things:
  - For some reason, the beginning part of the string comes at the end (and seemingly with a new line prefixed to it)
  - Some existing spaces are getting stripped while some new spaces are getting inserted

The above 2 points might make even the comparison with Region.text() work in an inconsistent fashion. Ofcourse, Region.text() can be replaced with 'Copy and get the data from clipboard', but that might not work in all cases (especially with Read-only UIs).

Output of the script snippet with the input data specified in the original question:

4.15006/ 10/ 14
XYZ

4.62509/08/14
XYZ

5.37501/30/15
XYZ

5.37501/30/15
XYZ

5.12504/10/15
XYZ

5.00009/15/15
XYZ

3.37511/24/15
XYZ

6.25001/15/16
XYZ

5.75003/15/16
XYZ

2.87503/ 16/ 16
XYZ

4.75006/ 14/ 16
XYZ

5.62502/27/17
XYZ

5.62502/27/17
XYZ

1.62503/20/17
XYZ

6.50007/31/17
XYZ

1.25009/26/17
XYZ

5.37512/05/ 17
XYZ

5.37512/05/ 17
XYZ

1.50002/ 19/ 18
XYZ

5.00006/04/18
XYZ

4.62507/ 15/ 18
XYZ

8.12511/26/18
XYZ

8.12511/26/18
XYZ

5.45006/ 10/ 19
XYZ

4.37503/ 16/ 21
XYZ

If you see that the recognition error is consistent you can attempt to perform some "post processing".

So for example if you consistently getting an unneeded space or new line, you can check that and make some corrections.