pycbc_plot_segments fails if the segment table is empty

[duncan-brown]

The job at

http://sugar-dev3.phy.syr.edu/pegasus/u/steven.reyes/r/4/w/1/j/21/ji/14

is failing with

OverflowError: cannot convert float infinity to integer

http://sugar-dev3.phy.syr.edu/pegasus/u/steven.reyes/r/4/w/1/j/21/ji/14/stdout

[duncan-brown]

Looks like this is due to reading a veto file with an empty segment table

        <Table Name="segment:table">
                <Column Type="ilwd:char" Name="segment:process_id"/>
                <Column Type="ilwd:char" Name="segment:segment_def_id"/>
                <Column Type="ilwd:char" Name="segment:segment_id"/>
                <Column Type="int_4s" Name="segment:start_time"/>
                <Column Type="int_4s" Name="segment:start_time_ns"/>
                <Column Type="int_4s" Name="segment:end_time"/>
                <Column Type="int_4s" Name="segment:end_time_ns"/>
                <Stream Delimiter="," Type="Local" Name="segment:table">
                </Stream>
        </Table>

[ahnitz]

Duncan,

I think this is due to all vetoes being empty, which is strange but possible, and should not fail. We had caught this issue here and it should be already fixed. Can you make sure your run includes this commit?

https://github.com/ligo-cbc/pycbc/commit/536a8a463a8be16828a10f72c3043da7da40858f

[duncan-brown]

Thanks, Alex, I think Steve had an older release.

@stevereyes01 please can you rebase to master and see if this works now?

[stevereyes01]

Will rebase to master. Thanks for the help Duncan and Alex. On Sep 21, 2015 8:20 AM, "Duncan Brown" notifications@github.com wrote:

Thanks, Alex, I think Steve had an older release.

@stevereyes01 https://github.com/stevereyes01 please can you rebase to master and see if this works now?

— Reply to this email directly or view it on GitHub https://github.com/ligo-cbc/pycbc/issues/318#issuecomment-141955023.

[duncan-brown]

Yes, fixed thanks: http://sugar-dev3.phy.syr.edu/pegasus/u/dbrown/r/55/w?wf_uuid=0df2506b-6a09-47b9-bdff-957dda965235. @stevereyes01 didn't have that fix.