fix the setting of beam_energy and AMITag to avoid saving more than one value...

fix the setting of beam_energy and AMITag to avoid saving more than one value in the TagInfo metadata

Apparently, the fix for https://its.cern.ch/jira/browse/ATLASSIM-4501 has some secondary effects for Athena MP, as reported in https://its.cern.ch/jira/browse/ATLASRECTS-5480.

It turns out that the fix in IOVDbMetaDataTool essentially allowed to write out multiple "different" values for a TagInfo tag/value pair. But normally, one should only write out ONE value for each tag. So it turns out that in RecExCommon_topOptions.py, the beam_energy was writing out values as 6500000 and 6500000.0 which are mathematically the same but created two values for beam_energy and this was causing crashes. Similarly, the AMITag was being written as q221 and q221_q221, etc. also causing crashes. So the changes make sure that only one value is written.

These 'multiple' values seem to just happen in MP which must be related to splitting up the jobs and merging the files together. I admit I didn't exactly understand what is happening in MP, but with these fixes we no longer have more than one value for these tags, which is what we want.

test_q221_mp.sh runs fine until it cannot find art.py which I suppose is normal for interactive running.

Adding in @tsulaia @gemmeren and @wlampl to look at RecEx JO.