Improve support for event formats

[kilian]

We agreed that this is the remaining major issue for an alpha pre-release of 2.2.0.

The most important formats, in order of priority: [ 1. Raw format (done, but still subject to incompatible changes and improvements). Done, except for separate ticket #627] [ 2. LHEF . Done, except for separate ticket #570.] [ 3. HepMC (see #506), done except for ticket #647.]

4. Stdhep and various ASCII formats, if still in use. Done except for

reading and reconstructing. [ 5. LCIO (see #439) and other new formats. -> separate ticket, probably done by WK's BSc]

Minimum is 1. and 2.

[kilian]

r4613: LHEF implemented for output. This should suffice for a first workable version.

[kilian]

In r4652, I added support for (potentially varying) alpha_s values in the LHEF output.

The corresponding test lhef_2 also digs up a bug: the color connections of initial particles are missing. Should be tackled first, then HepMC output. LHEF/HepMC input can wait.

[kilian]

The color connections have been fixed in r4654. The next issue is HepMC (for output).

[kilian]

As of r4672, there are both LHEF and HepMC available for output.

Minor issues remaining:

• The functional tests for HepMC do not check the file contents. I should filter out the HepMC version number, to get a portable check.
• HepMC does not respect keep_beams (that is probably ok, to be confirmed)
• The checks rely on the HepMC event format being fixed and include all digits from the numerics. They may fail for quadruple precision or for other HepMC versions. If this turns out to be non-portable, the check would have to rely on reading the file again and comparing the event data in WHIZARD's own format.

The updated priority list is as follows:

1. Address the above HepMC issues.
2. Implement further formats as supported by 2.1
3. Add input form for LHEF and HepMC
4. New event formats (LCIO)

[Juergen Reuter]

LHEF references:

​http://phystev.in2p3.fr/wiki/2013:groups:tools:eventformats

​http://arxiv.org/abs/arXiv:1003.1643

[Juergen Reuter]

Stalled. Too many features have been broken away from 2.1.1.

[Juergen Reuter]

It looks like Hepevt, Athena, Mokka, long and short ASCII formats are successfully reimplemented. LHA still is not correct, we have new ASCII default and debug formats. File endings have to be reenabled, and the two different stdhep formats.

[Juergen Reuter]

File endings done in r4803. Next are checks about LHA and the verbose versions of LHA and HEPEVT, then STDHEP and STDHEP_UP.

[Juergen Reuter]

LHA is working as intended. Next will be the two verbose output formats for HEPEUP and HEPEVT data blocks.

[kilian]

Some reshuffling in r4804 should establish the roles of the various sqme and weight numbers floating around. The handling of event is then well defined (in code, at least), and I'll leave the formats to JR for now, d'accord?

[Juergen Reuter]

I wanted to reassign that ticket to myself, but didn't dare to :P Now, verbose HEPEVT and LHA are done. Next is STDHEP with HEPEVT and HEPEUP based setup. However, I don't really know how to write a unit test for this.

[Juergen Reuter]

Finally, STDHEP is implemented again in both versions, including tests. Unit tests are very simplistic as there is no XDR reader for those files to check them.

Open points are: 1) XDR reader for proper STDHEP unit tests (shall we do this at all?) 2) Reading events back in from ASCII and STDHEP formats.

Both are enhancements/tasks left for 2.2.1.

[Juergen Reuter]

I'll try my luck with this one now.

[Juergen Reuter]

This has been split in separate open tickets, #647, #439, #662)