HPS-SOFTWARE Archives

Software for the Heavy Photon Search Experiment

HPS-SOFTWARE@LISTSERV.SLAC.STANFORD.EDU

Options: Use Forum View

Use Monospaced Font
Show Text Part by Default
Show All Mail Headers

Message: [<< First] [< Prev] [Next >] [Last >>]
Topic: [<< First] [< Prev] [Next >] [Last >>]
Author: [<< First] [< Prev] [Next >] [Last >>]

Print Reply
Subject:
From:
"McCormick, Jeremy I." <[log in to unmask]>
Reply To:
Software for the Heavy Photon Search Experiment <[log in to unmask]>
Date:
Mon, 19 Oct 2015 20:49:56 +0000
Content-Type:
text/plain
Parts/Attachments:
text/plain (34 lines)
Hi,

I've noticed that we have a lot of data files required by various parts of hps-java now, some of which are in our SVN trunk with the code and others of which are not, and that the way we access them is inconsistent and often a bit clunky.

These data files include:

1) Many detector data directories and files

2) Magnetic field maps in text files

3) Test data including LCIO and EVIO

The current way of accessing the magnetic field seems pretty awkward to me, as it requires manually downloading a file from the web to a specific local directory.  I would propose that instead we import the magnetic field map(s) into our SVN in a node that is separate from hps-java e.g.

hps/data/magfield-data

where it could be packaged as a Maven dependency.  Then the text data could be imported as a class resource at runtime.

We have many "loose" test files in an lcsim.org directory that are downloaded mostly during integration-tests.  I am not crazy about this setup, especially because we then have no real versioning system for our test data.  I would propose instead that the test data be checked into SVN (under hps/data/test-data) and bundled into a zip using the Maven assembly program so that it can be easily referenced and accessed by other projects.  Before we do this, we should probably cut down on the number of LCIO and EVIO files that we are using in our testing so that they are minimal. 

Finally, the detector data I believe should not be in the hps/java/trunk SVN node but would be better relocated someplace else like hps/data/detector-data as it is consistents of many data files which are not code, some of which are pretty large (e.g. the LCDD files) and take a long time to checkout.  This would likely be the biggest proposed change and it could have some ramifications in terms of build/release procedures, so I will not change this until there's a thorough discussion about it (essentially though the project would be unchanged and could just be moved using 'svn cp').

The other two changes I believe can be done without disruption, so I would eventually like to do this (perhaps after the collaboration meeting?).

Let me know what you think about these issues and if they are worth spending time on improving.

--Jeremy

########################################################################
Use REPLY-ALL to reply to list

To unsubscribe from the HPS-SOFTWARE list, click the following link:
https://listserv.slac.stanford.edu/cgi-bin/wa?SUBED1=HPS-SOFTWARE&A=1

ATOM RSS1 RSS2