Wire Cell Toolkit Configuration Data
This area holds "configuration data" for WCT. Confusingly, this is distinct from but similar to WCT software configuration (see wire-cell-cfg package) as much of if is in the same JSON format as configuration. This is also not "data" in the sense of what WCT simulation, signal processing or reconstruction produces.
Rather, "configuration data" is large, "bulk" input information which WCT needs to do its job but which is too voluminous for humans to type out. Rather, these configuration data files are generated with some external mechanism and then converted to ready-to-consume JSON by code in wire-cell-python).
2 Using this data
To use this data simply clone this repository into a directory and
place that directory in your
WIRECELL_PATH. The history of this
repository is not particularly useful so to save time and disk space a
shallow clone is enough:
$ git clone --depth 1 https://github.com/WireCell/wire-cell-data.git $ export WIRECELL_PATH=$(pwd)/wire-cell-data
3 Tour of data files
All WCT configuration data files are in JSON format and optionally compressed with BZip2. The files are described here broken into various categories of purpose. Be wary that the descriptions that follow may not be comprehensive nor up to date. Also, only WCT configuration data files are committed to the repository. The "upstream" files are named below and may be available here. There are also some diagnostic plots related to these files here.
Wires (wire segments) in WCT are described by their physical endpoints and their logical locations in terms of various numbers and indices. A "WCT wire file" provides an exhaustive list of this info for each wire. Note, WCT can generate wire files in a parameterized way given basic information about an anode plane. See
$ wirecell-util make-wires --help
However, some wires are taken from other programs and then converted to WCT wire files. See
$ wirecell-util convert-oneside-wires --help
ChannelWireGeometry_v2.txta dump of MB wires from larsoft in so called "celltree wire format"
- ./microboone-celltree-wires-v2.json.bz2 the conversion of the above
- ./microboone-celltree-wires-v2.1.json.bz2 the conversion of the above, fixing wire index in plane
See this plot for graphical representation of these wires. (Note, a transform may be applied to their locations in WCT, so don't take the positions shown here as gospel.)
3.2 Field Response
The field responses are generally calculated by Garfield (so called "truth responses" may be from some other source). The data produced by Garfield are in the form of a family of files named by their impact position and plane letter. Each file spans that impact position relative to many wires. To process this data into "WCT field response files" see:
$ wirecell-sigproc convert-garfield --help
The required "origin" and "speed" are not recorded in the files
directly and must be provided on the command line of this converter.
Note the converter takes as input a
.tar.gz of the Garfield
<impact>_<letter>.dat files (and no other files).
dune_4.71.tar.gzDUNE APA Garfield data
- ./garfield-1d-3planes-21wires-6impacts-dune-v1.json.bz2 WCT field response file from above.
MB has many sets of field response files. The nominal set is called
ub_10 (uboone wire model with fields starting at 10cm from the
ub_10.tar.gznominal MB Garfield data
- ./ub-10.json.bz2 WCT field response conversion
Alternatives in two dimensions have been created in order to understand the MB shorted wires and a Garfield normalization issue. The shorted wire alternatives have labels
- the U and V wires are shorted
- the V and Y wires are shorted
The normalization issue is that early Garfield runs seem to produce collection responses which when integrated gives a total charge of about 2.04 electrons instead of the expected 1.0. Initially the responses were normalized such that the average collection wire integral was 1.0. This technique utterly fails for shorted wires where "collection" becomes ill defined. The variants are labeled as:
- normalize by the average collection (W wire) integral method
- take Garfield normalization as-is
- simply scale by 0.5
These are the files:
The WCT simulation can produce proper intrinsic noise waveforms based on a measured noise amplitude distribution expressed in frequency domain. These are provided as simple text files which are then converted to "WCT noise files". For info on this conversion see
$ wirecell-sigproc convert-noise-spectra --help
As the procedures are improved the
vN version label is increased.
Note, strictly, these types of files should not be included here and may be removed in the future. The current WCT drift and detector response simulation relies on an external interaction/tracking simulation (ie, Geant4 based) to provide initial distribution of energy depositions or of ionization electrons. A couple samples of these are provided.