NEMO's tabplot (and also tabhist) are very useful for quick analysis of tabular
data without having to open complex programs and a complex user interface.
For a given table (or CSV file) it's usually as simple as
tabplot file.tab
which would plot columns 1 and 2. However, in NEMO's C program there is no further
interaction possible. In python's matplotlib there is an additional pan and zoom
option. We would like to make a simple version of tabplot and tabhist that will use
python. There are the following design issues:
1) The command line user interface?
Clone NEMO's key=value
Use the unix style --key value, e.g. using parseargs
Use docopt's self-describing, it also allows -key=val
something else
2) How to read the table. We have simple SSV (space separated values), the TSV, CSV
and perhaps a few others.
3) In matplotlib it's a bit tricky to run interactive (this should be the default) and
in a pipeline. In NEMO this is usually done via the yapp_pgplot style, e.g.
yapp=2/xs would open XS window 2, but yapp=plot2.png/png would write it to a PNG file.
4) Only a few basic options should be implemented (at first). The current programs have a rather large set of options that we probably don't need. However, the layout= keyword could be an interesting one that needs some thought.
NEMO's tabplot (and also tabhist) are very useful for quick analysis of tabular data without having to open complex programs and a complex user interface. For a given table (or CSV file) it's usually as simple as
which would plot columns 1 and 2. However, in NEMO's C program there is no further interaction possible. In python's matplotlib there is an additional pan and zoom option. We would like to make a simple version of tabplot and tabhist that will use python. There are the following design issues:
1) The command line user interface?
2) How to read the table. We have simple SSV (space separated values), the TSV, CSV and perhaps a few others.
3) In matplotlib it's a bit tricky to run interactive (this should be the default) and in a pipeline. In NEMO this is usually done via the yapp_pgplot style, e.g. yapp=2/xs would open XS window 2, but yapp=plot2.png/png would write it to a PNG file.
4) Only a few basic options should be implemented (at first). The current programs have a rather large set of options that we probably don't need. However, the layout= keyword could be an interesting one that needs some thought.