hgrep.man

.TH hgrep 1
.
.
.SH NAME

hgrep \- extract values from the HDF5 QMC output file
.
.
.SH SYNOPSIS

.B hgrep
.RI "(" filename " | " \fBlatest\fR ")" " " \fBlist\fR

.B hgrep
.RI "[OPTIONS] (" filename " | " \fBlatest\fR ") " "quantity" " [[" index \fB=\fR] values " ...]"
.
.
.SH DESCRIPTION

.B hgrep 
is a python script which extracts data sets from the binary result file of a DMFT
calculation.  These files are essentially a collection of quantities together 
with some metadata describing the run and the quantities in general.

To retrieve all quantities stored in an output file, use the first form above.  
To print a specific quantity, use the second form above.  You can select only 
parts of the data by specifying further values (see \fBSelecting data\fR).

You can use \fBlatest\fR instead of \fIfilename\fR to use the .hdf5 file in
the current directory that was last modified.  You can join multiple input files
by a colon.
.
.SS Output

By default, hgrep prints the selected quantity as a set of column-aligned 
tuples. Each tuple corresponds to one datapoint and contains the respective
indices (iteration, band, tau bin, etc.), the result value and, where 
available, an error estimate. It will also print header lines starting with `#'
that describe the data being plotted. The output is compatible with most
plotting tools and CSV import filters.
.
.SS Selecting data

You can limit the output to only parts of the quantity by specifying ranges for
the different indices. The general syntax is:
.IP
.RI [ index \fB=\fR] selection [\fB,\fR selection " ...]"
.PP
If \fIindex\fR is omitted, then the selections refer to the indices in the order
that they are printed. You may not mix named and unnamed indices.

Each \fIselection\fR (separated by \fB,\fR but no whitespace) may either be a
a single \fIvalue\fR or a range from \fIstart\fR to \fIstop\fR, both inclusive:
.IP
.IR selection " ::= " value " | [" start "]\fB:\fR[" stop "][\fB:\fR" step "]"
.PP
\fIvalue\fR, \fIstart\fR and \fIstop\fR may be positive or negative: if positive
\fIn\fR, they denote the \fIn\fR-th element of the index, starting with 1. If 
negative \fI\-n\fR, they denote the \fIn\fR-th last element (\-1 being the last
element). If \fIstart\fR or \fIstop\fR are omitted, they default to the 1 and 
\-1, respectively. \fIstep\fR is an optional positive stepsize and defaults to 1.
For example, \fB:\fR is shorthand for \fB1:-1:1\fR and selects all elements (see
\fBExamples\fR).

Selecting iterations works a bit differenty: there are two special iterations,
\fBstart\fR and \fBfinish\fR, which store data for the beginning and the end of
the run. If you omit the iteration selection, or use \fB:\fR for it, then these
iterations are included as well. If you specify only parts of the iterations,
then they are not included and you will have to select the manually, e.g., you
can use: \fBiter=-3:,finish\fR.
.
.
.SH OPTIONS
.TP
.BR \-d ", " \-\-debug
Prints lots of debug output.
.TP
.BR \-t ", " \-\-tabular
Print the quantity in a pivot table like fashion instead of a list of tuples,
but also separate them by tabs. Prints separate tables all fields (value, error,
real and imaginary part). These tables are separated by a pair of newlines,
which makes them available to the \fBindex\fR keyword in \fBgnuplot\fR.

.PI 
Then, the quantity's indices are distributed to rows and columns based on the 
nature of the quantity: for instance, typical plotting indices (like tau) are
used as row indices to aid plotting.

.PI
Unless \fB--skip-header\fR is specified, each table begins with header lines
(prefixed by `#') describing how the different columns map to combinations of the
column index values. Then, for each row in the table, first the row index values
are printed and then the field value for each column.
.
.
.SH EXAMPLES
.TP
hgrep foo.hdf5
Print available quantities from foo.hdf5.
.TP
hgrep foo.hdf5 mu
Print the chemical potential (mu) for all iterations in foo.hdf5.
.TP
hgrep foo.hdf5 siw 1,\-1 1 3: 2 0:20
Print the self-energy in Matsubara frequencies (siw) for the first and last 
iteration (1,\-1), first atom (1), all but the first two bands (3:) and 
second spin (2) for the frequency range 0 to 20 of foo.hdf5.
.TP
hgrep --tabular foo.hdf5 gtau iter=1,-3:-1 spin=1
Print the Green's function in tau (gtau) in a tabular fashion (--tabular)
for the first and the last three iterations (iter=1,-3:-1) and the first
spin (spin=1)
.TP
hgrep foo.hdf5 time finish
Print the total runtime
.
.
.SH AUTHOR

Markus Wallerberger <markus.wallerberger@ifp.tuwien.ac.at>