msr2data
1 Introduction
msr2data
(originally written by B. M. Wojek) is a program implemented in
C++
. Its purpose is to process multiple
musrfit
msr files with the same parameters and summarize the fitting results either in a TRIUMF
DB or a
column ASCII file. It evolved from a script called
mlog2db
which has been written only for collecting fit results of
LEM data stored in mlog files generated by
musrfit
and putting them together in a file with the
DB format for further analysis. Therefore, any peculiarities, e.g. in supplying options to
msr2data
are owed its original purpose.
msr2data
is
free software and licensed under the
GNU GPL version 2 or any later version (at your option).
Before the usage of msr2data will be explained in greater detail first be aware of the following:
If using msr2data always bear in mind its limitations!
Notes
:
2 Basic Types of Usage
Apart from numerous
optional parameters that might be set or not, in principle there are four different ways of calling
msr2data
. These differ in how the list of runs which should be processed is supplied:
- msr2data <run> <extension> [optional parameters]
- A single run number.
- msr2data <firstRunNo> <lastRunNo> <extension> [optional parameters]
- An interval of run numbers is specified through the first and the last run number. The condition <firstRunNo> < <lastRunNo> is not necessary.
- msr2data \[<run1> <run2> ... <runN> \] <extension> [optional parameters]
- A space separated list of run numbers can be enclosed in [ and ]. Every run that should be processed has to be specified explicitly by its run number.
- msr2data <runListFileName> <extension> [optional parameters]
- An ASCII file containing a list of run numbers and optional external parameters is passed to
msr2data
. For the structure of the ASCII file see below.
All four basic types of calling
msr2data
contain the
mandatory file-name
<extension> passed right after the list of runs. The meaning of this
<extension> should become clear after giving examples for all four cases:
msr2data 8472 _tf_h13
generates the
DB file
out.db
(can be changed by using
the -o option) from
8472_tf_h13.msr
.
msr2data 8472 8474 _tf_h13
generates the
DB file
out.db
(can be changed by using
the -o option) from
8472_tf_h13.msr
,
8473_tf_h13.msr
, and
8474_tf_h13.msr
.
msr2data [8472 8470] _tf_h13
generates the
DB file
out.db
(can be changed by using
the -o option) from
8472_tf_h13.msr
and
8470_tf_h13.msr
.
msr2data run.list _tf_h13
generates the
DB file
out.db
(can be changed by using
the -o option) from all runs listed in the ASCII file
run.list
in the working directory. In this file it is also possible to include external parameters which should be put in the resulting
DB file. The structure of the
run.list
is the following:
RUN VAR1 VAR2 VAR3 ...
8460 200 27.1 46.2 ...
8472 205 27.1 46.3 ...
8453 210 27.2 45.9 ...
· · · ·
· · · ·
· · · ·
The first not commented and not empty line determines the parameter names and labels and has to be present!
It is allowed to add comments (with a preceding
#) or empty lines to the run-list file.
The following should be mentioned together with the above examples:
- The output files in the examples above are only newly created if they did not exist before invoking msr2data. If the files were already present the msr-file data would be appended!
- If the files have been newly created, also the DB file header is written. If the files were present before, only the data blocks are appended. The output of the header can either be forced or completely suppressed with the header and noheader options as shall be seen later.
- If the
musrfit
output files do not have an <extension> as specified above like 8472.msr
one has to call msr2data
like in the following example:
msr2data 8472 8460 ""
3 Optional Parameters
As mentioned already above there are some optional parameters which change the behavior of
msr2data
and can be passed in any order. Here is a complete list:
- data
- The output file format is changed to a simple column ASCII file (default output file name:
out.dat
).
- new
- An existing output file is deleted before new information is written to it.
- header
- Force the output of the file headereven if the output file was present before.
- noheader
- The output of the file header is suppressedalso if the output file is newly created.
If either both or none of the header options are given, msr2data
writes the file header only to new files and it solely appends the data blocks to an existing output file assuming that the header is present already.
- nosummary
- There will be no attempt to read additional information like the temperature or the applied magnetic field from the data files even if these information were present there.
- -o<outputFileName>, -o <outputFileName>
- The processed data will be written to the file <outputFileName> instead of the default
out.db
or out.dat
. If <outputFileName> is equal to none (case-insensitive) the parameter data are not appended to any output file.
- fit
- Additionally to the final data collection
msr2data
will invoke musrfit to fit the specified runs. All msr files are assumed to be present, none is newly generated!
- fit-<template>[!]
- Additionally to the final data collection
msr2data
will generate msr files for the runs specified in the list of runs and invoke musrfit
for performing fits of the data. As template for the first run the file <template><extension>.msr
(or if not available: <template><extension>.mlog
) is used; the subsequent input files will be created using the msr output of the last processed runs ("chain fit"). However, if for all runs only the given template should be used one has to append an exclamation mark (!) to the <template>.
- msr-<template>
- The same as fit-<template>[!], without calling
musrfit
and the final data collection, i.e. only the msr files for the given runs are generated.
- -k
- If specified together with the fit-<template> option, the --keep-mn2-output option is passed to
musrfit
. In the case no fits should be done, this option is ignored.
- -t
- In case this option is given additionally to the fit-<template> option,
musrfit
is called with the --title-from-data-file option. If no fitting is done, this option is ignored.
In order to illustrate the usage of these parameters a few examples with explanations are given below:
msr2data 8400 8460 _tf_h13 -oABC.db fit-8472
Using
8472_tf_h13.msr
as first template,
msr2data
generates subsequent msr input files
8400_tf_h13.msr
through
8460_tf_h13.msr
, calls
musrfit
to perform a fit of these files and collects the results of the fits together with the
DB header in the new file
ABC.db
. Additionally, some information about external parameters like the temperature will be passed to
ABC.db
if it is present in the data files.
msr2data [8500 8505] _zf fit-8472 noheader nosummary -o DEF.db
Using
8472_zf.msr
as first template,
msr2data
generates subsequent msr input files
8500_zf.msr
and
8505_zf.msr
, calls
musrfit
to perform a fit of these files and collects the results of the fits in the file
DEF.db
without writing the
DB file header or attempting to read additional information from the data files.
msr2data 8595 8585 "" noheader fit-8472! -oGHI.dat data nosummary -k
Using
8472.msr
as template for
all runs,
msr2data
generates the msr input files
8595.msr
through
8585.msr
, calls
musrfit
with the option
--keep-mn2-ouput to perform a fit of these files and collects the results of the fits in the column-structured ASCII file
GHI.dat
without writing any file header or attempting to read additional information from the data files.
msr2data 8472 8475 "" fit -o none
Take the
given msr files
8472.msr
through
8475.msr
and call
musrfit
without finally summarizing the results.
msr2data 8472 8475 _tf_h13 msr-8471!
Using
8471_tf_h13.msr
as template for
all runs,
msr2data
generates the msr input files
8472_tf_h13.msr
through
8475_tf_h13.msr
.
No fitting will be performed and no DB or ASCII output will be generated!
4 The Global Mode
Apart from all the options described
above there is another program option:
global. This option changes the general behavior of
msr2data
in that way that instead of processing one msr file for each run it combines all specified runs in one single msr file with the possibility to define common parameters for all runs as well as run-specific parameters. When writing the obtained parameters to a DB file or a column-structured ASCII file that single msr file is read and the parameters valid for each run are extracted. The
global option can be used in conjunction with any of the described
invocations of
msr2data
and together with all options stated
above.
4.1 File Generation
The general idea of this mode is to generate a global msr file on the basis of a working single-run msr file. For this purpose a single-run template containing information about common and run-specific parameters should be created. These parameters are identified through their parameter names:
- run-specific parameters
- these parameters are tagged with the current run number in the format %0Xu, i.e. X digits with leading zeros, at the end of the parameter name, e.g. for a 4-digit-formatted run number alpha0123 if the run number was 123 or for a 8-digit-formatted run number alpha00123456 if the run number was 123456. X has to be at least 4.
- common parameters
- all parameters that are not run specific
The FITPARAMETER block of an exemplary template file
8472_example.msr
could therefore look like:
FITPARAMETER
# No Name Value Step Pos_Error Boundaries
1 Phase 35.8359 -3.94496 3.93749
2 Asy8472 0.04501 -0.00208 0.00211 0 0.33
3 Field 143.212 -0.27960 0.27885 100 200
4 Rate8472 0.14245 -0.02501 0.02279 0 1
Here the parameters
2 and
4 would be treated as run-specific whereas the parameters
1 and
3 would be common to the original and all newly added runs.
Normally, within the template file there should
not appear explicitly any run-specific parameters in the THEORY and FUNCTIONS blocks. If however, those parameters are met,
msr2data
will try to substitute them by mapped parameters and add them accordingly to the map contained in each RUN block.
When
msr2data
is called to generate a global msr file, e.g.
msr2data 8471 8470 _example msr-8472 global
a new msr file
8471+global_example.msr
is created. As can be seen in the example, the name of the global msr file always starts with the first specified run number followed by the
+global identifier and the template
<extension>.
The example's global FITPARAMETER block would be:
FITPARAMETER
# No Name Value Step Pos_Error Boundaries
# Common parameters for all runs
1 Phase 35.8359 -3.94496 3.93749
2 Field 143.212 -0.27960 0.27885 100 200
# Specific parameters for run 8471
3 Asy8471 0.04501 -0.00208 0.00211 0 0.33
4 Rate8471 0.14245 -0.02501 0.02279 0 1
# Specific parameters for run 8470
5 Asy8470 0.04501 -0.00208 0.00211 0 0.33
6 Rate8470 0.14245 -0.02501 0.02279 0 1
This shows that the fit parameters are reorganized in a way that the common parameters appear at the beginning of the parameter list and they are followed by copies of the parameters specific to each run (in the specified order!). Additionally, for each specified run new RUN blocks are created—for each run as many as found for the template run.
During this reorganization all the affected parameter occurrences are changed accordingly!
Please be aware of the fact that comments in the template msr file are not propagated to the newly generated global msr file!
4.2 Parameter Extraction
After fitting some model to the specified data the fit parameters can be extracted from the global msr file to a
DB or column-structured ASCII file—as usual this includes also parameters stored in the run data files or externally specified parameters given in a
run-list file.
In order to reach this goal the global msr file has to obey certain rules:
- The order of the parameters has to match the one described above, meaning the common parameters are listed first followed by the same number of parameters specific to each run tagged by the according run numbers at the end of the parameter names and having the same order as the specified list of runs.
- The RUN blocks have to be ordered according to the list of runs to be processed.
Following these rules—which is achieved most easily by generating the global msr file using
msr2data
as shown
above—the parameters can be extracted e.g. like
msr2data 8471 8470 _example global data -o globalFit.dat
This will read in the file
8471+global_example.msr
, extract for each run all relevant parameters from the msr file as well as from the according data files (if available) and append all of them in columns to the ASCII file
globalFit.dat
.
4.3 The Extended Global Mode
If a new global input file is generated, it is also possible to do an automatic pre-analysis for each single run using the specified template first; afterwards the run-specific parameters of these single-run msr files are collected in the global msr file. In special cases this might be useful to obtain a better set of starting values for the parameters, however, in most cases it will not replace the "manual review" of the generated global input file. The option is activated by choosing the keyword
global+. For example
msr2data 8471 8470 _example global+ msr-8472
Here,
8472_example.msr
is first used as template to generate the file
8471-OneRunFit_example.msr
, then
musrfit
is called for it, the result is used to generate
8470-OneRunFit_example.msr
and
musrfit
is called for that file. Finally, the global fit file
8471+global_example.msr
is produced—including the fit results of the OneRunFit files for the run-specific parameters.
By appending an exclamation mark
! to the
global+ option, the given template will be used for every new file generation (similar to the
fit option explained
before). The
+[!] extension will be ignored, if no new global input file is generated.
The single run msr files are not deleted at the moment. The information contained in them might be useful for some people. Of course the data can also be collected by
msr2data
. E.g. in order to produce a
DB file
OneRunFits.db
one could call
msr2data 8471 8470 -OneRunFit_example -o OneRunFits.db
Please be aware that the program in this mode always generates new single-run msr files and always calls musrfit
for them. In case there are already single-run fits present, these cannot be used in conjunction with this option. The program on purpose behaves in this way in order to ensure the file integrity and correct parameter order within these files.
5 Known Limitations
- The indexing run number of the msr file has to be at the begin of every filename.
- Within the data file name the RUN# has the format %0Xu, i.e. X digits with leading zeros, and has to be the rightmost number given in this format in the file name. X has to be at least 4. The highest treatable run number is 232-1 = 4294967295.
- In order to keep
msr2data
working properly the msr files should only contain one STATISTIC block at the end of the file and one FITPARAMETER block right after the TITLE—musrfit
itself allows to have more creative msr files...
- The msr-file generation from a template takes only care of runs given on the first line of a RUN block. ADDRUN statements are simply copied! Since this is most probably not what one likes to do, it is suggested not to use the fit-<template> and msr-<template> options if ADDRUN statements were present in the template file.
-
msr2data
will write only up to two successive empty lines in newly generated msr files. In case more subsequent empty lines are encountered in a template file, these are not copied! Actually, this measure is not a limitation but has been introduced to keep the msr files in a reasonable shape.
6 The Graphical User Interface Provided by musrgui
/ musredit
musrgui/musredit
, editors designed especially for the manipulation of
musrfit
msr files and graphical front ends to
musrfit
, offer an almost self-explanatory graphical user interface to
msr2data
depicted below:
- 1–3
- Choose one of the ways to specify your list of runs.
- 4
- Give the file extension here, e.g. _zf for files like
8472_zf.msr
. If the files do not have an extension this field stays empty. musrgui
/ musredit
take care of passing the "" to msr2data
as mentioned above.
- 5
- Activates the fit-<template> option if <template> is entered. In case the option "Chain Fit" is not set the given template will be used for the input-file generation for all runs to be fitted—otherwise the output of the first fit serves as template for the second and so on. The template field stays empty if no fits should be performed!
- 6
- Activates the -o<outputFileName> option if <outputFileName> is entered. If nothing is entered the default output file
out.db
or out.dat
is used.
- 7
- If not set, the noheader option is passed to
msr2data
.
- 8
- If set, the nosummary option is passed to
msr2data
.
- 9
- If set, the -k option is passed to
msr2data
.
- 10
- If set, the data option is passed to
msr2data
.
- 11
- If set, the output file is deleted before
msr2data
is called. Otherwise the data is appended to the file if it existed before msr2data
has been called.
- 12
- If set, after processing by
msr2data
all specified msr files are opened in tabs in musrgui
/ musredit
.
- 13
- If set, the -t option is passed to
msr2data
.
- 14
- If set,
msr2data
is called with the option msr-<template> (instead of fit-<template>) and hence is only producing the msr input files without any fitting.
- 15
- If set,
msr2data
is called with the option fit, i.e. no template will be used to generate new msr input files; only existing files are fed to musrfit
.
- 16
- If set,
msr2data
is operating in the global mode.
- 17
- If set,
msr2data
is operating in the extended global mode.
7 Bugtracking
For reporting bugs or requesting new features and improvements please use the
PSI Tracker or send an e-mail to A. Suter.
--
BMW &
AS