Command-line Client Guide
The main Soaplab2 command-line client is a typical testing tool. It
covers the full functionality of the Soaplab2 features but it is,
after all, only a command-line program. Which means that its
user-friendliness is limited, but it can be very well used for testing
and for batch and automated Soaplab2 services calls.
CmdLineClient
=============
Usage:
build/run/run-cmdline-client -h[elp]
build/run/run-cmdline-client <find-arguments> <list-options>
build/run/run-cmdline-client <find-arguments> [options] [inputs] [results]
where:
-h[elp] ... print this help
-q ... quiet mode (useful when you want to redirect
a result into a file)
-stack ... in case of any error it prints the stack trace
(without this option, it prints it only when a
serious runtime error occures)
where <find-arguments> defines what service to use, where to find it
and what protocol to use to access it (some of these parameters may
also be specified in the Soaplab client configuration property file -
by default named soaplab.client.properties and put on the CLASSPATH):
-protocol jaxws ... service will be accessed as a Web Service,
using JAX-WS toolkit (default protocol)
-protocol local ... service will be accessed as local class
(no remote server or any Web Service used)
-protocol axis1 ... service will be accessed as a Web Service,
using old Axis 1 toolkit (such protocol may
not be supported by the given Soaplab server)
-host <host> ... host name (default: localhost)
-port <port> ... port number (defaullt: 8080)
-ctx <context> ... context part of the URL (default: soaplab2)
For Analysis service requests:
-name <service> ... a name of a service to be called
(mostly mandatory, unless -e given or
only <list-options> used)
-e <url> ... a full service URL (an endpoint);
if given it has precedence over -host, -port,
-ctx, and -name;
default is created from other parameters as:
http://<host>:<port>/<context>/services/<service>
For List service requests:
-lname <service> ... a name of a list service to be called
(default 'AnalysisFactory' for axis1 protocol,
otherwise 'list')
-el <url> ... a full List service URL (an endpoint);
if given it has precedence over -host, -port, -ctx,
and -lname;
default is created from other parameters as:
http://<host>:<port>/<context>/services/<list-service>
-sl <filename> ... where to save produced lists;
by default the lists are shown on standard
output (which is usually a screen)
where <list-options> defines list request:
-l ... show all analyses and categories
-la ... show all analyses
-lad ... show all analyses with short descriptions
-lc ... show all categories
-category <category-name>
show analyses of the given category
where the main <options> are one of the following (the behaviour may
depend on whether also parameter '-job' is given):
-j ... create a job, but not start it
-x ... create a job and start it, but do not wait for it
-x -job <job-ID> ... start given job, do not wait for it
(note that the options -j and -x have none or only
limited use for '-protocol local')
-w ... create a job, start it, wait for its completion
-w -job <job-ID> ... wait for completion of the given job
-k -job <job-ID> ... kill the given job
-jp <file> ... store a job ID into given file
(default: print it on standard error)
where other <options> are (they are associated either with a job
created in the same invocation, or with a job defined by a '-job' parameter):
-s ... print job status
-c ... print job time characteristics
-le ... print the last event associated with a job
(TBD: more options related to notification)
-clean ... destroy given job and all its resources
(which includes removing also results)
where options showing metadata (not associated with any particular job) are:
-d ... show analysis full description
(it prints an XML file with all details)
-t ... show analysis type and few other properties
-i ... show names of analysis inputs
-ii ... show names and attributes of analysis inputs
-o ... show names of possible results
-oo ... show names and types of possible results
where <inputs> is a list of input names and its values:
available names can be found using option -i,
<input-name> <input-value>
for boolean inputs the <input-value> should be
'true' or 'false' (but see also below -<input-name>);
some inputs allow multiple (repetitive) values - they can be
entered using a bar character as separator '|', for example:
name 'Kim|Blanka'
(note that the quote characters must be there to avoid the UNIX
shell to misinterpret the bar character);
you need to quote the value if the bar character should be part
of the value, for example (again, double quoting is because of
the UNIX shell):
name "'this|is|not|a|pipe'" (means just single value)
<input-name> :<input-value>
a value starting with ':' is treated as a file name
and the contents of this file is used as an input;
multiple (repetitive) values with file names are also allowed:
name ':file1|:file2|not-a-file'
<input-name> ::<input-value>
a value starting with two colons ('::') is treated as a file
name whose contents is a list of line-separated file names;
these files are read and their contents is sent as an array
to the service as an input (of course, it works only for the
services that expects such input);
Note that this feature cannot be combined with a
repeatable value;
-<input-name>
can be used for boolean parameters and it means the same as:
<input-name> true
where <results> defines if any results will be returned back:
-ri ... return info (such as size) about all results
-r ... return all results
-r <list of names> ... what results to retrieve,
available result names can be found using option -o
(no list means getting all available results);
Note: the binary results are written into a file,
or into several files, with unique names
(unless given an almost undocumented
"binary print" option -bp)
-sr saves results (except 'report') to the current
working directory
Last modified: Mon Feb 4 17:27:32 2008