This method attaches SpecTcl to a new data source. Data analysis from any prior source is terminated. If successful, it's still necesary to start analysis on the data source. The method returns None
The actual operation of the method is determined by
entirely by keyword parameters. No positional parameters
are allowed. With the exception of
type and name, all
other keywords are options. Allowed keywords are:
typeThe value of this keyword must be a string that contains either file, if data are to be analyzed from a file, or pipe if data are analyzed from a pipe program.
name
The value of this keyword is a string. It's
interpretation depends on the value of
type above.
If type='file',
the name value is a path
(absolute or relative) to the file from which
data will be analyzed.
If type='pipe'
the name value is a program
and its arguments. The specified program must
output valid data to its stdout. The specified
program's stdout will be connected to a pipe
from which SpecTcl will read data.
Note that pipe data sources are how SpecTcl takes data from online systems.
size
The value of this keyword must be an integer.
SpecTcl will read data from its data source
using a buffer with as many bytes as
specified by size.
This keyword is optional and its value
defaults to 8192.
It is recommended that the size be large enough to hold several typical events.
formatThe value of this keyword is a string that specifies the data format that SpecTcl assumes the data is in. The legal values are ring10 or ring11. This keyword is optional and defaults to ring11.
Note that if Spectcl sees ring format items but ring10 was specified, it will override the specification to the data format specified by the format item. If the source is online and the run has already been started, the initial ring format item will not be seen and the specified format will be used.
Example 3-1. Attaching a file data source
set filename [file normalize [file join ~ stagearea experiment run1 run-0001-00.evt]]
python exec {
spectcl.attach(type='file', name='$filename')
}
Note that Tcl is used to generate and normalize
the event filename in the stagearea. The resulting
filename is used by the python script as the
name for the attach.
Example 3-2. Taking data from an NSCLDAQ ringbuffer
python exec {
import os
import getpass
daqbin=os.getenv('DAQBIN')
host =os.getenv('DAQHOST')
program = daqbin + '/ringselector --sample=PHYSICS_EVENT --non-blocking '
user = getpass.getuser()
uri = 'tcp://' + host + '/' + user
program = program + '--source=' + uri
spectcl.attach(type='pipe', name=program)
}
This example assumes thate:
The DAQBIN environment variable points at a valid bin subdirectory of an NSCLDAQ installation.
The environment variable DAQHOST contains the name of the system on which the ringbuffer from which we want data lives.
We want to take data from the default ringbuffer (the logged in username) from that host.
Note that the combination of
--sample and
--non-blocking on the
ringselector command
ensure that SpecTcl will not determine
the maximum data taking rate.
Example 3-3. Reading data from a gzipped event file
set filename [file normalize [file join ~ stagearea experiment run1 run-0001-00.evt.gz]]
python exec {
program = 'zcat $filename'
spectcl.attach(type='pipe', name=program)
}
The output of zcat, the uncompressed event file,
is attached via a pipe to SpecTcl. As in the
first example, we use Tcl to generate the filename.
We could equally well have done this using
os.path.join and
os.path.abspath
completely in the python script.