Difference between revisions of "Sequencer"

From MidasWiki
Jump to navigation Jump to search
m (Revised main text a bit)
Line 1: Line 1:
  
A new sequencer for starting and stopping runs was implemented in Jun 2011. Since then development has continued, with more commands being added, and the option of using a Midas Script Language (MSL) for commands, in addition to the XML language.
+
A new sequencer for starting and stopping runs was implemented in June 2011. Since then development has continued, with more commands being added, and the option of using a Midas Script Language (MSL) for commands, in addition to the XML language.
  
The sequencer runs independently from mhttpd, but an ODB [[/Sequencer ODB tree|Sequencer Tree]] is created if not already present once the Sequencer button on the main status page is pressed (if not shown see [[/Experiment ODB tree #/Experiment/Menu Buttons |Menu Buttons]]). The sequencer runs scripts in XML or MSL (Midas Script Language) format, which reside on the server (where mhttpd is running).
+
The sequencer runs inside mhttpd, but an ODB [[/Sequencer ODB tree|Sequencer Tree]] is created if not already present once the Sequencer button on the main status page is pressed (if not shown see [[/Experiment ODB tree #/Experiment/Menu Buttons |Menu Buttons]]). The sequencer runs scripts in XML or MSL (Midas Script Language) format, which reside on the server (where mhttpd is running).
  
The sequencer is stateless, meaning that even if mhttpd is stopped and restarted, the active sequence operation is still performed.
+
The sequencer state is completely stored in the ODB, meaning that even if mhttpd is stopped and restarted, the active sequence operation continues exactly where it has been stopped.
  
 
Refer to the [[/Sequencer ODB tree|Sequencer Tree]] for more details on its ODB organization.
 
Refer to the [[/Sequencer ODB tree|Sequencer Tree]] for more details on its ODB organization.

Revision as of 00:08, 19 December 2013

A new sequencer for starting and stopping runs was implemented in June 2011. Since then development has continued, with more commands being added, and the option of using a Midas Script Language (MSL) for commands, in addition to the XML language.

The sequencer runs inside mhttpd, but an ODB Sequencer Tree is created if not already present once the Sequencer button on the main status page is pressed (if not shown see Menu Buttons). The sequencer runs scripts in XML or MSL (Midas Script Language) format, which reside on the server (where mhttpd is running).

The sequencer state is completely stored in the ODB, meaning that even if mhttpd is stopped and restarted, the active sequence operation continues exactly where it has been stopped.

Refer to the Sequencer Tree for more details on its ODB organization.

Sequencer Commands

The following commands are implemented in the MIDAS sequencer. The upper syntax is for the XML file, the lower for the MSL (Midas Script Language).

Variable names are indicated in italic, options are enclosed in [brackets].

< !DOCTYPE RunSequence[< !ENTITY name SYSTEM "name.xml" >]> &name; INCLUDE name Include another XML file
<Call name="name"> a,b,c...</Call> CALL name, a, b, c, ... Call a subroutine. Optional parameters a,b,c... are passed to the subroutine, where they can be referenced via $1, $2, $3, etc. The subroutine can either reside in the current file, or in a library file which is included.
<Cat name="name"> a,b,c... </Set> CAT name, a, b, c, ... concatenates the strings a,b,c,... into a single variable name . Can be referenced with $name. If a string must contain a comma, it can be enclosed in quotes such as in <Cat name="title">Run $run, ",", $n events</Cat> <ODBSet path="/Experiment/Run parameters/Title">$title</ODBSet>
<Comment> comment </Comment> COMMENT comment a comment for this XML file, for information only. This comment is shown when one clicks on the XML file in the browser inside the web page and can be used to show some information about an XML file. This can be helpful if one has many XML files and the file name only is not enough to supply enough information.
<Goto line="n" / > GOTO n Jump to line n in scrip
<If condition="con" >...</if> IF con ... ENDIF Statements between <if> and </if> are only executed if condition con is true. The condition can use variables via $name and/or constants together with operators "<", "<=", ">", ">=", "==", "!=", "&" (bitwise AND).
<Library name="name" >...</Library> LIBRARY name Indicates that the current file is a library (which can be included by other files). A library usually consists of a set of subroutines.
<Loop [n="n"] [var="name" values="a,b,c,..."]> ... </Loop> LOOP n | a, b, c, ... ... ENDLOOP to execute a loop n times. For infinite loops, "infinite" can be specified as n. Alternatively, a list of values can be specified. In this case, the loop is executed once for each value, which is stored into the variable name so it can be accessed inside the loop via $name.
<Message [wait="1"]>message </Message> MESSAGE message [,1] Opens a message box in the browser containing the text message. If wait="1", then the sequencer waits until the box is closed by the user before continuing.
<ODBInc path="path" > delta </ODBInc> ODBINC path, delta to increment a value in the ODB.
<ODBSet [notify="0|1"] path="path" > value </ODBSet> ODBSET path, value [, 0|1] to set a value in the ODB. The notify flag specifies if possible hot-links to this ODB value are notified. This can be useful if a front-end program has many parameters, which must be set for a specific run. The front-end usually has a hot-link to its parameters, so on each modification a callback function in the front-end is called which usually modifies some hardware. If there are many ODBSet statements for all parameters, the callback would be executed on each ODBSet, so the hardware would be reconfigured many times slowing down the startup of a run. In order to prevent this, notify="0" can be specified for all ODBSet statements except the last one which contains notify="1", so the callback function is called only once at the end.
<ODBSubdir [notify="0/1"] path="path" >...</ODBSubdir> ODBSUBDIR path [,1] ...ENDODBSUBDIR If one wants to se several ODB values in the same directory, the ODBSet commands can be rather long if thepath is long. Using this command, the subdir can be specified for all commands between the opening and ending tags. So instead of writing

<ODBSet path="/Very/long/path/into/the/odb/value1">1</ODBSet> <ODBSet path="/Very/long/path/into/the/odb/value2">1</ODBSet> <ODBSet path="/Very/long/path/into/the/odb/value3">1</ODBSet> <ODBSet path="/Very/long/path/into/the/odb/value4">1</ODBSet> one can write <ODBSubdir path="/Very/long/path/into/the/odb"> <ODBSet path="value1">1</ODBSet> <ODBSet path="value2">1</ODBSet> <ODBSet path="value3">1</ODBSet> <ODBSet path="value4">1</ODBSet> </ODBSubdir>

<Param name="name" [comment="comment"][options="a,b,c,..."] [type="bool"] /> PARAM name, comment, [a,b,c | bool] When starting a script, a start page is shown where one can enter additional parameters for the script. Parameters can be defined either centrally for all scripts in the ODB under /Experiment/Edit on sequence. This subdirectory in the ODB can contain links to ODB values, which are queried at the start page. In addition, each script can define additional parameters appended to this list. They will be stored under /Sequencer/Variables and can be referenced inside the script via $name, where name can be any variable name specified in the parameter statement. The optional "comment" is shown on the start page below the parameter name. If only certain options are possible for the parameter, they can be defined via the options list. The web page will then contain a drop-down list showing the options. In case of type="bool", a checkbox will be shown.
<RunDescription> Description </RunDescription> RUNDESCRIPTION description a run description which is stored under /Experiment/Run Parameters/Run Description .
<Script [params="a,b,c,..."]> Script </Script> SCRIPT script [, a, b, c, ...] to call a script on the server side. Optionally, pass parameters to the script.
<Set name="name"> value </Set> SET name, value sets the variable name to value. Can be referenced with $name.
<Subroutine name="name">...</Subroutine> SUBROUTINE name ... ENDSUBROUTINE Declares a subroutine which can be called via <Call>.
<Transition>Start | Stop</Transition> TRANSITION start | stop | pause | resume to start, stop, pause or resume a run
<Wait for="events | ODBvalue | seconds" [path="ODB path"] [op=">=/>/<=/</==/!="]> value </Wait> WAIT events | ODBvalue | seconds, [ODB path], [op], [value] wait until a number of events is acquired (testing /Equipment/Trigger/Statistics/Events sent), or until a value in the ODB exceeds value, or wait for value seconds. If the operand op is given, the ODB value is compared with value using this operand. So one could wait until an ODB value is equal to value or becomes smaller than value. Here is an example of such a statement which waits until some high voltage is below 100 V.

WAIT ODBvalue, /Equipment/HV/Variables/Measured[3], <, 100

MSL Example

The following example is a simple script, which asks for a number of runs, then executes the runs, each lasting ten seconds.

comment "This is the MSL test file"
param runs
LOOP $runs
     TRANSITION START
     WAIT 10
     TRANSITION STOP
ENDLOOP

XML Example

Attached is a simple script which can be used as a starting point.

<?xml version="1.0" encoding="ISO-8859-1"?>
<RunSequence xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:noNamespaceSchemaLocation="">
<Comment>This is the XML test file</Comment>

<ODBSet path="/Experiment/Run Parameters/Comment">Test comment</ODBSet>
<RunDescription>Test Run</RunDescription>

<!-- do 10 runs -->
<Loop n="10">
   <!-- increment high voltage and wait for some time to settle -->
   <ODBInc path="/Equipment/HV/Variables/Demand[0]">10</ODBInc>
   <Wait for="seconds">10</Wait>

   <!-- start a run, acquire 3000 events -->
   <Transition>Start</Transition>
   <Wait for="events">3000</Wait>
   <Transition>Stop</Transition> -->
</Loop>
</RunSequence>