ODB Access and Use

From MidasWiki
Revision as of 18:13, 21 August 2014 by Suz (talk | contribs)
Jump to navigation Jump to search

Links

Purpose

This page gives details on how users may access and use the ODB.

Introduction

The MIDAS Online Data Base (ODB) is the main communication platform between MIDAS applications. It contains all the information needed to run an experiment. A list of all the ODB trees used by the system can be found in Online Database.

ODB Keys

The ODB is hierarchically structured, similar to a file system, with directories and sub-directories (or trees and subtrees). The data are stored in key/data pairs, similar to the Windows NT registry. ODB keys can be dynamically created and deleted. The data associated with a key can be of different types such as: byte, words, double words, float, strings, etc. or arrays of any of those. A key can also be a directory or a symbolic link. ODB Key names are case-independent, and may contain spaces.

Accessing the ODB

The MIDAS ODB can be accessed interactively in two ways:

For non-interactive access,

  • The MIDAS library provides a complete set of functions (in C/C++) to manage and operate on ODB keys (see odb functions).

Examples of many of these routines in use can be found in Frontend user code.

  • Scripts can easily access the odb using the odbedit -c argument. Examples of this can be found in Script Access
  • A set of Javascript functions are provided to allow mhttpd custom pages to access the ODB.



Creating the ODB

After installation of MIDAS, before any other applications are started, the ODB must be created by the user (see creating the ODB ). Running odbedit to create the ODB also causes the trees |/Runinfo, /Experiment, /System to be created and filled with default values. Other system applications will automatically create their own initial ODB structure filled with default values (for example, the data logger creates the /Logger ODB tree. The user may then modify these configurations to fit his/her requirements (see customizing). A list of the trees in the ODB used by the system can be found here.

Shared-memory files

Creating the ODB also creates all the shared-memory files needed for the experiment (see list below). By default, these files will be created in the area indicated in the exptab file for your experiment.

.ALARM.SHM  .ELOG.SHM  .HISTORY.SHM  .MSG.SHM  .ODB.SHM  .SYSMSG.SHM  .SYSTEM.SHM

Customizing the ODB

The user customizes the ODB for their experiment by modifying the values of "system" parameters (e.g. Keys in the ODB /Logger tree, or adding extra keys under existing trees as needed, e.g. "Edit on Start" parameters under the /Experiment ODB tree. The user may creates "optional" trees e.g. (/Script ODB tree, /Custom ODB tree) populating them with keys as required. The user is also free to create his/her own tree structure under root (/) for his/her own purposes.

Creating ODB keys

ODB Key names are case-independent and may contain spaces. Key names with spaces must be enclosed in quotes.

The easiest way to create, set (and delete if necessary) ODB keys is to use the mhttpd ODB page. Buttons are provided to create and delete keys, and the values can be set by clicking on the appropriate key.

Alternatively, odbedit can be used as demonstrated in the following example. Odbedit is more powerful, also allowing the user to for example re-order, copy, rename the keys (see odbedit command list.

Once a directory is created, keys can be created of the types supported by MIDAS, e.g. INT DWORD BOOL FLOAT DOUBLE STRING (defined in midas.h). Arrays of all these types can also be created.

Values are assigned to the keys either by using the ODB Page or the odbedit set command.

Example

This example shows how to create an ODB tree (/Custom) using odbedit. Creating the same tree using the ODB Page is self-explanatory.

The odbedit commands are shown in red. Comments are shown in green and preceded by a "#".

$ odbedit   
[local:midas:S]/>pwd    # show current directory (tree) 
/
[local:midas:S]/>mkdir custom/images    # make directory "custom" and subdirectory "images"
[local:midas:S]/>ls  # list all keys
System
Programs
Experiment
Runinfo
Alarms
Custom
[local:midas:S]/>cd custom   # change directory to "/custom" 
[local:midas:S]/custom>ls
images
[local:midas:S]/custom>create string "custom page&"  # Create a key of type STRING. Note use of quotes as key name contains a space 
String length [32]: 128
[local:midas:S]/custom>set "custom page&" /home/midas/custom/custom.html  # Set key to a value. 
#             Could type instead "set cus" followed by tab to demonstrate tab-completion 
[local:midas:S]/custom>ls -r  # list recursively 
custom
   images
   custom page&           /home/midas/custom/custom.html

Creating arrays

Creating arrays in the ODB is simple using the ODB Page.

The following example demonstrates the use of odbedit to create an array, set its elements to various value, and expand it. An array can be truncated using the odbedit trunc command.

[local:pol:S]/test>create int fred[5]
[local:pol:S]/test>set fred[*] 5
[local:pol:S]/test>set fred[1..3] 6
[local:pol:S]/test>set fred[8] 9
[local:pol:S]/test>ls
fred
                               5
                               6
                               6
                               6
                               5
                               0
                               0
                               0
                               9


Reorder ODB keys

Sometimes it may be desired to reorder the ODB keys once created. The [odbedit] move command is used for this purpose as in the following example:

The key "custom" can be moved to the top (or bottom) of the list, e.g.

[local:midas:S]/>ls
System
Programs
Experiment
Runinfo
Alarms
Custom

The key "Custom" can be moved to the top

[local:midas:S]/>move custom top
[local:midas:S]/>ls
Custom
System
Programs
Experiment
Runinfo
Alarms

or to any position, e.g.

[local:midas:Stopped]/>move custom 1
[local:midas:Stopped]/>ls
System
Custom
Programs
Experiment
Runinfo
Alarms

Create a link in the ODB

Creating a link (softlink) is easy to do using the ODB Page.

An example of creating softlinks with odebdit can be found in Creating edit-on-start parameters.

Save and reload the ODB

The ODB can be saved at its current position in several formats (ascii, xml as a c structure) using the odbedit save command, and reloaded from a saved file with the odbedit load command.

Since the ODB may become corrupted, a copy is saved automatically at the beginning of each run. Since this file is overwritten each time, users often also save a copy (that will not be overwritten) at the end of each run (see ODB Dump) along with the data for that run. Users can then return their ODB to the state it was in for any particular run.

The entire database need not be loaded. Saved ASCII files can be made of just a part of the database, and these can be reloaded into the database. Since the full path is given in the saved file, the file can be loaded from any position in the database.

The saved ASCII file may of course be edited prior to loading, if keynames or values need to be changed. If the keys in the load file do not exist, they will be created. If they do exist, the values from the file will be loaded.



experim.h include file

The odbedit make command creates "experim.h", a file containing C structures which can be included into frontend and analyzer code to enable easy access to the ODB parameters (see also Frontend). The file is created in the current directory. e.g.

[local:midas:S]/>make
Analyzer "Analyzer" not found in ODB, skipping analyzer parameters.
"experim.h" has been written to /home/midas/online

The file experim.h contains C structures for the ODB trees /Experiment, /Analyzer/Parameters, and /Equipment/<equipment-name>/Settings if these trees are present. All C structures are accompanied with a string representation which can be used in the db_create_record function to setup an ODB structure which matches the required C structure, ensuring that the required keys are present in the ODB.

In order to include the analyzer section, the ODB key /<Analyzer>/Parameters has to be present, where <Analyzer> is the name of the analyzer. The command used is then "make <Analyzer> ".

An example experim.h can be found at [1].

An experim.h file is often included in the Frontend code, and also when setting up a Hot-Link.



Recover from corrupted ODB

If the ODB becomes corrupted, odbedit may no longer work, and other clients will also fail to open the database. In this case, the old ODB should be deleted and a new one created. The contents of the ODB can be reloaded from a (see #Save and reload the ODB|saved ODB file]].

To delete the corrupted ODB, delete the *.SHM files for your experiment.

$ ls .*.SHM
.ALARM.SHM  .ELOG.SHM  .HISTORY.SHM  .MSG.SHM  .ODB.SHM  .SYSMSG.SHM  .SYSTEM.SHM
$ rm .*.SHM

Create new *.SHM files by re-creating the ODB, then load a saved file containing the latest copy of the ODB contents.


Accessing the ODB from a script

This external command feature of odbedit allows for sophisticated scripts to be created that can manipulate the odb. Such scripts can for example

  • check ODB parameters prior to beginning of run
  • send run parameters to the electronic logbook
  • set some ODB parameters when a script button on the mhttpd main status page is pressed
  • load a particular saved odb file when a button on a custom page is pressed (e.g. load a "tune")


Shell scripts sending odbedit commands


Example script

The following is part of a script run at end-of-run, where it reads some parameters from the odb and saves them in a temporary file to be sent to the elog.

To make the script run at end of run, the name of the script is entered in the "Execute on stop run" key in the The ODB /Programs tree .

#!/bin/tcsh
# This script is started at the end of each run. It takes some parameters
# from the odb and creates an entry in the elog 
.....
# Start collecting information from ODB first
set Run_number = `odb -e $MIDAS_EXPT_NAME -c 'ls "/Runinfo/Run number"'`
set number = `echo $Run_number | awk '{print $3}'`
set sample = `odb -e $MIDAS_EXPT_NAME -c 'ls "/Experiment/Edit on Start/sample"'`
set Sample = `echo $sample | awk '{print $2}'`
.....
# Now create the temporary file to be sent to the elog
echo "Run # $number" >> $fin
odb -e $MIDAS_EXPT_NAME -c 'ls "/Runinfo/Start time"' >> $fin
odb -e $MIDAS_EXPT_NAME -c 'ls "/Runinfo/Stop time"' >> $fin
echo "$Sample at T = $T K, H = $H T and RF = $RF mW">> $fin
.....