Editing Archive (section)

==UCT Reading Data Troubles==


===HowTo: fix the '''UCT_EVALUATION''', if given an error message regarding '''failure to read data'''.===  

Note: This problem is specifically refering to using the "VOI" logical to  run the evaluation. In this case, pressing "Evaluation 3D" should simply extract a normal and segmented AIM.

:*An example log file of this:	DISK1:[MICROCT.SCRATCH]UCT_EVALUATION_VOI.LOG;56

:*It was actually version 55; that crashed.  It tried to read 261 slices, but the ISQ only contains 260 slices. 

::-Check what the eval (or voi) program shows for that scan ?

::-Did you start the voi with normal pushing, or did you start it 'by hand' and gave -1 for dimension ?

::-The 'missing last slice' problem is known, but the VOI should not crash if you do not select that slice in the white box.

:::?????????? what does this mean, the missing last slice prob is known??

===HowTo: Restart the SCEVAL_FAST queue===

How to: Restart the SCEVAL_FAST queue, if somehow stopped, on our sceval.ucalgary.ca workstation.  Do I need to be user "calibration" to do this?

:* As any user, you may restart it.  Use the following command:

::<code>$ start sceval_fast</code>

:::However, it should be started every time you reboot, so difficult to say why it was stopped...




===HowTo: Select multiple samples quickly===

:* Hold the Shift and/or Ctrl key.


===How to: Search for a particular samplename===

:*You can use the samplename filter field on top, to display only the names containing the filter string.


===HowTo: Start the move/copy===

:*Bby clicking on 'Start...'

::Note: The 'Script...' button is for advanced users.


===HowTo: Use the 'Script' button===

:*This will this will generate a script:  <code>sys$scratch:archive.com</code>

:*You you may submit this in 'batch mode' later (e.g. over night) by typing:

::<code>	 $ scanco_submit_command sys$scratch:archive.com/after=23:59 </code>

::* In a multi-computer cluster setup, you may want to specify the exact queue name of the computer that has the tape hooked up by giving, e.g. 

:::<code>	/que=sceval_fast</code>

:*Insert a tape, but do NOT mount the tape in the normal program.

::Note: If necessary: if it's a new tape, initialize it as usual.

:*The batch job will be started one minute before midnight in the example above.

:*The following will show you that it is holding until then:

::<code>	 $ show entry</code>


:*For more help on this,  try the following help commands:

::<code>	$ help submit /after 

::	$ help dcl_tips date_time example</code>


===HowTo: Do Weekly and Full System BackUps===
:It is possible under some circumstances that the hard disk may crash.  Therefore a regular copy should be made of the UCT database for data security.  
:In other words, a backup should be performed each week of the system’s data for that week and each month a full backup of the system should be made.  The following describes the process to do this.
    
:* There are two special backup tapes that are used alternately for both weekly and full backups, labeled BCKP0A and BCKP0B.  
:*The BCKP stands for backup and the A and B are the tape identifiers. 
:*The reason for using two tapes is to ensure that older data is not overwritten.


[[Image: UCT Disk2.jpg|center]]


:To perform weekly backup:
:*	On Friday afternoon, insert into the tape device, the special backup tape opposite from the one used the previous week (ie. use Tape A, if Tape B was used the week before).  To determine this, check on the Weekly Backup schedule sheet.
:*	Open up the UCT_BACKUP tool, shown in the Figure above.:*•	Under the heading “Disk Information�? and “Tape:�? click on Mount.
:*	When “Tape: mounted as BCKP0_�? appears under “Disk Information�?, click on Weekly Backup at the bottom of the window.  A message will appear inquiring if the special tape for backup is inserted in the device, and click on Yes.


===HowTo: Continue writting when a tape is full===

:When the tape is full, there should be a message displayed about continuing on another tape.

:*Hit ctrl+C to interrupt the program, take a fresh tape, initialize it, and then select the remaining scans to archive again.

:*Then hit 'start'.


'''Important Reminders:'''

:*While the tape is running and being written to, DO NOT LOG OUT, DO NOT INTERRUPT THE PROGRAM !

:*This would make that tape unusable from this point on - you could still read from that tape up to this point, but could not add more scans to this tape.

:*WE STRONGLY URGE YOU TO DO THE WEEKLY BACKUP (SAFETY BACKUP OF USER DISK1 WITH YOUR DATABASES ON IT).

:*2 users DID experience disk crashes last year, and without the weekly backup, you may still have all scans on the data disk (e.g. disk2). But without the databases, you cannot tell which scans belong the which sample!

----

===HowTo: Install vtkMy with Qt on Linux===

:*qt-devel must be downloaded and installed from any rpm website and then cmake must be run with the following setting for qt:
<code>
::QT_INCLUDE_DIR                   /usr/lib/qt-3.3/include
::QT_MOC_EXECUTABLE                /usr/bin/moc
::QT_QASSISTANTCLIENT_LIBRARY      /usr/lib/qt-3.3/lib/libqassistantclient.a
::QT_QT_LIBRARY                    /usr/lib/qt-3.3/lib/libqt-mt.so
::QT_UIC_EXECUTABLE                /usr/bin/uic
</code>

===HowTo: Avoid errors due to a one-off reading ISQ problem===

:*The last slice in the ISQ cannot be read on occasion during the evaluation process - i.e., creating the AIM.  
::(Scanco is aware of this problem.)
:*The interm solution is to make sure that the last slice in the ISQ is not selected when you start the evaluation.  
:*When using the white box to set your ROI, make sure that the box does not include the last slice of the ISQ.


===HowTo: Tabulate evaluation data directly from LOG files===
: This also applies for data from the 3D database.

:* If the sample was properly imported into our database (<code>$uct_import</code>) and if the database was activated in the ipl, all the results go into our 3D database, and you can just get them out with the usual
::<code>$ uct_list</code>

:* Otherwise, the command to 'tabulate' .log files:

::<code>$ search xxxx.log "ov "</code>

:* Or, whatever you want to search for,
::<code>$ search/output=sys$scratch:ov.txt xxxx.log "ov "</code> 
:: to put the results into a .txt file.

====HowTo: Searching multiple log files====
:*To search more than one log file, and get output into one single file:

::<code>$ fda/output=xxx.txt search *.log "ov "</code>

::: The <code>fda</code> means 'file do all' and is a script of ours to do one command (search) on a wildcard of files (*.log)

:*Then put all the ov.txt, tv.txt, name.txt etc. files into excel.