Appion

Notes from Recon meeting » History » Revision 2

Revision 1 (Amber Herold, 06/22/2011 03:39 PM) → Revision 2/8 (Amber Herold, 06/22/2011 03:59 PM)

h1. Notes from Recon meeting 

 Moving forward, refinements will all be split into 2 steps, prep and run. 

 h2. Prepare refine 

 When the user selects to prep a refinement, a web form is provided to select the: 
 # refinement method - eman, xmipp, frealign, etc...  
 # stack 
 # model 
 # run parameters - runname, rundir, description 
 # stack prep params - lp, hp, last particle, binning 

 The web then calls prepRefine.py located on the local cluster to prepare the refinement. 

 h2. Run Refine 

 When the user selects to run a prepared refinement, a web form is provided to select the: 
 # prepped refine  
 # cluster parameters - ppn, nodes, walltime, cputime, memory, mempernode 
 # refine params, both general and method specific 

 The web server will then: 
 # verify the cluster params by checking default_cluster.php 
 # if needed, copy the stack and model to a location that can be accessed by the selected cluster 
 # verify the user is logged into the cluster 
 # pass the list of commands to runJob.py (extended from the Agent class), located on the remote cluster 

 runJob.py will: 
 # format the command tokens in a dictionary of key-value pairs 
 # set the job type which was passed in the command 
 # create an instance of the job class based on the job type 
 # create an instance of the processing host class 
 # launch the job based via the processing host 
 # update the job status in the appion database (do we have db access from the remote cluster?) 

 h2. Object Model 

 h3. Processing Host 

 Each processing host (eg. Garibaldi, Guppy, Trestles) will define a class extended from a base ProcessingHost class.  
 The extended classes know what headers need to be placed at the top of job files and they know how to execute a command based on the specific clusters requirements. 
 The base ProcessingHost class could be defined as follows: 
 <pre> 
 abstract class ProcessingHost(): ProcessingHost: 
     def generateHeader(jobObject) # - abstract, extended classes should define this, returns a string 
     def executeCommand(command) # - abstract, extending classes define this 
     def createJobFile(header, commandList) # - defined in base class, commandList is a 2D array, each row is a line in the job file. 
     def launchJob(jobObject) # - defined in base class, jobObject is an instance of the job class specific to the jobtype we are running 
         header = generateHeader(jobObject) 
         jobFile = createJobFile(header, jobObject.getCommandList()) 
         executeCommand(jobFile) 
 </pre> 

 h3. Job 

 Each type of appion job (eg Emanrefine, xmipprefine) will define a class that is extended from a base Job class. 
 The extending classes know parameters that are specific to the job type and how to farmat the parameters for the job file. 
 The base Job class could be defined as follows: 
 <pre> 
 class Job(): 
     self.commandList 
     self.name 
     self.rundir 
     self.ppn 
     self.nodes 
     self.walltime 
     self.cputime 
     self.memory 
     self.mempernode 
     def __init__(paramDictionary) 
         self.commandList = self.createCommandList(paramDictionary)    
     def createCommandList(paramDictionary) # defined by sub classes, returns a commandList which is a 2D array where each row corresponds to a line in a job file 
 </pre>