[ 
https://issues.apache.org/jira/browse/UIMA-60?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Richard Eckart de Castilho updated UIMA-60:
-------------------------------------------
    Labels: Stale  (was: )

This issue is marked as "stale" due to inactivity for 5 years or longer. If no 
further activity is detected on this issue, it is scheduled be closed as 
'unresolved' in 3 months time from now (Dec 2016).

> Improve how progress is reported in CPE
> ---------------------------------------
>
>                 Key: UIMA-60
>                 URL: https://issues.apache.org/jira/browse/UIMA-60
>             Project: UIMA
>          Issue Type: Improvement
>          Components: Collection Processing
>            Reporter: Adam Lally
>            Priority: Minor
>              Labels: Stale
>
> The CPE.getProgress() method is implemented by calling the 
> CollectionReader.getProgress() method.  This has a number of problems.  For 
> one thing the CollectionReader reports the number of CASes it has read from 
> the collection, but this does not indicate how many CASes have been fully 
> processed by the CPE.  Because of this, in the CPE GUI the progress bar "runs 
> ahead" of the actual number of elements processed by about the size of the 
> CAS Pool.  (If you have a CAS Pool of size 10, the CPE progress will report 
> that it has processed 10 documents as soon as the CollectionReader has read 
> all 10 CASes.
> It does not make much sense to rely on the CollectionReader for this 
> information, since the CPE *knows* how many elements have been processed 
> (indeed it already sends callbacks upon completion of each CAS).
> Also the getProgress() API returns an array of Progress objects, which gets 
> ugly.  The original idea was that the CollectionReader might report progress 
> in different units (i.e. the number of documents proccessed and the number of 
> bytes processed).  No one ever uses this and it makes the API more 
> complicated to use and implement. 
> The only thing we may want to ask the CollectionReader for is an estimate of 
> how many elements there are in the collection.  So a simple way to go might 
> be to replace CollectionReader.getProgress() with something like 
> CollectionReader.getCollectionSize().  The CollectionReader would be allowed 
> to return -1 to indicate an unknown size.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Reply via email to