[ 
https://issues.apache.org/jira/browse/TIKA-2276?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Tim Allison updated TIKA-2276:
------------------------------
    Description: 
If we run the AutoDetectParser() against the files in our unit tests (around 
600 files*), there are 701 new instantiations of TikaConfig.  The time is 
around 20 seconds.  If we modify AutoDetectParser to pass its TikaConfig via 
the ParseContext if one isn't already specified, that drops to 234 
instantiations, and parse time goes to ~17 seconds.

Let's make this simple change and look for other areas to decrease the number 
of times our parsers are creating a new TikaConfig.

*Note I did not include the testCHM2.chm monster in these runs.

  was:
If we run the AutoDetectParser() against the files in our unit tests (around 
600 files), there are 701 new instantiations of TikaConfig.  The time is around 
20 seconds.  If we modify AutoDetectParser to pass its TikaConfig via the 
ParseContext if one isn't already specified, that drops to 234 instantiations, 
and parse time goes to ~17 seconds.

Let's make this simple change and look for other areas to decrease the number 
of times our parsers are creating a new TikaConfig.


> Try to be more parsimonious creating TikaConfigs
> ------------------------------------------------
>
>                 Key: TIKA-2276
>                 URL: https://issues.apache.org/jira/browse/TIKA-2276
>             Project: Tika
>          Issue Type: Improvement
>            Reporter: Tim Allison
>
> If we run the AutoDetectParser() against the files in our unit tests (around 
> 600 files*), there are 701 new instantiations of TikaConfig.  The time is 
> around 20 seconds.  If we modify AutoDetectParser to pass its TikaConfig via 
> the ParseContext if one isn't already specified, that drops to 234 
> instantiations, and parse time goes to ~17 seconds.
> Let's make this simple change and look for other areas to decrease the number 
> of times our parsers are creating a new TikaConfig.
> *Note I did not include the testCHM2.chm monster in these runs.



--
This message was sent by Atlassian JIRA
(v6.3.15#6346)

Reply via email to