[
https://issues.apache.org/jira/browse/TIKA-3812?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17612617#comment-17612617
]
David Pilato commented on TIKA-3812:
------------------------------------
I'm still having issues with 2.5.0.
Basically my OCR tests are working as expected with 2.4.0. It's breaking in
2.4.1 and 2.5.0.
I'm trying to do OCR on a PNG image.
The image I'm trying to run OCR on is available at
[https://github.com/dadoonet/fscrawler/blob/master/test-documents/src/main/resources/documents/test-ocr.png]
I don't see any metadata extracted with 2.5.0. I have no idea what is happening
and if I need to change something on my end...
> Parser Order: image get parsed by GDALParser instead of TesseractOCRParser
> --------------------------------------------------------------------------
>
> Key: TIKA-3812
> URL: https://issues.apache.org/jira/browse/TIKA-3812
> Project: Tika
> Issue Type: Bug
> Components: parser
> Affects Versions: 2.4.1
> Reporter: Eugen Caruntu
> Priority: Minor
> Fix For: 2.5.0
>
> Attachments: parser-diffs.tgz
>
>
> The selected parser seems to be different in 2.4.1. For example sending an
> image (jpg/png) that was previously (2.4.0) processed by TesseractOCRParser,
> now gets parsed by GDALParser.
> Seems that when multiple parsers support same file types, the selected parser
> depends on the order in which they get loaded.
> For example the GDALParser, ImageParser and TesseractOCRParser all support
> image/jpeg, image/png, image/gif ...
> A recent change is reversing the parser order (TIKA-3750).
> Re-configuring the GDALParser by excluding the image mime types might work,
> but there could be other duplicated parsers.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
