You are viewing a plain text version of this content. The canonical link for it is here.

Posted to user@tika.apache.org by Tim Allison <ta...@apache.org> on 2021/05/20 15:36:21 UTC

Side effects of turning off parsers...

All,

Let's say I don't want to parse old iWorks files (zip-based file
format).  I can exclude that parser via TikaConfig, but then it gets
parsed by the PackageParser.  So, then I have to decorate the
PackageParser with mime-exclude=application/iworks... or add an
EmptyParser that handles application/iWorks.

Do we have a way to say: I only want the PackageParser to process
actual zip files and not zip-based files generally ... without having
to enumerate the zip-based files.

Thank you.

Cheers,

          Tim

Re: Side effects of turning off parsers...

Posted by Tim Allison <ta...@apache.org>.

This feels a bit like a bug and not a feature? Maybe...

Nick, what do you think?

The PackageParser winds up parsing the file with this one:
<properties>
  <parsers>
    <parser class="org.apache.tika.parser.DefaultParser">
      <mime-exclude>application/vnd.apple.keynote</mime-exclude>
      <parser-exclude class="org.apache.tika.parser.iwork.IWorkPackageParser"/>
    </parser>
  </parsers>
</properties>

and this one:

<properties>
  <parsers>
    <parser class="org.apache.tika.parser.DefaultParser">
      <mime-exclude>application/vnd.apple.keynote</mime-exclude>
      <parser-exclude class="org.apache.tika.parser.iwork.IWorkPackageParser"/>
      <parser-exclude class="org.apache.tika.parser.pkg.PackageParser"/>
    </parser>
    <parser class="org.apache.tika.parser.pkg.PackageParser">
      <mime-exclude>application/vnd.apple.keynote</mime-exclude>
    </parser>
  </parsers>
</properties>

Again, the problem is that the DefaultParser backs off to
application/zip because application/vnd.apple.keynote is a subclass of
zip.

Not sure what the fix should be...



On Thu, May 20, 2021 at 2:24 PM Furkan KAMACI <fu...@gmail.com> wrote:
>
> Hi Tim,
>
> Seems that we can exclude mime types parser only [1]
>
> How about globally excluding such mime-types? Is there any way to define it?
>
> [1] https://tika.apache.org/1.26/configuring.html
>
> Kind Regards,
> Furkan KAMACI
>
> On Thu, May 20, 2021 at 6:36 PM Tim Allison <ta...@apache.org> wrote:
>>
>> All,
>>
>> Let's say I don't want to parse old iWorks files (zip-based file
>> format).  I can exclude that parser via TikaConfig, but then it gets
>> parsed by the PackageParser.  So, then I have to decorate the
>> PackageParser with mime-exclude=application/iworks... or add an
>> EmptyParser that handles application/iWorks.
>>
>> Do we have a way to say: I only want the PackageParser to process
>> actual zip files and not zip-based files generally ... without having
>> to enumerate the zip-based files.
>>
>> Thank you.
>>
>> Cheers,
>>
>>           Tim

Re: Side effects of turning off parsers...

Posted by Furkan KAMACI <fu...@gmail.com>.

Hi Tim,

Seems that we can exclude mime types parser only [1]

How about globally excluding such mime-types? Is there any way to define it?

[1] https://tika.apache.org/1.26/configuring.html

Kind Regards,
Furkan KAMACI

On Thu, May 20, 2021 at 6:36 PM Tim Allison <ta...@apache.org> wrote:

> All,
>
> Let's say I don't want to parse old iWorks files (zip-based file
> format).  I can exclude that parser via TikaConfig, but then it gets
> parsed by the PackageParser.  So, then I have to decorate the
> PackageParser with mime-exclude=application/iworks... or add an
> EmptyParser that handles application/iWorks.
>
> Do we have a way to say: I only want the PackageParser to process
> actual zip files and not zip-based files generally ... without having
> to enumerate the zip-based files.
>
> Thank you.
>
> Cheers,
>
>           Tim
>