Indexing Issue| Email (application/vnd.ms-outlook - msg) with Attachment | Uploaded Via Explorer

[milochanzy]

Hi,

I'm struggling to get one thing sorted out in the Alfresco vanilla version 4.2.1.7. The context is, I'm trying to upload via "explorer" an email which has attachments to it. Now I'm able to see that indexing is happening on the email (application/vnd.ms-outlook - msg) content and metadata however the attachment content is not getting indexed. The attachment is simple word (2010) document (but it could be excel or pdf or text).

I browsed through the implementation and I see Tika parse OutlookExtractor is used to parse the email and attachments in the email. The URL (http://localhost:8080/alfresco/service/mimetypes?mimetype=application/vnd.ms-outlook#application/vnd...) gives following:

application/vnd.ms-outlook - msg
Extractors: org.alfresco.repo.content.metadata.MailMetadataExtracter
Transformable To:
application/xhtml+xml = org.alfresco.repo.content.transform.MailContentTransformer
text/html = org.alfresco.repo.content.transform.MailContentTransformer
text/plain = org.alfresco.repo.content.transform.MailContentTransformer
text/xml = org.alfresco.repo.content.transform.MailContentTransformer
Transformable From: Cannot be generated from anything else

So ideally the email attachment should get indexed too, right? But its not . I would like to get your opinion on how can I fix this issue? Please let me know if you like to see any specific config file I've used. I've not modified anything however from default implementations.

I'm using Tika 1.5 and POI 3.10.

Thanks,
Milan.

[mrogers]

No.   Sorry.  

"Compound types" are a problem that I think still needs solving.   And its been that way for a while.   We also have the same issue with contents of a zip file.

[milochanzy]

Hi Rogers,

Thanks for your attention. So is it fair to conclude that Alfresco does NOT support indexing of email attachment contents in version 4.2.1.x? Also, if so, is there another version that DOES support this functionality, Say V 5.x?

How would you suggest developer should approach to customize to add this feature?

Appreciate your time!

Thanks,
Milan.