manifoldcf-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Furkan KAMACI <furkankam...@gmail.com>
Subject Re: Linking E-Mail and Its Attachments
Date Wed, 05 Apr 2017 17:47:04 GMT
Hi Karl,

Similar systems to ManifoldCF link emails and its attachments. Let's assume
that one selected Elasticsearch as output connector. One cannot access to
original document (or cannot have necessary information to access it) when
he/she makes a search and if it hits to an attachment.

I've come up with an idea about what we can do about that problem within
ManifoldCF with its restrictions. We should add these metadata:

To Attachment:

   - From
   - To
   - Date

(metadata of email it belongs to)

To Email:

   - Filename of attachments

So, one can get necessary information about the email of attachment and
vice versa.


Kind Regards,
Furkan KAMACI



On Thu, Mar 9, 2017 at 8:21 PM, Karl Wright <daddywri@gmail.com> wrote:

> What URL is used for the attachments?  If the URL contains a way to get to
> the primary document then there would not be any need for any other kind of
> linkage.
>
> I have no problem with adding more of the metadata for attachments to the
> attachment RepositoryDocument but I don't see offhand how this can help.
>
> Karl
>
>
>
> On Thu, Mar 9, 2017 at 1:13 PM, Furkan KAMACI <furkankamaci@gmail.com>
> wrote:
>
> > Hi,
> >
> > Currently, E-mail and its attachments are indexed individually. However,
> we
> > should link them somehow because there is not a direct link between them
> > (one can not access to parent - e-mail or child - attachment from one to
> > other).
> >
> > First of all, when we index e-mail documents, we have optional metadata
> > fields for attachments:
> >
> > Encoding of Attachment
> > Mime type of attachment
> >
> > it may be useful to store name of attachments (also ids of attachments)
> to
> > softly link them. I used the term 'softly', because a hardly linked
> > solution is similar to nested documents of Solr but it may break the
> >  genericness.
> >
> > What do you think about it?
> >
> > Kind Regards,
> > Furkan KAMACI
> >
>

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message