Re: [Tracker] Tracker for networked home document management

[Ivan Frade <ivan frade gmail com>]

Hi,

 Tracker works *per-user*. It is designed to run inside a user
session. Each user in the computer has its own databases.

 That being said, maybe with some adjustments here and there you can
get what you want. The PDF indexing is not a problem.

 More answers below:

On Wed, Apr 4, 2012 at 9:57 AM, Aaron Whitehouse
<lists whitehouse org nz> wrote:
> Hi,
>
> I have a few questions about Tracker that I was hoping you could answer:
> 1) Would we be able to store the documents on the server and access them
> from another computer? I understand that Tracker stores its index
> separately from the files, so it wouldn't be possible to simply share
> the "Documents" directory across the network and point both Tracker
> instances to it.

 Tracker stores its information under ~/.local/share/tracker and
~/.cache, but you can modify this with environment variables.

  You could create a folder with read/write permissions for both users
and point tracker to write its files in that folder. Then when tracker
boots (in any user session on the computer) they will use the same
files. This could work as long as you don't have more than one user
logged at the same time.

 Maybe some adjustments in the file permissions are needed after
creating the database, because the first time it will have only user
permissions.

> 2) We are currently intending to have just one laptop, so could store
> the files on the laptop (in a folder accessible by both users on the
> laptop). Would two users then be able to use the same sets of tags etc?

 If they share the database with the trick above yes. Tags are in the database.

 Another option is to put all your files in a folder (visible to all
users) and let tracker index it per-user. In that case each user will
have its own databases. Then the information in the database but not
in the files (E.g. tags added by the user) won't be shared.

> 3) Another feature of OpenKM that appealed was its ability to OCR
> documents and allow the user to search the full text. Is this possible
> with Tracker?

 Yes, we have a PDF extractor. It is not the fastest thing in the
universe but it should work.

> 4) Most of our documents were OCRd when we scanned them and are
> searchable PDFs. I assume that Tracker could index these like any other
> PDF. Does this sound right?

 Yes. Tracker will find the PDF, try to extract the text and index it.
There are few options to enable/disable/fine-tune it.

 Regards,

Ivan