Mayan for home/private use: Experiences, hints, best practices?

[Manuel Reiter]

I want to use Mayan as a document management system for my private/home stuff (bills, bank statements, etc.) to try and get rid of some paper. However, I'm somewhat overwhelmed by the possibilities and options available.

So, if any of you are using Mayan in a similar fashion, could you describe your setup a little? What document types do you have, which experiences did you make, any setup errors that took you a long time to correct, hints, best practices you've found, handy tips you'd like to share?

Thanks in advance for your contributions!

[Manuel Reiter]

One more thing I forgot: What's your storage and backup policy?

[Mirco Hansen]

Hi,

I use it for my private stuff. I run the docker image on an Intel NUC with SSD and 8 GB RAM. I scan my paperstuff and save it as PDF. OCR works very well. I would not run Mayan on a Raspberry as OCR can take some time.

The most complex thing were authorizations because besides me my girlfriend uses Mayan and we have some stuff we both have to access. But after all there is nothing complex about Mayan. Just install it and use it :-)

Regarding the backup: I do backups in the local filesystem with rsnapshot and encrypted remote backups in a cloud with duplicity.

Hope this helps :-)

Cheers
Mirco

[Manuel Reiter]

Hi Mirco,

thanks for sharing your thoughts on this! So far, I'm actually surprised
how well Mayan runs on the Pi - I've only scanned and uploaded a few
sample documents though. I guess I'll see how far the Pi will get me and
upgrade to a beefier system once I hit its limits.

If I may ask, how do you access/categorize your files? Via folders?
Indices (which I still don't completely understand)? Tags? A
combination? Which approach has proved most practical for you?

All the best,

Manuel

[Mirco Hansen]

Hi Manuel,

who can access a document is controlled by the document type (my girlfriend, me or both). And then I run a mixture between folders and tags. Folders are very toplevel (e.g. banking) and tags are more precise (e.g. bank statement account 1, bank statement account 2). So folders are some top level sort criteria and the detailled criteria are the tags.
The disadvantage of this approach is that everybody can see all tags so you have to be precise and set up a naming convention if don't use it on your own. Would be nice if there were ACLs for tags.

Currently I don't see any benefit in indices for me and I must assume that I don't understand them (but I didn't spend so much time on this topic).

Quite interesting that it runs well on a Pi... I had a document which was the terms and conditions of my cable TV (13 pages, 4 columns in font size 3 or so ;-)) and the OCR run one hour on it.

Cheers
Mirco

[Manuel Reiter]

Hi Mirco,

thanks, I'll experiment a bit with these approaches. May I also ask how your scanning process is organised? Do you use staging or watch folders? Do you process your scans in any way before adding them to Mayan?

As for running well on the Pi: Well, so far I've only added one page documents, so we'll see how well it goes with larger one. But I'll have to get the scanning sorted out first. And find out how to remove a tag from a document...

All the best,

Manuel

[Mirco Hansen]

Hi Manuel,

my scan "process" is quite simple :)

My scanner stores the scans on a shard folder on my NAS. From there I add them via drag n drop to Mayan via add new document. So I have control over the tags.
I do not any preprocessing of the scans... I just add them to Mayan via the webinterface and do not use any kind of staging or watch folders.

Cheers
Mirco