netvandal /home/michele


Trajan, all your files (and relation) only a click away in your browser

What is trajan: Trajan is a small http server, written with libsoup and tracker, that expose a small search engine, that permit you to search your data, see it and look at the relation between resources with your browser. This at the moment, in future can be used as feed reader, to look at your photos, contact and so on in a completely new way, using the power of sparql query language and tracker.

Why "trajan": The name was suggest by a madbob (my mentor, my "musa", and a good friend :P ), for it's assonance with "tracker" and in reference of the famous "trajan bridge", as a link beetween two very distance word, the web and the desktop.

The "Use Case": I have a small file server, i save there all my files, all my "really important" stuff.

Usually i don't keep file longer than necessary on my laptop, so i've a huge repository of file on my file server, and when i need something i've to remember where i put it, wich hierarchy system i have in my mind when i saved it.

I usually don't remember it, so i use find/grep for get my files. But it's a pain.

So here came tracker+trajan. Tracker are able to crawl my files and directory, index it and extract usefull metadata from my files. I can use also some different miner for fetching varius thing from the web, like rss, facebook message and so on.

All great, but... usually my file is on my file server, so i don't have nautilus integration, t-s-t or any fancy doc bar.

I can log with ssh, and use console tools but my girlfriend can't, and is not so exiting.

So i start writing trajan, a small http server, that expose a small search engine based on sparql and tracker, that allow me to search, find and view (and preview) my files, my data, and see the relation between them, and also browse semantically through linked resource. In future relevant metadata could be displayed in relevant way (geodata on maps, contact rich feautures page, and so on).  Combining tracker, sparql, metadata and html 5 we can build exiting new things.

Actually the code is a bit messed up, and probably leaking here and there.

The code is here:

- Search results: Schermata
- Video preview: Schermata-3
- Docs preview: Schermata-1
- Metadata dump: Schermata-2

By the way: the fster mentioned in the images is another cool tracker-based project, look at this page for informations"




March 2017
« Oct