new python GPS wrapper (liblocation.py)
 

I'm attaching my python code to access the GPS. It is a thin wrapper around the Maemo 4.x liblocation API described here:
http://maemo.org/development/documen....html#Location

I am running an n810 with the latest OS 2008 firmware and was not able to make the python-gpsbt module work as advertised. I decided to wrap the official location API using ctypes instead. The example usage script should "just work" to turn on your GPS and start printing location data once it locks.

This code "works for me", but should be considered alpha since I am the only one who has used it. Eventually I'll try releasing this someplace better. The example is placed in the public domain and the wrapper is licensed under LGPL v3.

The files should be renamed to *.py after downloading.

Re: new python GPS wrapper (liblocation.py)
 

Thank you so much for your python wrapper. It's very useful for me.
I've detected a really weird error:
When I load your liblocation from console applications or graphic applications everything works well. But if I try to load it from some "hildon desktop home plugin" (aka widget) I get the error:
File "/usr/lib/hildon-desktop/gps_N800.py", line 1, in <module>
import liblocation
ImportError: dynamic module does not define init function (initliblocation)

I don't have any Idea about how to fix it. It doesn't have any sense¡

Re: new python GPS wrapper (liblocation.py)
 

I reply myself:
Adding
It works.
I should read better the error messages. Sometimes are useful.

Re: new python GPS wrapper (liblocation.py)
 

Any thoughts on getting this working for N900? I'm trying to build a cool location-based social network leveaging an existing site's API, but I'm not finding any information on how to connect to the built in location based services of the N900 (either through internal GPS or the cell-triangulation?). I tried using your examples but no dice.

Much appreciated.

Re: new python GPS wrapper (liblocation.py)
 

@hypnotik, are you focused on Python, or can you use c/c++ ?

Re: new python GPS wrapper (liblocation.py)
 

Well, considering I'm coming from a web-dev background and that I know Python, yeah, I don't have the wherewithal to learn c/c++ (gives me a headache - pointers, memory management and what not).

I don't understand why there isn't something out there already as an example, esp if there is this push for PyQt to get rapid development on the N900.

I've already started building a PyQt based app for the N900... but not being able to read gps info from the phone is a critical stumbing block.

Re: new python GPS wrapper (liblocation.py)
 

The GPS module is being worked by the Pymaemo guys. They have something in the SVN and they will be pushing a package to the repository soon enough.

http://wiki.maemo.org/PyMaemo/Using_Location_API
https://bugs.maemo.org/show_bug.cgi?id=5748#c2

Re: new python GPS wrapper (liblocation.py)
 

I asked, because we have a beta sensors library for c++. For GPS you just create an object of a receiver and then you have access to the coordinates. But yeah, c++ is not as good for development as python (although not so bad as you could imagine, especially with Qt).

Re: new python GPS wrapper (liblocation.py)
 

hopbeat: Where can I find out more about the C++ library?

Re: new python GPS wrapper (liblocation.py)
 

@sjaensch, PM me, as it is not yet ready for public tests :)

Re: new python GPS wrapper (liblocation.py)
 

Pity that liblocation API must have changed. I went with that since it seemed to be the most "officially documented" way of accessing the GPS in maemo and therefore the least likely to change.

I not going to attempt to migrate the code because I don't have an n900 device to test with. If someone else would like to try it, it should involve making the python ctypes structures and function calls line up with whatever is now changed in the C API. This would involve chasing through the new liblocation docs and header files. The python wrapper is very thin so there's not a whole lot that can be going wrong, unless the new API is radically different.

Re: new python GPS wrapper (liblocation.py)
 

I finally had a chance to look at this, and it seems there are some problems:

using the example code:
it runs, however, the result returns immediately with:

No matter what I change the preferred_method= to it always returns instantaneously - even if I reboot the phone to force a cold-start.

Lat is 1.4 degrees south of my current position, and Long is zero which isn't right either....

Re: new python GPS wrapper (liblocation.py)
 

Follow up to this (I got it working)

I posted my question to the pymaemo developer list and one of the developers, David, pointed out the problem with the code exiting the first time on_changed() gets called when it hasn't fulling acquired a fix. Commenting out the data.stop() line allows a 2nd call to acquire the full fix:

Also, I had updated my location server from supl.nokia.com to supl.google.com per this post: http://talk.maemo.org/showpost.php?p...3&postcount=14

Allowed me to get a quick network-based fix in the Ovi Maps app indoors. Before using the old server I could never get a fix.

Re: new python GPS wrapper (liblocation.py)
 

Hi
Do you have and example like this but using ctypes?
I need a faster refresh and this wrapper only allows 1 sec or higher

thanks

Re: new python GPS wrapper (liblocation.py)
 

Have you tried playing with the python-location? See http://wiki.maemo.org/PyMaemo/Using_Location_API for an example .

Re: new python GPS wrapper (liblocation.py)
 

lizardo, that is the one i have used. but it only allows a max rate of 1 position refresh per second and i need a position refresh of at least 0.1 of second.
thanks