Several improvements for your nominatim module

[iburadempa]

• polishing of code and documentation
• add few unittests; run them from the project home with: python -m unittest discover
• changed default Nominatim instance to http://open.mapquestapi.com/nominatim/v1 to avoid overloading of openstreetmap (s. http://wiki.openstreetmap.org/wiki/Nominatim_usage_policy )
• add option osm_id plus osm_type to NominatimReverse
• add zoom aliases (NominatimReverse)
• add countrycodes selection
• add url options to choose Nominatim instance
• add NominatimException
• factor out the HTTP request to NominatimRequest
• fix lat==0 or lon==0 leading to no result
• fix URL encoding

[damianbraun]

thanks for help :)

[iburadempa]

hi damian,

do you have plans for further improvement of the nominatim module?

in my view it would be nice to have functions for extracting useful information from the query results, in particular if there are many of them.

best, ibu

[damianbraun]

Hi, this module works good enough for me. I use it only for my gpxseg https://github.com/damianbraun/gpxseg project.

If you have time and willingness im happy for your improvements. Only thing i will ask you is to commit more often (commit per feature). I consider it good practise and try to do it myself.

About this functions. Do you mean something like getStreet(), getAdress()? Which would return only specific strings or rather unicodes?

How u use this module?

[iburadempa]

On 07/10/2014 05:41 PM, damianbraun wrote:

Hi, this module works good enough for me. I use it only for my gpxseg https://github.com/damianbraun/gpxseg project.

I see, you're using it for reverse geocoding and with data where you can reasonably expect to get a good result.

If you have time and willingness im happy for your improvements. Only thing i will ask you is to commit more often (commit per feature). I consider it good practise and try to do it myself.

Sure, I'll try. - Initially I was adapting it to my own needs only.

About this functions. Do you mean something like getStreet(), getAdress()? Which would return only specific strings or rather unicodes?

How u use this module?

I have short texts containing valid addresses, invalid/misspelled addresses, parts of valid or invalid addresses, or no address information at all. What I want is: extract as much address information as possible while keeping the number of requests small. The general case would perhaps be a probability distribution of locations for each zoom level, or probabilities for (a location to be within) a set of (disjunct) regions, where regions could be continents, countries, quarters of cities, or any other regions on any reasonable scale.

This involves reverse requests, too, because if I have found a location possibly matching the text string, I want to get the correct address at this location to assess text similarity. And I want to know which parts of the text do probably contain which part of the found address.

Most of this will go to nominatim/addressminer.py. I also want to allow for other location information (such as inferred from the language (which is already part of Nominatim, but the module currently only allows for one language, not an iterable, which I'd like to change some time), or from other information provided externally (like by other text miners)).

If you think that doesn't fit into your module, let me know. I also had some doubt, but in order to extract maximal address information, the miner should know the data structures of the geocoder very well.

best, ibu

2014-07-10 12:43 GMT+02:00 iburadempa notifications@github.com:

hi damian,

do you have plans for further improvement of the nominatim module?

in my view it would be nice to have functions for extracting useful information from the query results, in particular if there are many of them.