[test] Customized Taginfo : 620 areas - every country and some new experimental features ( ~ 1 week test )

classic Classic list List threaded Threaded
3 messages Options
Reply | Threaded
Open this post in threaded view
|

[test] Customized Taginfo : 620 areas - every country and some new experimental features ( ~ 1 week test )

Imre Samu
This is a Proof of Concept of my vision [ customizing taginfo for countries, regions ] 
in my experience - It can be useful for finding local tagging errors.


dev site:         http://taginfo-dev.opengeodata.hu  find your area/country
1 week test:   shutdown time:  ~ 2018-aug-20 ( GMT 23:00h )


Main changes:

-  620 areas  - not refreshing 
      = 620 docker services running in a simple cloud machine.
                    32Gb RAM,   slow CPU :  Intel(R) Atom(TM) CPU  C2750  @ 2.40GHz,   8 core ,  ~ 600Gb Disk )

          
-  2 new experimental reports:

      "QA-Normalized name differences (Experimental)"
            example:  http://eu-at.taginfo-dev.opengeodata.hu/reports/normalized_names
            The result can be download as an xlsx file:  http://eu-at.taginfo-dev.opengeodata.hu/download/normalized_names.xlsx

            ( I hope - this will be useful for the localized https://github.com/osmlab/name-suggestion-index 
                     ( see https://github.com/osmlab/name-suggestion-index/issues/11 )   


      "QA-Problematic tags (Experimental)"     [ still a lot of bugs,   for example:  checking access type of tags  is not perfect yet, sorry ]  

- `name` support for tags  ( Experimental )  

     examples:
              
     Spain   amenity=place_of_worship     ( names in Spanish  + Català (ca), Galego (ga) and Euskera (eu) )


       or Switzerland   amenity=bank
  
      the "name:*" tags configured on the https://wiki.openstreetmap.org/wiki/Multilingual_names info but not perfect yet.

      so "Belgium has three official languages (Dutch, French and German) ...."    so the amenity=pub  names is:
    
     or in Ireland (Republic )  -  amenity=pub  names  in a Gaeltacht (name:ga)  : http://eu-ie.taginfo-dev.opengeodata.hu/tags/amenity=pub#tagnames_lang3      

   
      It is interesting for checking in your area the tag"names" of    *=yes  tags -  some of them easy to fix.
       * shop=yes  ;  amenity=yes   ;    man_made=yes ; natural=yes ; sport=yes  ; leisure=yes

          shop=yes       in the UK ( http://eu-gb.taginfo-dev.opengeodata.hu/tags/shop=yes#tagnames_lang1 )


Housenumbers ...

 It is interesting for me - checking the frequent    addr:housenumber   values.

 In Taiwan ( http://as-tw.taginfo-dev.opengeodata.hu/keys/addr%3Ahousenumber#values )    not so much  number "4"   
 compare to european countries ...   (  hint:  https://en.wikipedia.org/wiki/Tetraphobia )
 
 in Switzerland - check the 13   :)   ( https://en.wikipedia.org/wiki/Triskaidekaphobia )

The South America - has a different frequent housenumbers - compare to Europe:
Peru :           TOP3   100,200,199 :  http://sa-pe.taginfo-dev.opengeodata.hu/keys/addr%3Ahousenumber#values
Brazil:           TOP3   100,50,35 :     http://sa-br.taginfo-dev.opengeodata.hu/keys/addr%3Ahousenumber#values
Chile :           TOP3   500,600,401  http://sa-cl.taginfo-dev.opengeodata.hu/keys/addr:housenumber#values

in Europe:

The Vatican is the best.     only 1 housenumber:   http://eu-va.taginfo-dev.opengeodata.hu/keys/addr%3Ahousenumber#values


----- work in progress ...
source code - for the dockerization:       https://github.com/taginfo/dockerized-taginfo  ( issues )
    This is an experimental - and not official changes,   probably some changes will be converted as a "taginfo plug-in".
-----  

Regards,
   Imre

// ImreSamu  ,  member of the OSM  Hungary





_______________________________________________
talk mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/talk
Reply | Threaded
Open this post in threaded view
|

Re: [test] Customized Taginfo : 620 areas - every country and some new experimental features ( ~ 1 week test )

Frédéric Rodrigo-2
Hello,

It's an impressive job. It can really help.
Just a note about France, you split France on admin_level=6, this result
on around 100 pieces, it does not make sens. For France admin_level=4 is
the only right level of sub area.

Like for France or other countries, sub levels is interesting but the
country as a whole also..

Frédéric.


Le 14/08/2018 à 20:26, Imre Samu a écrit :

> This is a Proof of Concept of my vision [ customizing taginfo for
> countries, regions ]
> in my experience - It can be useful for finding local tagging errors.
>
>
> dev site: http://taginfo-dev.opengeodata.hu 
> <http://taginfo-dev.opengeodata.hu/> find your area/country
> 1 week test:   shutdown time: *~ 2018-aug-20 ( GMT 23:00h )*
>
>
> Main changes:
>
> *-  620 areas  - not refreshing *
>       = 620 docker services running in a simple cloud machine.
>                     32Gb RAM,   slow CPU :  Intel(R) Atom(TM) CPU 
> C2750  @ 2.40GHz,   8 core ,  ~ 600Gb Disk )
>
> *-  2 new experimental reports*:
>
>       "QA-Normalized name differences (Experimental)"
>             example:
> http://eu-at.taginfo-dev.opengeodata.hu/reports/normalized_names
>             The result can be download as an xlsx file:
> http://eu-at.taginfo-dev.opengeodata.hu/download/normalized_names.xlsx
>
>             ( I hope - this will be useful for the localized
> https://github.com/osmlab/name-suggestion-index
>                      ( see
> https://github.com/osmlab/name-suggestion-index/issues/11 )
>
>
>       "QA-Problematic tags (Experimental)"     [ still a lot of bugs, 
>  for example:  checking access type of tags  is not perfect yet, sorry ]
>              example:
> http://eu-at.taginfo-dev.opengeodata.hu/reports/problematic_tags
>              .xlsx result:
> http://eu-at.taginfo-dev.opengeodata.hu/download/problematic_tags.xlsx
>
> *- `name` support for tags  ( Experimental ) *
>
>      examples:
>      Spain   amenity=place_of_worship     ( names in Spanish  + Català
> (ca), Galego (ga) and Euskera (eu) )
>          name       =
> http://eu-es.taginfo-dev.opengeodata.hu/tags/amenity=place_of_worship#tagnames_lang1
>          name:es  =
> http://eu-es.taginfo-dev.opengeodata.hu/tags/amenity=place_of_worship#tagnames_lang2
>          name:eu  =
> http://eu-es.taginfo-dev.opengeodata.hu/tags/amenity=place_of_worship#tagnames_lang3
>          name:ca  =
> http://eu-es.taginfo-dev.opengeodata.hu/tags/amenity=place_of_worship#tagnames_lang4
>          name:gl   =
> http://eu-es.taginfo-dev.opengeodata.hu/tags/amenity=place_of_worship#tagnames_lang5
>
>
>        or Switzerland   amenity=bank
>          name       =
> http://eu-ch.taginfo-dev.opengeodata.hu/tags/amenity=bank#tagnames_lang1
>          name:en  =
> http://eu-ch.taginfo-dev.opengeodata.hu/tags/amenity=bank#tagnames_lang2
>          name:de  =
> http://eu-ch.taginfo-dev.opengeodata.hu/tags/amenity=bank#tagnames_lang3
>          name:fr    =
> http://eu-ch.taginfo-dev.opengeodata.hu/tags/amenity=bank#tagnames_lang4
>          name:it    =
> http://eu-ch.taginfo-dev.opengeodata.hu/tags/amenity=bank#tagnames_lang5
>       the "name:*" tags configured on the
> https://wiki.openstreetmap.org/wiki/Multilingual_names info but not
> perfect yet.
>
>       so "Belgium has three official languages (Dutch, French and
> German) ...."    so the *amenity=pub * names is:
>        name     =
> http://eu-be.taginfo-dev.opengeodata.hu/tags/amenity=pub#tagnames_lang1
>        name:fr  =
> http://eu-be.taginfo-dev.opengeodata.hu/tags/amenity=pub#tagnames_lang2
>        name:nl =
> http://eu-be.taginfo-dev.opengeodata.hu/tags/amenity=pub#tagnames_lang3
>        name:en =
> http://eu-be.taginfo-dev.opengeodata.hu/tags/amenity=pub#tagnames_lang4
>        name:de =
> http://eu-be.taginfo-dev.opengeodata.hu/tags/amenity=pub#tagnames_lang5
>      or in Ireland (Republic )  -  amenity=pub  names  in a Gaeltacht
> (name:ga)  :
> http://eu-ie.taginfo-dev.opengeodata.hu/tags/amenity=pub#tagnames_lang3
>
>       It is interesting for checking in your area the tag"names" of   
> *=yes  tags -  some of them easy to fix.
>        * shop=yes  ;  amenity=yes   ;    man_made=yes ; natural=yes
> ; sport=yes  ; leisure=yes
>
>           amenity=yes  in the UK (
> http://eu-gb.taginfo-dev.opengeodata.hu/tags/amenity=yes#tagnames_lang1 )
>           shop=yes       in the UK (
> http://eu-gb.taginfo-dev.opengeodata.hu/tags/shop=yes#tagnames_lang1 )
>
> *
> *
> *Housenumbers ...*
>
>  It is interesting for me - checking the frequent
> * addr**:**housenumber *  values.
>
>  In Taiwan (
> http://as-tw.taginfo-dev.opengeodata.hu/keys/addr%3Ahousenumber#values 
> )    not so much  number "4"
>  compare to european countries ...   (  hint:
> https://en.wikipedia.org/wiki/Tetraphobia )
>  in Switzerland - check the 13   :)   (
> https://en.wikipedia.org/wiki/Triskaidekaphobia )
> http://eu-ch.taginfo-dev.opengeodata.hu/keys/addr%3Ahousenumber#values
>
> The South America - has a different frequent housenumbers - compare to
> Europe:
>
>     Argentina  :  TOP3   199,201,200 :
>     http://sa-ar.taginfo-dev.opengeodata.hu/keys/addr%3Ahousenumber#values
>     Peru :           TOP3   100,200,199 :
>     http://sa-pe.taginfo-dev.opengeodata.hu/keys/addr%3Ahousenumber#values
>     Brazil:           TOP3   100,50,35 :
>     http://sa-br.taginfo-dev.opengeodata.hu/keys/addr%3Ahousenumber#values
>     Chile :           TOP3   500,600,401
>     http://sa-cl.taginfo-dev.opengeodata.hu/keys/addr:housenumber#values
>
>     in Europe:
>     Denmark    TOP3   4,3,5    :
>     http://eu-dk.taginfo-dev.opengeodata.hu/keys/addr%3Ahousenumber#values
>     Austria        TOP3  1,3,2    :
>     http://eu-at.taginfo-dev.opengeodata.hu/keys/addr:housenumber#values
>     Estonia      TOP3    4,3,1   :
>     http://eu-ee.taginfo-dev.opengeodata.hu/keys/addr:housenumber#values
>     Finland      TOP3    3,4,1   :
>     http://eu-fi.taginfo-dev.opengeodata.hu/keys/addr:housenumber#values
>
>
>     The Vatican is the best.     only 1 housenumber:
>     http://eu-va.taginfo-dev.opengeodata.hu/keys/addr%3Ahousenumber#values
>     and no pubs :
>     http://eu-va.taginfo-dev.opengeodata.hu/keys/amenity#values
>
>
>
> ----- work in progress ...
> source code - for the dockerization:
> https://github.com/taginfo/dockerized-taginfo ( issues )
> taginfo changes:
> https://github.com/taginfo/taginfo/compare/master...ImreSamu:name_tabs_v2
>     This is an experimental - and not official changes,  probably some
> changes will be converted as a "taginfo plug-in".
> -----
>
> Regards,
>    Imre
>
> // ImreSamu  ,  member of the OSM Hungary
>
>
>
>
>
> _______________________________________________
> talk mailing list
> [hidden email]
> https://lists.openstreetmap.org/listinfo/talk



_______________________________________________
talk mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/talk
Reply | Threaded
Open this post in threaded view
|

Re: [test] Customized Taginfo : 620 areas - every country and some new experimental features ( ~ 1 week test )

Imre Samu
Hi Frédéric,

Thank you for the feedback.
I will recheck the big countries config files  ( France,  Russia, Canada, Germany, U.S, Japan ) in the next version.

background:  I have started the development with 8GB RAM - and this is too low for processing some countries.   

Thanks,
  Imre


Frédéric Rodrigo <[hidden email]> ezt írta (időpont: 2018. aug. 16., Cs, 15:24):
Hello,

It's an impressive job. It can really help.
Just a note about France, you split France on admin_level=6, this result
on around 100 pieces, it does not make sens. For France admin_level=4 is
the only right level of sub area.

Like for France or other countries, sub levels is interesting but the
country as a whole also..

Frédéric.


Le 14/08/2018 à 20:26, Imre Samu a écrit :
> This is a Proof of Concept of my vision [ customizing taginfo for
> countries, regions ]
> in my experience - It can be useful for finding local tagging errors.
>
>
> dev site: http://taginfo-dev.opengeodata.hu
> <http://taginfo-dev.opengeodata.hu/> find your area/country
> 1 week test:   shutdown time: *~ 2018-aug-20 ( GMT 23:00h )*
>
>
> Main changes:
>
> *-  620 areas  - not refreshing *
>       = 620 docker services running in a simple cloud machine.
>                     32Gb RAM,   slow CPU :  Intel(R) Atom(TM) CPU 
> C2750  @ 2.40GHz,   8 core ,  ~ 600Gb Disk )
>
> *-  2 new experimental reports*:
>
>       "QA-Normalized name differences (Experimental)"
>             example:
> http://eu-at.taginfo-dev.opengeodata.hu/reports/normalized_names
>             The result can be download as an xlsx file:
> http://eu-at.taginfo-dev.opengeodata.hu/download/normalized_names.xlsx
>
>             ( I hope - this will be useful for the localized
> https://github.com/osmlab/name-suggestion-index
>                      ( see
> https://github.com/osmlab/name-suggestion-index/issues/11 )
>
>
>       "QA-Problematic tags (Experimental)"     [ still a lot of bugs, 
>  for example:  checking access type of tags  is not perfect yet, sorry ]
>              example:
> http://eu-at.taginfo-dev.opengeodata.hu/reports/problematic_tags
>              .xlsx result:
> http://eu-at.taginfo-dev.opengeodata.hu/download/problematic_tags.xlsx
>
> *- `name` support for tags  ( Experimental ) *
>
>      examples:
>      Spain   amenity=place_of_worship     ( names in Spanish  + Català
> (ca), Galego (ga) and Euskera (eu) )
>          name       =
> http://eu-es.taginfo-dev.opengeodata.hu/tags/amenity=place_of_worship#tagnames_lang1
>          name:es  =
> http://eu-es.taginfo-dev.opengeodata.hu/tags/amenity=place_of_worship#tagnames_lang2
>          name:eu  =
> http://eu-es.taginfo-dev.opengeodata.hu/tags/amenity=place_of_worship#tagnames_lang3
>          name:ca  =
> http://eu-es.taginfo-dev.opengeodata.hu/tags/amenity=place_of_worship#tagnames_lang4
>          name:gl   =
> http://eu-es.taginfo-dev.opengeodata.hu/tags/amenity=place_of_worship#tagnames_lang5
>
>
>        or Switzerland   amenity=bank
>          name       =
> http://eu-ch.taginfo-dev.opengeodata.hu/tags/amenity=bank#tagnames_lang1
>          name:en  =
> http://eu-ch.taginfo-dev.opengeodata.hu/tags/amenity=bank#tagnames_lang2
>          name:de  =
> http://eu-ch.taginfo-dev.opengeodata.hu/tags/amenity=bank#tagnames_lang3
>          name:fr    =
> http://eu-ch.taginfo-dev.opengeodata.hu/tags/amenity=bank#tagnames_lang4
>          name:it    =
> http://eu-ch.taginfo-dev.opengeodata.hu/tags/amenity=bank#tagnames_lang5
>       the "name:*" tags configured on the
> https://wiki.openstreetmap.org/wiki/Multilingual_names info but not
> perfect yet.
>
>       so "Belgium has three official languages (Dutch, French and
> German) ...."    so the *amenity=pub * names is:
>        name     =
> http://eu-be.taginfo-dev.opengeodata.hu/tags/amenity=pub#tagnames_lang1
>        name:fr  =
> http://eu-be.taginfo-dev.opengeodata.hu/tags/amenity=pub#tagnames_lang2
>        name:nl =
> http://eu-be.taginfo-dev.opengeodata.hu/tags/amenity=pub#tagnames_lang3
>        name:en =
> http://eu-be.taginfo-dev.opengeodata.hu/tags/amenity=pub#tagnames_lang4
>        name:de =
> http://eu-be.taginfo-dev.opengeodata.hu/tags/amenity=pub#tagnames_lang5
>      or in Ireland (Republic )  -  amenity=pub  names  in a Gaeltacht
> (name:ga)  :
> http://eu-ie.taginfo-dev.opengeodata.hu/tags/amenity=pub#tagnames_lang3
>
>       It is interesting for checking in your area the tag"names" of   
> *=yes  tags -  some of them easy to fix.
>        * shop=yes  ;  amenity=yes   ;    man_made=yes ; natural=yes
> ; sport=yes  ; leisure=yes
>
>           amenity=yes  in the UK (
> http://eu-gb.taginfo-dev.opengeodata.hu/tags/amenity=yes#tagnames_lang1 )
>           shop=yes       in the UK (
> http://eu-gb.taginfo-dev.opengeodata.hu/tags/shop=yes#tagnames_lang1 )
>
> *
> *
> *Housenumbers ...*
>
>  It is interesting for me - checking the frequent
> * addr**:**housenumber *  values.
>
>  In Taiwan (
> http://as-tw.taginfo-dev.opengeodata.hu/keys/addr%3Ahousenumber#values
> )    not so much  number "4"
>  compare to european countries ...   (  hint:
> https://en.wikipedia.org/wiki/Tetraphobia )
>  in Switzerland - check the 13   :)   (
> https://en.wikipedia.org/wiki/Triskaidekaphobia )
> http://eu-ch.taginfo-dev.opengeodata.hu/keys/addr%3Ahousenumber#values
>
> The South America - has a different frequent housenumbers - compare to
> Europe:
>
>     Argentina  :  TOP3   199,201,200 :
>     http://sa-ar.taginfo-dev.opengeodata.hu/keys/addr%3Ahousenumber#values
>     Peru :           TOP3   100,200,199 :
>     http://sa-pe.taginfo-dev.opengeodata.hu/keys/addr%3Ahousenumber#values
>     Brazil:           TOP3   100,50,35 :
>     http://sa-br.taginfo-dev.opengeodata.hu/keys/addr%3Ahousenumber#values
>     Chile :           TOP3   500,600,401
>     http://sa-cl.taginfo-dev.opengeodata.hu/keys/addr:housenumber#values
>
>     in Europe:
>     Denmark    TOP3   4,3,5    :
>     http://eu-dk.taginfo-dev.opengeodata.hu/keys/addr%3Ahousenumber#values
>     Austria        TOP3  1,3,2    :
>     http://eu-at.taginfo-dev.opengeodata.hu/keys/addr:housenumber#values
>     Estonia      TOP3    4,3,1   :
>     http://eu-ee.taginfo-dev.opengeodata.hu/keys/addr:housenumber#values
>     Finland      TOP3    3,4,1   :
>     http://eu-fi.taginfo-dev.opengeodata.hu/keys/addr:housenumber#values
>
>
>     The Vatican is the best.     only 1 housenumber:
>     http://eu-va.taginfo-dev.opengeodata.hu/keys/addr%3Ahousenumber#values
>     and no pubs :
>     http://eu-va.taginfo-dev.opengeodata.hu/keys/amenity#values
>
>
>
> ----- work in progress ...
> source code - for the dockerization:
> https://github.com/taginfo/dockerized-taginfo ( issues )
> taginfo changes:
> https://github.com/taginfo/taginfo/compare/master...ImreSamu:name_tabs_v2
>     This is an experimental - and not official changes,  probably some
> changes will be converted as a "taginfo plug-in".
> -----
>
> Regards,
>    Imre
>
> // ImreSamu  ,  member of the OSM Hungary
>
>
>
>
>
> _______________________________________________
> talk mailing list
> [hidden email]
> https://lists.openstreetmap.org/listinfo/talk



_______________________________________________
talk mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/talk