Export OSMNames¶
When exporting OSMNames the output files get created. This documents describes the implementation of the export. Details about the output format can be found in the introduction.
create functions¶
This step creates the SQL functions later used for the export.
Besides the following descriptions of the functions are the unit tests of Python a good entry point to understand how the functions work.
determine_class¶
Returns a class for a given type. For example, the type city leads to the class place.
The full mapping can be found in the code.
get_parent_info¶
This function makes use of the hierarchy and the place rank to return the following information for an element:
- city
- county
- state
- country_code
- display name
Whereas the display name is a concatenation of the name of the element and all other information.
Note
More information about the impelementation of the function can be found in the PR #82
Example
These elements exists:
| Type | ID | Name | Parent ID |
|---|---|---|---|
| Linestring | 1 | Oberseestrasse | 2 |
| Polygon | 2 | Rapperswil-Jona | 3 |
| Polygon | 3 | Wahlkreis See-Gaster | 4 |
| Polygon | 4 | Sankt Gallen | 5 |
| Polygon | 5 | Schweiz | - |
When calling the function get_parent_info with the parent id and the name of linestring Oberseestrasse following information will be returned:
| Attribute | Value |
|---|---|
| city | Rapperswil-Jona |
| county | Wahlkreis See-Gaster |
| state | Sankt Gallen |
| country_code | ch |
| display name | Oberseestrasse, Rapperswil-Jona, Wahlkreis See-Gaster, Sankt Gallen, Switzerland |
Note
The decision which polygon is the city, county or state is based on the corresponding place rank.
get_country_name¶
Returns the name of a country for a given country code. The name will be returned in the first language present, following the precedence: [English -> native name -> French -> German -> Spanish -> Russian -> Chinese].
The names are read from the helper table country_name (see Import Helper Tables).
get_importance¶
This function returns an importance for an element by its URL to a wikipedia article if present or its place rank.
If a feature has a wikipedia URL a matching entry in the wikipedia helper table is taken for calculating the importance with the following formula:
importance = log (totalcount) / log( max(totalcount))
where totalcount is the number of references to the article from other wikipedia articles. In case there is no wikipedia information or no match was found, the following formula is applied:
importance = 0.75 - (place_rank/40)
Since every feature has a rank, it is ensured that every feature also has an importance.
get_country_language_code¶
Returns the default language for a country. The value is read from the helper table country_name (see Import Helper Tables).
get_housenumbers¶
Returns a comma separated string of all house numbers, associated to the given osm_id.
get_bounding_box¶
This functions takes a geometry, a country code and an admin_level as attribute and determines a bounding box. It is only used for polygons to handle these special cases:
- Some countries do have colonies where are big bounding box is returned. Since this is inconvenient from a user perspective, a smaller bounding box, only covering the main country is returned. See Issue #57 for more details.
- When a polygons intersects the antimeridian, a unintuitive bounding box is returned. In this case the bounding box is shifted manually. See Issue #94 for more details.
create views¶
This function creates the views, which are later used to export the geonames and house numbers. The columns of the views equals the output format of OSMNames.
export geonames¶
This function exports all rows of the polygon, linestring and point view to the file <import-file-name>_geonames.tsv. This by making use of the PostgreSQL function COPY.
export housenumbers¶
This function exports all rows of house number view to the file <import-file-name>_housenumbers.tsv. This by making use of the PostgreSQL function COPY.
Note
House numbers unable to associated to a street or place when preparing the data, are not exported.