Fresh, clean, patch against the ocitysmap repo for simplifying the
Turkish translation. This patch contains only the language manipulation.
Signed-off-by: Maxime Petazzoni <maxime.petazzoni@bulix.org>
Cleanup some of the variable definitions, add -C 5000 by default to the
osm2pgsql process to benefit from better cache hit rate, and fix logging
and output of the maposmatic_admin table update.
Signed-off-by: Maxime Petazzoni <maxime.petazzoni@bulix.org>
Renamed ocitysmap2-render as render.py since we don't need/want
"ocitysmap2" in the name anymore. Rename conf template as
ocitysmap.conf.dist for the same reason. Updated INSTALL instructions
accordingly.
Signed-off-by: Maxime Petazzoni <maxime.petazzoni@bulix.org>
In ocitysmap2/__init__.py:_get_geographic_info(), when one is looking in
the database, the returned "records" can be "[(None,)]". In that case
wkt obtains value None, which breaks the "shapely.wkt.loads(wkt)" call.
We add a protection against that.
Signed-off-by: David MENTRE <dmentre@linux-france.org>
Due to the order of spanish determinants, a street such as "Calle de
los Huertos" was indexed under "los Huertos (Calle de)" instead of
"Huertos (Calle de los)". Same problem for "de la" or "de las". By
reordering the determinants so that the most specific ones are
considered first, we can achieve the desired behaviour.
See
http://www.maposmatic.org/rendered//046017_2012-04-27_11-51_MirafloresdelaSierraSpain.png
for an example of many current problems.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@enix.org>
As the changes in the base stylesheet begins to be heavy, specific
stylesheets files have been created. Reference files (labels.mms for
labels-ocitysmap.mss and style-toner-standard.mml for
style-toner-ocitysmap.mml) have been added to ocitysmap repository to
track changes on the toner main repository.
style-toner-ocitysmap.mml and labels-ocitysmap.mss have been updated
to display names of localities, hamlets and villages.
INSTALL instructions have been updated.
In concerned SQL queries this patch apply a 0-buffer to malformed ways.
As malformed ways seem pretty uncommon and as this operation is heavy
for performance issue the query is first run without the 0-buffer and
is re-run with it if the query fails.
With PostGIS 1.5 the failure message is not explicit so we can only
catch the generic psycopg2.InternalError.
Hakan Tandogan
