Going through python for the interface

Author: Tristramg

README

@@ -4,19 +4,20 @@ adapted for routing applications.
 == INPUT FORMAT ==
 The input is an OpenStreetMap XML file. The file can be read:
 * from a plain .osm file
-* from a bz2 file
-* from stdin
+* from a bzip2 file
+* from a gzip file
 
 == OUTPUT FORMAT ==
 The output can be:
 * a csv file
-* postgis database (geographical extension to postgres).
+* database (postgres, mysql, sqlite, postgis)
 
 In both output you'll get two files/tables:
 * nodes that contain
     * id (64 bit integer)
     * longitude (decimal real)
     * latitude (decimal real)
+    * geometry (only in postgis)
 
 * edges that contain
     * id (64 bit integer)
@@ -58,26 +59,39 @@ The integers mean:
 
 == INSTALL ==
 You need:
-* the GCC C++ complier (g++)
 * Boost
 * expat (XML parser)
-* libbz2 (Bzip2)
-* libpq (postgres)
+* python
+* swig
 
-Run make and it should compile.
+# Create a virtual environment and activate it
+python virtualenv.py env
+source env/bin/activate
+
+# Get python dependencies and build the application
+python setup.py develop
+
+# Run it
+python parse.py --help
 
 
 == USAGE == 
 Get the .osm XML file of the region that interests you.
 For limited regions, use the export tools from the web interface.
 For bigger regions you might find what you want at http://download.geofabrik.de/osm/
+Osmosis can help you to have a smaller region from a big dump http://wiki.openstreetmap.org/wiki/Osmosis
 
 To know the options, run:
-./osm4routing --help
+python parse.py --help
 
+== PERFORMANCE ==
+OSM data can get very big and can be very consuming, don't try to parse the whole world ;)
+On my laptop from 2006 (core2duo 1.66Ghz, 2Gb Ram, slow hard drive),
+it takes 20 minutes to parse 8Gb uncompressed (0.5Gb as bzip2) and represents France in June 2010
 
-== Postgres output ==
-The database must be a postgis database.
+== Postgis output ==
+Only if you want to use the spatial abilities of postgis, please read those extra informations
+in order to a have spatial database
 The usual way to get is to execute the following commands (the location of
 lwpostgis.sql and spatial_ref_sys.sql depend on your installation).