Osmread is a simple library for reading OpenStreetMap data files in Python. It supports XML and PBF file formats as inputs. It is not designed for fast processing of big planet dumps but can be used for simple processing of smaller files such as regional extracts.
If you need fast processing of large files look at imposm.parser library.
Simply run setup.py or use pip or easy_install.
$ python setup.py install
from osmread import parse_file, Way
highway_count = 0
for entity in parse_file('foo.osm.bz2'):
if isinstance(entity, Way) and 'highway' in entity.tags:
highway_count += 1
print("%d highways found" % highway_count)
All Node, Way and Relation instances have id, version, changeset,
timestamp, uid and tags attributes. Node coordinates stored in lon and
lat. Way nodes stored in nodes, relation members in members.
Relation members are array of tuples (role, class, id), where class is
Node, Way or Relation.
This library uses very slow protobuf implementation - pure python protobuf library. As a result parsing PBF files is slower than XML. This may change in future versions.