A simple library for parsing OSM data. Supports simple OSM XML files as well as OSM GZ, OSM BZ2 and OSM PBF.
Please be aware that osmiter uses Google's protobuf library, written in pure Python, which isn't particularly fast.
import osmiter
shop_count = 0
for feature in osmiter.iter_from_osm("some_osm_file.osm"):
if feature["type"] == "node" and "shop" in feature["tag"]:
shop_count += 1
print(f"this osm file containes {shop_count} shop nodes")
For each feature (node/way/relation) it yields a dict containing element attributes
(like id
, lat
or timestamp
) and 2 additional items: key "type"
holding "node"/"way"/"relation"
and key "tag"
holding a dict with feature tags (this dict may be empty).
Additionally nodes will contain keys "lat"
and "lon"
with node coordinates,
ways will contain key "nd"
with a list of all node_ids references by this way,
and relations contain a key "member"
with a list of dicts of each member's attributes.
Almost all attributes are returned as strings with the exception for:
id
,ref
,version
,changeset
,uid
andchangeset_count
→ intlat
,lon
→ floatopen
andvisible
→ booltimestamp
→ aware datetime.datetime item.
osmiter preforms almost no data validation, so it is possible to recieve ways with no nodes, relations with no members, empty tag values, invalid coordinates, references to non-existing items, or duplicate ids※.
However, several data assumptions are made:
- Each feature has an
id
attribute.
(※) For OSM PBF files, if an object is missing an id-1
will be assigned, per the osmformat.proto definition. This can result in multiple objects with an id equal to-1
. - Each node has to have both
lat
andlon
defined. - Every attribute defined in the table on attribute type conversion has to be convertible to its type.
So,id == 0x1453
,changeset_count == AAAAAA
,ref == 12.433
orlat == 1.23E+10
will cause an exception;
timestamp
value has to be either ISO8601-compliant or epoch time represented by an integer. - Boolean atributes are only considered truthy if they're set to
true
(case-insensitive). Values1
,on
,yes
,TRUE
will all evaluate to False.
Bare-minimum node:
{
"id": int,
"type": "node",
"lat": float,
"lon": float,
"tag": Dict[str, str], # May be empty
}
Bare-minimum way:
{
"id": int,
"type": "way",
"tag": Dict[str, str], # May be empty
"nd": List[int],
}
Bare-minimum relation:
{
"id": int,
"type": "relation",
"tag": Dict[str, str], # May be empty
"member": List[ dict ]
}
See the corresponding OSM XML examples.
{
"type": "node",
"tag": {}
"id": 298887269,
"lat": 54.0901746,
"lon": 12.2482632,
"user": "SvenHRO",
"uid": 46882,
"visible": True,
"version": 1,
"changeset": 676636,
"timestamp": datetime.datetime(2008, 9, 21, 21, 37, 45, tzinfo=datetime.timezone.utc)
}
{
"type": "node",
"tag": {"name": "Neu Broderstorf", "traffic_sign": "city_limit"},
"id": 1831881213,
"version": 1,
"changeset": 12370172,
"lat": 54.0900666,
"lon": 12.2539381,
"user": "lafkor",
"uid": 75625,
"visible": True,
"timestamp": datetime.datetime(2012, 7, 20, 9, 43, 19, tzinfo=datetime.timezone.utc),
}
{
"type": "way",
"tag": {"highway": "unclassified", "name": "Pastower Straße"},
"id": 26659127,
"user": "Masch",
"uid": 55988,
"visible": True,
"version": 5,
"changeset": 4142606,
"timestamp": datetime.datetime(2010, 3, 16, 11, 47, 8, tzinfo=datetime.timezone.utc),
"nd": [292403538, 298884289, 261728686]
}
{
"type": "relation",
"tag": {
"name": "Küstenbus Linie 123",
"network": "VVW",
"operator": "Regionalverkehr Küste",
"ref": "123",
"route": "bus",
"type": "route"
},
"id": 56688,
"user": "kmvar",
"uid": 56190,
"visible": True,
"version": 28,
"changeset": 6947637,
"timestamp": datetime.datetime(2011, 1, 12, 14, 23, 49, tzinfo=datetime.timezone.utc),
"member": [
{"type": "node", "ref": 294942404, "role": ""},
{"type": "node", "ref": 364933006, "role": ""},
{"type": "way", "ref": 4579143, "role": ""},
{"type": "node", "ref": 249673494, "role": ""},
]
}
iter_from_osm(
source: Union[str, bytes, os.PathLike, int, IO[bytes]],
file_format: Union[str, NoneType] = None,
filter_attrs: Union[Iterable[str], NoneType] = None) -> Iterator[dict]
Yields all items from provided source file.
If source is a str/bytes/os.PathLike (path) the format will be guess based on file extension.
Otherwise, if source is an int (file descriptior) or a file-like object,
the file_format
argument must be provided.
File-like sources have to be opened in binary mode. Format has to be one of "xml", "gz", "bz2", "pbf".
osmiter spends most of its time parsing element attributes. If only specific attributes are going to be used, pass an Iterable (most prefereably a set) with wanted attributes to filter_attrs.
No matter what attributes you define in filter_attrs, some attributes are always parsed:
- "id", "lat" and "lon": for nodes
- "id": for ways and relations
- "type", "ref" and "role": for members
filter_attrs
is ignored for pbf files.
iter_from_xml_buffer(
buff: IO[bytes],
filter_attrs: Union[Iterable[str], NoneType] = None) -> Iterator[dict]
Yields all items inside a given OSM XML buffer.
filter_attrs
is explained in osmiter.iter_from_osm documentation.
iter_from_pbf_buffer(buff: IO[bytes]) -> Iterator[dict]
Yields all items inside a given OSM PBF buffer.
Same as osmiter.iter_from_xml_buffer
.
An exception (inheriting from RuntimeException
) used to represent issues with XML data.
iter_from_pbf_buffer(buff: BinaryIO) -> Iterator[dict]
Yields all items inside a given OSM PBF file.
Same as osmiter.iter_from_pbf_buffer
.
Internal object used to parse PBF files. Don't use.
An Exception (inheriting from RuntimeException
) used to represent issues with OSM PBF files.
osmiter is provided under the MIT license, included in the license.md
file.