OSM Online Loader¶

Run this notebook in Google Colab:

Remember to install the srai library before running the notebook:

%pip install srai[all]

In [1]:

Copied!





from functional import seq

from srai.loaders.osm_loaders import OSMOnlineLoader
from srai.loaders.osm_loaders.filters import GEOFABRIK_LAYERS, HEX2VEC_FILTER
from srai.loaders.osm_loaders.filters.popular import get_popular_tags
from srai.plotting.folium_wrapper import plot_regions
from srai.regionalizers import geocode_to_region_gdf
from functional import seq

from srai.loaders.osm_loaders import OSMOnlineLoader
from srai.loaders.osm_loaders.filters import GEOFABRIK_LAYERS, HEX2VEC_FILTER
from srai.loaders.osm_loaders.filters.popular import get_popular_tags
from srai.plotting.folium_wrapper import plot_regions
from srai.regionalizers import geocode_to_region_gdf

Filters¶

Filters are dictionaries used for specifying what type of objects one would like to download from OpenStreetMap.
There is currently one predefined filter (from Hex2Vec paper) and one way to download a filter - using popular tags from taginfo API.
They can also be defined manually in code.

Additionally, few predefined grouped filters are available (eg. BASE_OSM_GROUPS_FILTER and GEOFABRIK_LAYERS). Grouped filters categorize base filters into groups.

Get popular tags from taginfo API¶

In [2]:

Copied!





all_popular_tags = get_popular_tags()
num_keys = len(all_popular_tags)
num_values = seq(all_popular_tags.values()).map(len).sum()
f"Unique keys: {num_keys}. Key/value pairs: {num_values}"
all_popular_tags = get_popular_tags()
num_keys = len(all_popular_tags)
num_values = seq(all_popular_tags.values()).map(len).sum()
f"Unique keys: {num_keys}. Key/value pairs: {num_values}"

Out[2]:

'Unique keys: 1230. Key/value pairs: 12955'

In [3]:

Copied!

seq(all_popular_tags.items()).take(10).dict()
seq(all_popular_tags.items()).take(10).dict()

Out[3]:

{'4wd_only': ['recommended', 'yes'],
 'CLC:code': ['112', '211', '221', '231', '242', '243', '311', '312', '313'],
 'CLC:explanation': ['See http://wiki.openstreetmap.org/wiki/Romania_CLC_Import.'],
 'CLC:year': ['2006'],
 'FMMP_modified': ['no'],
 'FMMP_reviewed': ['no'],
 'GNS:dsg_code': ['PPL', 'WAD'],
 'GNS:dsg_name': ['populated place', 'wadi'],
 'GNS:dsg_string': ['populated place'],
 'GeoBaseNHN:DatasetName': ['02LG000', '02LH000']}

In [4]:

Copied!

frequent_in_wiki_only_tags = get_popular_tags(in_wiki_only=True, min_fraction=0.001)
frequent_in_wiki_only_tags
frequent_in_wiki_only_tags = get_popular_tags(in_wiki_only=True, min_fraction=0.001)
frequent_in_wiki_only_tags

Out[4]:

{'access': ['private'],
 'addr:country': ['DE'],
 'building': ['house', 'residential', 'yes'],
 'highway': ['crossing',
  'footway',
  'path',
  'residential',
  'service',
  'track',
  'unclassified'],
 'landuse': ['farmland'],
 'natural': ['tree', 'water', 'wood'],
 'oneway': ['yes'],
 'power': ['pole', 'tower'],
 'service': ['driveway'],
 'source': ['BAG', 'Bing', 'NRCan-CanVec-10.0'],
 'surface': ['asphalt', 'unpaved'],
 'wall': ['no'],
 'waterway': ['stream']}

Import a predefined filter¶

In [5]:

Copied!

hex_2_vec_keys = len(HEX2VEC_FILTER)
hex_2_vec_key_values = seq(HEX2VEC_FILTER.values()).map(len).sum()
f"Unique keys: {hex_2_vec_keys}. Key/value pairs: {hex_2_vec_key_values}"
hex_2_vec_keys = len(HEX2VEC_FILTER)
hex_2_vec_key_values = seq(HEX2VEC_FILTER.values()).map(len).sum()
f"Unique keys: {hex_2_vec_keys}. Key/value pairs: {hex_2_vec_key_values}"

Out[5]:

'Unique keys: 15. Key/value pairs: 725'

In [6]:

Copied!





geofabrik_layers_keys = len(GEOFABRIK_LAYERS)
geofabrik_layers_key_values = (
    seq(GEOFABRIK_LAYERS.values()).flat_map(lambda filter: filter.items()).map(len).sum()
)
f"Unique groups: {geofabrik_layers_keys}. Key/value pairs: {geofabrik_layers_key_values}"
geofabrik_layers_keys = len(GEOFABRIK_LAYERS)
geofabrik_layers_key_values = (
    seq(GEOFABRIK_LAYERS.values()).flat_map(lambda filter: filter.items()).map(len).sum()
)
f"Unique groups: {geofabrik_layers_keys}. Key/value pairs: {geofabrik_layers_key_values}"

Out[6]:

'Unique groups: 28. Key/value pairs: 116'

Using OSMOnlineLoader to download data for a specific area¶

Download all parks in Wrocław, Poland¶

In [7]:

Copied!





loader = OSMOnlineLoader()
parks_filter = {"leisure": "park"}
wroclaw_gdf = geocode_to_region_gdf("Wrocław, Poland")
parks_gdf = loader.load(wroclaw_gdf, parks_filter)
parks_gdf
loader = OSMOnlineLoader()
parks_filter = {"leisure": "park"}
wroclaw_gdf = geocode_to_region_gdf("Wrocław, Poland")
parks_gdf = loader.load(wroclaw_gdf, parks_filter)
parks_gdf

Out[7]:

	geometry	leisure
feature_id
relation/1348101	POLYGON ((17.07422 51.08476, 17.07431 51.08479...	park
relation/3654662	MULTIPOLYGON (((17.05674 51.12007, 17.05682 51...	park
relation/4552866	POLYGON ((16.87549 51.13633, 16.87538 51.1364,...	park
relation/6629819	POLYGON ((16.97217 51.08334, 16.97213 51.08324...	park
relation/6727464	POLYGON ((16.97677 51.0944, 16.97681 51.0944, ...	park
...	...	...
way/1316998958	POLYGON ((17.03157 51.1295, 17.03162 51.12933,...	park
way/1318777588	POLYGON ((16.92317 51.13584, 16.92343 51.13581...	park
way/1320420245	POLYGON ((16.94538 51.10385, 16.94556 51.10381...	park
way/1360073315	POLYGON ((16.90986 51.17, 16.90885 51.17011, 1...	park
way/1366196846	POLYGON ((17.03293 51.10299, 17.0335 51.10292,...	park

305 rows × 2 columns

In [8]:

Copied!

folium_map = plot_regions(wroclaw_gdf, colormap=["lightgray"], tiles_style="CartoDB positron")
parks_gdf.explore(m=folium_map, color="forestgreen")
folium_map = plot_regions(wroclaw_gdf, colormap=["lightgray"], tiles_style="CartoDB positron")
parks_gdf.explore(m=folium_map, color="forestgreen")

Out[8]:

Make this Notebook Trusted to load map: File -> Trust Notebook

Uses grouped filters as an example.

In [9]:

Copied!





barcelona_gdf = geocode_to_region_gdf("Barcelona")
barcelona_filter = {
    "tourism": {"building": "hotel", "amenity": ["bar", "cafe", "pub"]},
    "sport": {"sport": "soccer", "leisure": ["pitch", "sports_centre", "stadium"]},
}

barcelona_objects_gdf = loader.load(barcelona_gdf, barcelona_filter)
barcelona_objects_gdf
barcelona_gdf = geocode_to_region_gdf("Barcelona")
barcelona_filter = {
    "tourism": {"building": "hotel", "amenity": ["bar", "cafe", "pub"]},
    "sport": {"sport": "soccer", "leisure": ["pitch", "sports_centre", "stadium"]},
}

barcelona_objects_gdf = loader.load(barcelona_gdf, barcelona_filter)
barcelona_objects_gdf

Out[9]:

	geometry	tourism	sport
feature_id
node/216330105	POINT (2.17484 41.3858)	amenity=bar	NaN
node/432592965	POINT (2.17077 41.38035)	amenity=cafe	NaN
node/499122827	POINT (2.17432 41.43045)	amenity=bar	NaN
node/499122828	POINT (2.17371 41.43198)	amenity=bar	NaN
node/499122829	POINT (2.17252 41.43052)	amenity=cafe	NaN
...	...	...	...
way/1371050943	POLYGON ((2.12837 41.39457, 2.12847 41.39449, ...	NaN	leisure=pitch
way/1371344761	POLYGON ((2.14249 41.36265, 2.14366 41.36264, ...	NaN	leisure=pitch
way/1371344763	POLYGON ((2.14242 41.36157, 2.14369 41.36165, ...	NaN	sport=soccer
way/1371344764	POLYGON ((2.14236 41.36242, 2.14362 41.36246, ...	NaN	sport=soccer
way/1371900982	POLYGON ((2.21076 41.41428, 2.21086 41.41435, ...	NaN	leisure=pitch

4299 rows × 3 columns