meridian 0.1.0

Easy geospatial data processing.

meridian

Higher-level geospatial data abstractions for Python.

meridian lets you treat your geospatial dataset like you would a normal Python data structure, backed by a spatial index for high-performance geospatial queries.

Usage

The core data structure implemented is the SpatialDataset, which takes an iterable (list, generator, etc) of GeoJSON-structured dicts and builds up the dataset and index. The records in your dataset are stored within the SpatialDataset as namedtuples with all the input field names as attributes. Once it's done loading, you have a familiar-feeling data structure you can use to query your dataset.

meridian is fully compatible with the GeoJSON-like objects produced by the fiona library, which makes it very easy to use:

import fiona

from meridian import SpatialDataset
from shapely import geometry

with fiona.open('/path/to/my/shapefile.shp') as src:
    dataset = SpatialDataset(src)


poi = geometry.shape({
    'type': 'Point',
    'coordinates': [-72.319261, 43.648956]
})

for record in dataset.intersection(poi):
    print(record)

All of the spatial query methods on a SpatialDataset require only that the query object has a bounds property which returns a 4-tuple like (xmin, ymin, xmax, ymax). As long as that exists, meridian is agnostic of query geometry implementation, however it does use shapely geometry on the backend for the records stored within.

The records in a SpatialDataset are spatialdata namedtuples with three special properties:

for record in dataset:
    print(record.id)  # The id in the `id` field of the input geojson
    print(record.geom)  # The `shapely` geometry representation of the record
    print(record.bounds)  # The bounds of the geometry
    print(record.my_fancy_property)  # All other properties in the geojson feature will be exposed as attributes on the namedtuple

Since the id field is not part of the GeoJSON spec it is optional to include; the library will function just fine without it. However, it does give the user a means to uniquely identify the records within each dataset.

Installation

meridian requires GEOS (for the shapely library) and libspatialindex to create the spatial index used for querying. On most systems, libspatialindex must be compiled from source. These instructions should work on Linux & macOS:

wget -qO- http://download.osgeo.org/libspatialindex/spatialindex-src-1.8.5.tar.gz | tar xz -C /tmp
cd /tmp/spatialindex-src-1.8.5 && ./configure; make; make install

On Linux, you might be to run ldconfig afterwards to ensure that the rtree python library can find the library correctly.

Gotcha's

Data takes up memory. Depending on the number & size of the geometries you're trying to work with, you might run out of memory. On my machine, a 2016 MacBook Pro, I found that a dataset with 350k records with an average of 6 nodes per polygon used ~500mb of memory footprint. YMMV.

meridian is opinionated and believes that data should be immutable. If you need your data to change, you should create new data representing your input + processing instead of changing old data. Thus, a SpatialDataset is more like a frozenset in behavior than a list.

Planned features

• Format compat. Built-in tools to help load data from other formats (Postgres, WKT, etc)
• Helper function for common geospatial comparisons between spatialdata objects