Add support for array-based bulk insert #340
Conversation
for more information, see https://pre-commit.ci
|
Test failures are expected as it depends on new functionality for libspatialindex. For my case of inserting ~1 million points time goes from ~6s to ~1s. |
|
These tests should be resilient to different versions of libspatialindex. Rtree releases are not pinned to specific libspatialindex releases. |
FreddieWitherden
force-pushed
the
master
branch
from
02efce9 to
e4a3bd9
Compare
for more information, see https://pre-commit.ci
Have updated the PR so we catch the exceptions. Does this approach seem reasonable (assuming it checks out test wise?) |
FreddieWitherden
force-pushed
the
master
branch
from
1ffc13b to
1fd5065
Compare
for more information, see https://pre-commit.ci
|
Just a short note about the implementation: the inputs for this are required to be NumPy arrays however I've been careful to not import NumPy so it is not a dependency of the library. This should be reasonably flexible and anyone looking to seriously take advantage of the bulk capabilities really needs to be using NumPy (or a compatible alternative) anyway. |
This is the counterpart to libspatialindex/libspatialindex#267