dymaxionlabs/dask-rasterio

Read and write rasters in parallel using Rasterio and Dask

dask-rasterio

dask-rasterio provides some methods for reading and writing rasters in parallel using Rasterio and Dask arrays.

Usage

Read a multiband raster

>>> from dask_rasterio import read_raster

>>> array = read_raster('tests/data/RGB.byte.tif')
>>> array
dask.array<stack, shape=(3, 718, 791), dtype=uint8, chunksize=(1, 3, 791)>

>>> array.mean()
dask.array<mean_agg-aggregate, shape=(), dtype=float64, chunksize=()>
>>> array.mean().compute()
40.858976977533935

Read a single band from a raster

>>> from dask_rasterio import read_raster

>>> array = read_raster('tests/data/RGB.byte.tif', band=3)
>>> array
dask.array<raster, shape=(718, 791), dtype=uint8, chunksize=(3, 791)>

Write a singleband or multiband raster

>>> from dask_rasterio import read_raster, write_raster

>>> array = read_raster('tests/data/RGB.byte.tif')

>>> new_array = array & (array > 100)
>>> new_array
dask.array<and_, shape=(3, 718, 791), dtype=uint8, chunksize=(1, 3, 791)>

>>> prof = ... # reuse profile from tests/data/RGB.byte.tif...
>>> write_raster('processed_image.tif', new_array, **prof)

Chunk size

Both read_raster and write_raster accept a block_size argument that acts as a multiplier to the block size of rasters. The default value is 1, which means the dask array chunk size will be the same as the block size of the raster file. You will have to adjust this value depending on the specification of your machine (how much memory do you have, and the block size of the raster).

Install

Install with pip:

pip install dask-rasterio

Development

This project is managed by Poetry. If you do not have it installed, please refer to Poetry instructions.

Now, clone the repository and run poetry install. This will create a virtual environment and install all required packages there.

Run poetry run pytest to run all tests.

Run poetry build to build package on dist/.

Issue tracker

Please report any bugs and enhancement ideas using the GitHub issue tracker:

https://github.com/dymaxionlabs/dask-rasterio/issues

Feel free to also ask questions on our Gitter channel, or by email.

Help wanted

Any help in testing, development, documentation and other tasks is highly appreciated and useful to the project.

For more details, see the file CONTRIBUTING.md.

License

Source code is released under a BSD-2 license. Please refer to LICENSE.md for more information.

dask-rasterio's Issues

the time to write a dask array in tif is too long?

I have a image which have 40000cols and 40000 rows ,while I use data = read_raster(filename, band=1), and the use witer_raster(out_filename, data), it need about one minutes, I want to know can it be quickly

TypeError: self._hds cannot be converted to a Python object for pickling

Seems that rasterio's _hds object is no more serializable

distributed.protocol.pickle - INFO - Failed to serialize ("('filled-2f9fe0560be0502eda038fa941309294', 0, 0)", <dask_rasterio.write.RasterioDataset object at 0x7f8f9deac828>, (slice(0, 748, None), slice(0, 22415, None)), <unlocked _thread.lock object at 0x7f8f9cb2af58>, False). Exception: self._hds cannot be converted to a Python object for pickling
---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
~/miniconda3/envs/jupyter/lib/python3.6/site-packages/distributed/protocol/pickle.py in dumps(x)
     37     try:
---> 38         result = pickle.dumps(x, protocol=pickle.HIGHEST_PROTOCOL)
     39         if len(result) < 1000:

~/miniconda3/envs/jupyter/lib/python3.6/site-packages/rasterio/_io.cpython-36m-x86_64-linux-gnu.so in rasterio._io.DatasetWriterBase.__reduce_cython__()

TypeError: self._hds cannot be converted to a Python object for pickling

Does dask-rasterio support masked array?

I'm working with dask masked array, and was wondering what would be the translation of these lines?

    import rasterio
    with rasterio.open(inputFile) as source:
        # this is a 3D numpy array, with dimensions [band, row, col]
        src_array = source.read(masked=True)

Thank you for your cool lib!