aiocogeo

Installation

pip install aiocogeo

# With S3 filesystem
pip install aiocogeo[s3]

# With tiler
pip install aiocogeo[tiler]

Usage

COGs are opened using the COGReader asynchronous context manager:

from aiocogeo import COGReader

async with COGReader("http://cog.tif") as cog:
    ...

Several filesystems are supported:

• HTTP/HTTPS (http://, https://)
• S3 (s3://)
• File (/)

Metadata

Generating a rasterio-style profile for the COG:

async with COGReader("https://async-cog-reader-test-data.s3.amazonaws.com/lzw_cog.tif") as cog:
    print(cog.profile)

>>> {'driver': 'GTiff', 'width': 10280, 'height': 12190, 'count': 3, 'dtype': 'uint8', 'transform': Affine(0.6, 0.0, 367188.0,
       0.0, -0.6, 3777102.0), 'blockxsize': 512, 'blockysize': 512, 'compress': 'lzw', 'interleave': 'pixel', 'crs': 'EPSG:26911', 'tiled': True, 'photometric': 'rgb'}

Lower Level Metadata

A COG is composed of several IFDs, each with many TIFF tags:

from aiocogeo.ifd import IFD
from aiocogeo.tag import Tag

async with COGReader("https://async-cog-reader-test-data.s3.amazonaws.com/lzw_cog.tif") as cog:
    for ifd in cog:
        assert isinstance(ifd, IFD)
        for tag in ifd:
            assert isinstance(tag, Tag)

Each IFD contains more granular metadata about the image than what is included in the profile. For example, finding the tilesize for each IFD:

async with COGReader("https://async-cog-reader-test-data.s3.amazonaws.com/lzw_cog.tif") as cog:
    for ifd in cog:
        print(ifd.TileWidth.value, ifd.TileHeight.value)

>>> 512 512
    128 128
    128 128
    128 128
    128 128
    128 128

More advanced use cases may need access to tag-level metadata:

async with COGReader("https://async-cog-reader-test-data.s3.amazonaws.com/lzw_cog.tif") as cog:
    first_ifd = cog.ifds[0]
    assert first_ifd.tag_count == 24

    for tag in first_ifd:
        print(tag)

>>> Tag(code=258, name='BitsPerSample', tag_type=TagType(format='H', size=2), count=3, length=6, value=(8, 8, 8))
    Tag(code=259, name='Compression', tag_type=TagType(format='H', size=2), count=1, length=2, value=5)
    Tag(code=257, name='ImageHeight', tag_type=TagType(format='H', size=2), count=1, length=2, value=12190)
    Tag(code=256, name='ImageWidth', tag_type=TagType(format='H', size=2), count=1, length=2, value=10280)
    ...

Image Data

The reader also has methods for reading internal image tiles and performing partial reads. Currently only jpeg, lzw, deflate, packbits, and webp compressions are supported.

Image Tiles

Reading the top left tile of an image at native resolution:

async with COGReader("https://async-cog-reader-test-data.s3.amazonaws.com/webp_cog.tif") as cog:
    x = y = z = 0
    tile = await cog.get_tile(x, y, z)
    
    ifd = cog.ifds[z]
    assert tile.shape == (ifd.bands, ifd.TileHeight.value, ifd.TileWidth.value)

Partial Read

You can read a portion of the image by specifying a bounding box in the native crs of the image and an output shape:

async with COGReader("https://async-cog-reader-test-data.s3.amazonaws.com/webp_cog.tif") as cog:
    assert cog.epsg == 26911
    partial_data = await cog.read(bounds=(368461,3770591,368796,3770921), shape=(512,512))

Internal Masks

If the COG has an internal mask, the returned array will be a masked array:

import numpy as np

async with COGReader("https://async-cog-reader-test-data.s3.amazonaws.com/naip_image_masked.tif") as cog:
    assert cog.is_masked

    tile = await cog.get_tile(0,0,0)
    assert np.ma.is_masked(tile)

Tiling

Aiocogeo has a tiling extra (pip install aiocogeo[tiling]) which adds coordinate system handling via rasterio and morecantile to primarily support dynamic tiling use cases. The interface is intentionally designed to be similar to rio-tiler-crs.

from aiocogeo.tiler import COGTiler

async with COGReader("http://oin-hotosm.s3.amazonaws.com/5a95f32c2553e6000ce5ad2e/0/10edab38-1bdd-4c06-b83d-6e10ac532b7d.tif") as cog:
    tiler = COGTiler(cog)
    tile = await cog.tile(x=691559, y=956905, z=21, tile_size=256)

Configuration

Configuration options are exposed through environment variables:

• INGESTED_BYTES_AT_OPEN - defines the number of bytes in the first GET request at file opening (defaults to 16KB)
• ENABLE_CACHE - determines if range requests are cached in memory (defaults to TRUE)
• HTTP_MERGE_CONSECUTIVE_RANGES - determines if consecutive ranges are merged into a single request (defaults to FALSE)
• BOUNDLESS_READ - determines if internal tiles outside the bounds of the IFD are read (defaults to TRUE)
• BOUNDLESS_READ_FILL_VALUE - determines the value used to fill boundless reads (defaults to 0)
• LOG_LEVEL - determines the log level used by the package (defaults to ERROR)
• VERBOSE_LOGS - enables verbose logging, designed for use when LOG_LEVEL=DEBUG (defaults to FALSE)
• AWS_REQUEST_PAYER - set to requester to enable reading from S3 RequesterPays buckets.

Refer to aiocogeo/config.py for more details about configuration options.

CLI

$ aiocogeo --help
Usage: aiocogeo [OPTIONS] COMMAND [ARGS]...

Options:
  --install-completion [bash|zsh|fish|powershell|pwsh]
                                  Install completion for the specified shell.
  --show-completion [bash|zsh|fish|powershell|pwsh]
                                  Show completion for the specified shell, to
                                  copy it or customize the installation.

  --help                          Show this message and exit.

Commands:
  create-tms  Create OGC TileMatrixSet.
  info        Read COG metadata.