Rio-tiler

* Make sure output data is of type `Uint8` when applying a colormap (https://github.com/cogeotiff/rio-tiler/pull/423)
* Do not auto-rescale data if there is a colormap (https://github.com/cogeotiff/rio-tiler/pull/423)

* update type information for mosaics functions (https://github.com/cogeotiff/rio-tiler/pull/409)

* add support for setting the S3 endpoint url via the `AWS_S3_ENDPOINT` environment variables in `aws_get_object` function using boto3 (https://github.com/cogeotiff/rio-tiler/pull/394)
* make `ImageStatistics.valid_percent` a value between 0 and 100 (instead of 0 and 1) (author param-thakker, https://github.com/cogeotiff/rio-tiler/pull/400)
* add `fetch_options` to `STACReader` to allow custom configuration to the fetch client (https://github.com/cogeotiff/rio-tiler/pull/404)

python
with STACReader("s3://...", fetch_options={"request_pays": True}):
pass


* Fix alpha band values when storing `Uint16` data in **PNG**. (https://github.com/cogeotiff/rio-tiler/pull/407)

* add auto-rescaling in `ImageData.render` method to avoid error when datatype is not supported by the output driver (https://github.com/cogeotiff/rio-tiler/pull/391)

python
before - exit with error
with open("img.png", "wb") as f:
f.write(ImageData(numpy.zeros((3, 256, 256), dtype="float32")).render())
>>> (ERROR) CPLE_NotSupportedError: "PNG driver doesn't support data type Float32. Only eight bit (Byte) and sixteen bit (UInt16) bands supported".

now - print a warning
with open("img.png", "wb") as f:
f.write(ImageData(numpy.zeros((3, 256, 256), dtype="float32")).render())
>>> (WARNING) InvalidDatatypeWarning: "Invalid type: `float32` for the `PNG` driver. Data will be rescaled using min/max type bounds".


**breaking changes**

* change type of `in_range` option in `ImageData.render` to `Sequence[Tuple[NumType, NumType]]` (https://github.com/cogeotiff/rio-tiler/pull/391)

python
img = ImageData(numpy.zeros((3, 256, 256), dtype="uint16"))

before - Tuple[NumType, NumType]
buff = img.render(in_range=(0, 1000, 0, 1000, 0, 1000))

now - Sequence[Tuple[NumType, NumType]]
buff = img.render(in_range=((0, 1000), (0, 1000), (0, 1000)))

* add warning when dataset doesn't have overviews (https://github.com/cogeotiff/rio-tiler/pull/386)
* add `width`, `height`, `count` and `overviews` infos in `COGReader.info()` (https://github.com/cogeotiff/rio-tiler/pull/387)
* add `driver` in `COGReader.info()` output (https://github.com/cogeotiff/rio-tiler/pull/388)
* add `valid_percent` in `stats` output (https://github.com/cogeotiff/rio-tiler/pull/389)

* use importlib.resources `.files` method to resolve the package directory (https://github.com/cogeotiff/rio-tiler/pull/379)