mapbox / grib-doctor

Utilities for handling quirks of weather data grib files.

gribdoctor

Utility functions for handling quirks of weather data rasters in grib2 format

Installation -------.. role:: console(code) :language: console

pip install gribdoctor --pre

Usage - smoosh

stack multiple gribs of the same or varying resolutions

Usage: gribdoctor smoosh [OPTIONS] [INPUTS]... OUTPUT

OPTIONS:

-dev, --develoment Dev tag for experimental features

-uw, --unwrap Unwrap GFS Rasters

--help Show this message and exit.

Use this subcommand to "smoosh" together any number of variable resolution gribs into one tiff. This also effectively performs the below globewrap operation on all inputs if indicated with --unwrap, and outputs a raster of 2x (due to unwrapping) the highest input resolution.

Usage - unwrap

gribdoctor unwrap [OPTIONS] INGRIB OUTRASTER

OPTIONS

-bt, --bandtags Flag to indicate printing of band tags / band metadata to stdout

-b, --bidx Bands to include in output raster. Default = all (use caution - you may inadvertently create ginourmous tif files)

Many gridded weather GRIBs, such as NOAA's Global Forecast System (GFS), cover a global extent of -0.5 to 359.5 (or similar, depending on resolution), with the antimeridian (-180 / 180) bisecting a column of pixels. This routine upsamples, slices, and merges these rasters as to convert them to "standard" -180 to 180 extent global grids. Outputs to any raster format supported by rasterio.

grib-doctor's Issues

Calculate wind direction and speeds from UGRID and VGRID

I am thinking the best approach would be simply as an option that one could pass to smoosh or unwrap, eg:

gribdoctor unwrap --wind {'UGRID':<band>,'VGRID':<band>}

cc: @ian29

Handling for non GFS data in smoosh

• By specifying an explicit flag, don't unwrap data.
• By specifying another -dev flag, autodetect file type and don't unwrap.

cc: @ian29

import cut_splice_globewrap as csg, click ModuleNotFoundError: No module named 'cut_splice_globewrap'

i 'm using gribdoctor but i got this issue. Can you help me??? Thank you.

Handle single band rasters in smoosh

While it is counterintuitive in some ways, this will enable better program flow.

singleband .grib in --> singleband .tif out.

cc @ian29

Update grib doctor to Rasterio 1.0

I am using Python 2.7.13 (64 bit) in windows 7 and I have successfully installed gdal, rasterio (following https://mapbox.github.io/rasterio/installation.html) and gribdoctor (using pip). So, whenever I am using gribdoctor smoosh -dev -uw raw.grib2 unwrap.tif command, I am getting the
error- "with rasterio.drivers(): TypeError: 'module' object is not callable." How can I fix this problem so that I can get converted output? Please note that raw.grib2 is the input GRIB file and unwrap.tif will be the output file.

"AttributeError: 'DatasetReader' object has no attribute 'affine'."

Hello, thanks for the tutorial!
I am using Python 3 on mac.
I have Installed gribdoctor using pip, but when I use gribdoctor smoosh -dev -uw raw.grib2 unwrap.tif command, I am getting the
error- "AttributeError: 'DatasetReader' object has no attribute 'affine'."
Do you know how I can fix the problem? I'm able to import the file on qgis but it's flipped.. Is there a way to fix it directly in qgis? thanks! I'll attach both my qgis screenshot and the terminal

Write better crs output for unwrap

Right now, writes existing crs even after unwrapping.

cc: @ian29

Command for upsampling, slicing, and flipping 0 - 360 grib rasters

gribnormalize unwrap INPUT OUTPUT

1. Open a grib2 of n bands
2. For each band, upsample, slice, and split
3. Write to new raster

@ian29