Testing H5Coro

conda-environment.yml

name: h5coro
channels:
- conda-forge
dependencies:
- gdal
- ipython
- proj
- pybind11
- python

h5coro-test.py

# Installation. Something like this should work

# # Create a conda environment
# conda env create h5coro
# conda install -c conda-forge python pybind11 gdal proj

# # Install script for cmake.
# # Run this from inside the `sliderule` source code directory:
#
# TARGETDIR="/path/to/sliderule-install"
# mkdir -p build
# cd build
#
# # Command below is *inside* the `build` directory
# cmake \
#   -DINSTALLDIR="$TARGETDIR" \
#   -DRUNTIMEDIR="$TARGETDIR/etc/sliderule" \
#   -DSHARED_LIBRARY=ON \
#   -DPYTHON_BINDINGS=ON \
#   -DUSE_AWS_PACKAGE=ON \
#   -DUSE_GEO_PACKAGE=ON \
#   -DUSE_NETSVC_PACKAGE=ON \
#   ..
#
# make
# # Or, for parallel install with verbose logging, `make -j8 VERBOSE=1`
# # ...and then, install files to `$TARGETDIR`
# make install

# # Run the python script below with:
# LD_LIBRARY_PATH="/path/to/sliderule-install/lib" python script.py
#
# # The LD_LIBRARY_PATH needs to contain the following:
# #   - libsliderule.so.X.X.X
# #   - srpybin.cypthon-<py_version>-<arch>-<os>-<glib_ver>.so

# This is needed to detect the `srpybin` library.
import sys
import os
# Change to:
# sliderule_lib = os.path.abspath("/path/to/sliderule-install/lib")
sliderule_lib = os.path.abspath("lib/")
sys.path.append(sliderule_lib)

# This is the sliderule Python bindings, including H5Coro
import srpybin

url = "s3://noaa-goes17/ABI-L2-FDCC/2018/239/00/OR_ABI-L2-FDCC-M3_G17_s20182390052191_e20182390054564_c20182390055159.nc"

asset = "awsodr"      # Dummy variable
resource = os.path.basename(url)
identity = "mystore"  # Dummy variable
driver = "s3"
path = os.path.dirname(url.lstrip("s3://"))
region = "us-east-1"
endpoint = "https://noaa-goes17.s3.amazonaws.com"

h5file = srpybin.h5coro(asset, resource, identity, driver, path, region, endpoint)
# Get metadata for specific variables.
# This works...
h5file.meta("Temp")
h5file.meta("Power")

# Try to read a particular variable.
# This fails with "invalid read location", perhaps because the data are not 
# strictly 2-dimensional and/or there are extra rows.
# Arguments: (Variable name, column, start row, number of rows)
h5file.read("Power", 0, 0, 1)

# Try to read multiple variables in parallel.
# This doesn't
dat = h5file.readp([
    ["Temp", 0, 0, 1],
    ["Temp", 0, 0, 1],
    ["Temp", 0, 0, 10]
])