aodn-public-notebooks

These Jupyter notebooks are developed by the AODN for public use.

To run the notebooks, you will need to install required dependencies pip install -r requirements.txt, it's recommended to use virual Python environment tools such as Conda.

conda create -n <your_env_name> python=<version> ipython --yes

Add your env to Jupyter Lab:

python -m ipykernel install --user --name <your_env_name> --display-name "<name_you_want>"

Below is a list of dependencies used for multi-purpose environment for the time being, feel free to remove the packages that may not be required in your scope of usage:

setuptools
xarray[complete]
tornado
jinja2==3.0.3
bokeh==2.4.3
boto3==1.20.24
botocore==1.23.24
jsonschema==4.7.2
toolz==0.11.2
git+https://github.com/intake/kerchunk.git@0.0.5
cloudpickle==2.0.0
dask==2022.1.1
distributed==2022.1.1
msgpack==1.0.2
toolz==0.11.2
pandas==1.3.4
numpy==1.21.5
graphviz==0.8.1
click==8.0.4
dask-cloudprovider==2022.1.0
jedi==0.18.1
s3fs==2022.1.0
fsspec==2022.1.0
h5netcdf==0.13.1
numcodecs==0.9.1
blosc==1.10.6
lz4==3.1.10
pickle5==0.0.11
retry==0.9.2
pyyaml==6.0
seawater==3.3.4
awswrangler==2.14.0
netCDF4==1.6.0
zarr==2.11.0
aioboto3==9.3.1
aiobotocore[boto3]==2.1.0
urllib3==1.26.7

If running Conda env, installing eccodes: conda install -c conda-forge -y eccodes fixes some library import warnings.

Also with Conda, if facing GDAL related errors during pip installing dependencies, run conda uninstall gdal then install again conda install gdal will help.

Our imos-data-pixeldrill is publicly available, sample method to access our datasets:

import s3fs
s3 = s3fs.S3FileSystem(anon=True)
store_path = 's3://imos-data-pixeldrill/csiro-argo.zarr/'
store = s3fs.S3Map(root=store_path, s3=s3)
data = xr.open_zarr(store)