Python packages for conda env

[emiliom]

@ocefpaf @bbest @jbosch-noaa Last week Don (@lsetiawan) started putting together a Docker file for the AWS cloud instance that'll host JupyterHub. He created a basic conda environment file for Python. It's here, but I'm pasting the listing of packages below, for your convenience.

If there are other Python packages you think we should include, please chime in via this issue. Remember that many of the packages below have a lot of dependencies that will be brought in by default, such as numpy, pandas, matplotlib, netcdf4, etc.

For example:

• @ocefpaf, is your erddapy package mature enough to be included?
• Is there such a thing as a Python Darwin Core reader / writer, or API to the biological taxonomy resources, that we should include?

@bbest feel free to open a similar github issue for R if you'd like, and ping Abby on that issue.

Here's the conda env file:

name: iooswkshp
channels:
- conda-forge
dependencies:
- python=3.6
- requests
- scipy
- xarray
- folium
- geopandas
- gdal=2.2
# Jupyter notebook
- ipykernel
- jupyter
- nb_conda_kernels

[ocefpaf]

@ocefpaf, is your erddapy package mature enough to be included?

Not really. I need to allocate some time to fix some remaining issues and issue a new release.

[emiliom]

@ocefpaf, are you saying that that's something you can do within the next couple of weeks, before the workshop?

I'd like to volunteer @lsetiawan to help @ocefpaf with testing and as a second pair of eyeballs 😃. Don, this will be good for you and us, as it'll give you more exposure to ERDDAP (client-side). And you can charge the BioData project and NANOOS DMAC (but we'll talk about that offline).

[bbest]

Hi @emiliom, I created another issue for R install: #5: Setup server-based R environment in Amazon Cloud. That should take care of dependencies for the most common geospatial and tidyverse packages.

[ocefpaf]

@ocefpaf, are you saying that that's something you can do within the next couple of weeks, before the workshop?

Depending on the functionalities you want to demonstrate, yes. There are some low hanging fruit that I can finish in 1 week and issue a new release. Please take a look at the examples for the current functionalities available and let me know you think is missing and/or what I should prioritize.

[emiliom]

Depending on the functionalities you want to demonstrate, yes. There are some low hanging fruit that I can finish in 1 week and issue a new release. Please take a look at the examples for the current functionalities available and let me know you think is missing and/or what I should prioritize.

@jbosch-noaa, any thoughts on what functionality an ERDDAP client library should prioritize for our workshop? @bbest, please chime in too if you have thoughts on this. I don't have Abby's github handle, but her input would be helpful too. Though maybe this topic should be split off as a separate issue, either on this repo or the erdapy repo?

Otherwise, my comments will be generic. And I'll be away for a week, so you'll be waiting for a while.

[jbosch-noaa]

I'm copying part of the email from @albenson-usgs into this thread:

Important R packages would be:
robis
obistools
tidyverse
uuid
taxizesoap
stringr
readr

[ocefpaf]

Thanks @jbosch-noaa! I'm adding those to conda-forge to make @lsetiawan's life easier.

xref.: conda-forge/staged-recipes#4834

[lsetiawan]

Thanks @ocefpaf

[AbbyBenson]

Another R package to add is finch.

[jbosch-noaa]

Hey folks,
I'm adding @samuelbosch to this conversation. He will be one of the instructors at the workshop from iOBIS.

[ocefpaf]

Another R package to add is finch.

Done in conda-forge/staged-recipes#4841

@albenson-usgs and @lsetiawan the r-finch will be available in conda-forge as soon as that PR is merged.

[lsetiawan]

@ocefpaf question about conda: Is there a way to install to root via an environment.yml file? Thanks