dweather-client 2.1.0

Python client for interacting with weather data on IPFS.

dWeather-Python-Client

Quickstart

Install dweather_client:

pip3 install dweather_client

Get valid dataset names and associated hashes:

http_client.get_heads()

Get the metadata for a given dataset name:

http_client.get_metadata('chirps_05-daily')

Get a rainfall dict for a gridded dataset:

http_client.get_rainfall_dict(41.175, -75.125, 'chirps_05-daily')

Get a rainfall dataframe:

df_loader.get_rainfall_df(41.125, -75.125, 'chirps_05-daily')

Get a station dataframe:

df_loader.get_station_rainfall_df('USW00024285')
df_loader.get_station_temperature_df('USW00024285')
df_loader.get_station_snow_df('USW00024285')

See further examples in tests

Development

cd dweather-python-client
python3 -m venv .
bin/pip3 install -r requirements.txt
bin/python3 -m pytest dweather_clieht/tests

Some dweather_client features require an ipfs daemon to work.

Download Go-IPFS version 0.6.0

See Assets list at the bottom of this page: https://github.com/ipfs/go-ipfs/releases/tag/v0.6.0 Download the build appropriate for your machine, or just download the source tar if you're not sure.

Install go-IPFS

Unzip the file that you downloaded.

tar xvfz {filename}.tar.gz

Move the binary into your path. sudo may be required for this.

mv go-ipfs/ipfs /usr/local/bin/ipfs

Initialize a ~/.ipfs directory. This is where your files and config will be saved.

ipfs init

Configure go-IPFS

Remove default peers for performance.

ipfs bootstrap rm --all

Add the dWeather server as a peer.

ipfs bootstrap add  "/ip4/198.211.104.50/tcp/4001/p2p/QmWsAFSDajELyneR7LkMsgfaRk2ib1y3SEU7nQuXSNPsQV"

Make sure go-IPFS it works

Start the IPFS daemon. You will need to have the daemon running to use some functionality of the dWeather client.

ipfs daemon

In a new window, confirm that you can pull content.

ipfs cat QmVsy2HZCi39ePJRpNqXEJvHgRMqjcyu1FLqgiFkPTMknq/USW00014704.csv.gz

Confirm that the dWeather server is a peer.

ipfs swarm peers

Install the Python dWeather Client

Create an isolated Python installation and install the dependencies.

git clone https://github.com/Arbol-Project/dWeather-Python-Client.git
cd dWeather-Python-Client
python3 -m venv .
bin/pip3 install -r requirements.txt

Run the tests, if you want.

bin/python3 -m pytest -s --log-cli-level=20 dweather_client/tests

Local data

Certain dweather functions will try to save query results locally to disk for faster performance on subsequent loads. This can be overridden by passing pin=False in these function calls.

Load the ipfs UI to browse what files are stored locally. Paste the following into a web browser. http://127.0.0.1:5001/webui

Navigate to Files, then click "pins." Content can be unpinned via the UI.

If you just want to remove everything, delete ~/.ipfs and rerun the installation and configuration from ipfs init.

rm -rf ~/.ipfs

Further documentation

See tests directory for example usage. Documented examples of usage should appear in a docs repository or in product-dev-notebook.