Note: Pyoos is very much a work in progress and should considered experimental until a 1.0 release is made!

Pyoos attempts to fill the need for a high level data collection library for met/ocean data publically available through many different websites and webservices.

Pyoos will collect and parse the following data services into the Paegan Discrete Geometry CDM:

• NERRS Observations - SOAP
• NDBC Observations - IOOS SWE SOS 1.0
• CO-OPS Observations - IOOS SWE SOS 1.0
• STORET Water Quality - WqxOutbound via REST (waterqualitydata.us)
• USGS NWIS Water Quality - WqxOutbound via REST (waterqualitydata.us)
• USGS Instantaneous Values - WaterML via REST
• NWS AWC Observations - XML via REST (http://www.aviationweather.gov)

# (minx, miny, maxx, maxy)
collector.filter(bbox=(-74, 30, -70, 38))

from dateime import dateime, timedelta
collector.filter(start=datetime.utcnow() - timedelta(hours=1))

from dateime import dateime
collector.filter(end=datetime.utcnow())

from dateime import dateime, timedelta
collector.filter(start=datetime.utcnow - timedelta(hours=24), end=datetime.utcnow())

It is highly dependent on the data provider how they identify unique features/stations/objects. Pyoos does its best job to figure out what you are passing in. For example, you may pass WMO ID's to the NDBC collector and Pyoos will request the correct complete URN to the NDBC SOS.

collector.list_features()

# Any iterable of strings
collector.filter(features=["21KY-BSW004"])

Pyoos does its best job to format any string into the correct format for the actual request. For example, you may pass typical standard_name string from CF-1.6 to the NDBC collector and Pyoos will turn it into a complete MMI URI.

collector.list_variables()

# Any iterable of strings
collector.filter(variables=["sea_water_temperature"])

collector.clear()

You may chain many filter calls together (it returns a collector object)

collection.filter(bbox=(-74, 30, -70, 38)).filter(end=datetime.utcnow())

You may also combine many filter types into one call to filter

collection.filter(bbox=(-74, 30, -70, 38), end=datetime.utcnow())

collector.collect()

collector.raw()

Each collector may implement a set of functions specific to that collection. Please see the Wiki for an explanation of this type of functionality.

You are using virtualenv, right?

1. Install virtualenv-burrito
2. Create virtualenv named "pyoos-dev": mkvirtualenv -p your_python_binary pyoos-dev
3. Start using your new virtualenv: workon pyoos-dev

Pyoos requires python 2.7.x and is available on PyPI.

The best way to install Pyoos is through pip:

pip install pyoos

Pyoos requires the following python libraries which will be downloaded and installed through pip:

• Paegan>=0.9.6
• OWSLib (custom fork... but will be intergrated soon!)
• Shapely>=1.2.17
• pytz>=2012h
• python-dateutil>=2.1
• requests

If your NetCDF4 and HDF5 libraries are in non-typical locations, you will need to pass the locations to the pip command:

NETCDF4_DIR=path HDF5_DIR=path pip install pyoos

There seems to be a problem installing numpy through pip dependency chains so you may need to install numpy before doing any of the above:

pip install numpy==1.6.2

• Development of a standardized Metadata concept, possibly through SensorML and/or ISO 19115-2

Submit a PR with your use case!

If you are having trouble getting any of the pyoos functionality to work, try running the tests:

git clone git@github.com:asascience-open/pyoos.git
cd pyoos
python setup.py test

• Kyle Wilcox kwilcox@asascience.com
• Sean Cowan scowan@asascience.com
• Alex Crosby acrosby@asascience.com