Depreciated version

This code runs under Python 2.7 and with SSR7 RadialMetric data. Please see updated official respository for qccodar with newer versions of SeaSonde software (SSR8) under Python 3, using Mike Smith's HFRadarPy toolbox for manipulating radial data files.
While we are contributing to this updated code it is now maintained by Teresa Updyke (@teresaupdyke).

qccodar

This python code applies several quality control (QC) functions based on CODAR SeaSonde (COS) Radialmetric data currently output in COS RadialSuite version 7.x. There are two modes to this code: an auto- and manual-mode. Auto-mode is for realtime processing. Manual-mode is for processing all RadialMetric files in a specified folder in post-processing. qccodar is intended to run beside (in parallel) to SeaSonde Analysis and not supplant any processing. In fact, qccodar uses the LLUVMerger.app provided by SeaSonde to merge the data back into standard SeaSonde processing methodology.

Quickstart

If you are familiar with Python and your system Python installation, you can install the latest version using pip. Otherwise follow the detailed installation procedures provided below. After installing pip you can install qccodar with this command:

   $ pip install qccodar

This will install qccodar together with all the required dependencies. After RadialMetric output is generated, you can run qccodar either in realtime (auto-mode) processing data found in CODAR's default data directory /Codar/SeaSonde/Data:

   $ qccodar auto

*Note: place command in crontab entry to rerun

or manually (manual-mode) to process data in other locations, for example after running CODAR's SpectraOfflineProcessor:

   $ qccodar manual --pattern IdealPattern --datadir ./reprocess_HATY/2014_08

and for a little help with available options:

   $ qccodar --help

Installation

qccodar is a python package and runs under Python 2. Eventhough Mac OS X comes with Python 2 installed or you can install Python directly from python.org, it is recommended to use the lightweight option from Conda called miniconda. Miniconda contains only Python and other libraries needed to run Conda itself; other packages will be downloaded and installed as requested. Conda has a package manager which makes this installation easy. Conda also supports virtual environments where qccodar can run independently from the system-installed Python.

The following instructions show how to install and configure Miniconda2 and qccodar to provide QC'd Radial data based on RadialMetric output.

While logged on as user codar, open a terminal to download (curl) and run the installer script.

   $ cd ~/Downloads
   $ curl https://repo.continuum.io/miniconda/Miniconda2-latest-MacOSX-x86_64.sh -o "miniconda.sh"
   $ bash ~/Downloads/miniconda.sh -b -p $HOME/miniconda
   $ export PATH="$HOME/miniconda/bin:$PATH"

As, user codar, edit ~/.bash_profile, add the export PATH="$HOME/miniconda/bin:$PATH", and then source your profile.

   $ source ~/.bash_profile

Refer to the Conda Installation Guide for more details on installing Conda, making the appropriate adjustments for Miniconda2.

Creating a conda environment allows the qccodar module and its dependencies to run isolated from the installed Python avoiding potential module version conflicts. While this is not a requirement to get qccodar running, it is recommended.

   $ conda create --name qccodar python

Activate the environment:

   $ source activate qccodar
   (qccodar) $ which python
   /Users/codar/miniconda/envs/qccodar/bin/python

1. Install qccodar using pip:

   (qccodar) $ pip install qccodar

2. Alternatively, grab source, unpack, and install using sourced setup.py

Use git to retrieve latest repo from the github repo:

   (qccodar) $ git clone https://github.com/nccoos/qccodar.git

Use curl to grab the source distribution:

   $ curl -L -o "qccodar.tar.gz" http://github.com/nccoos/qccodar/tarball/master/

Install from source

   (qccodar) $ cd qccodar
   (qccodar) $ python setup.py install

Either method will install qccodar together with all the required dependencies. After configuring CODAR RadialSuite to ouput RadialMetric data, you can then run qccodar in either auto- or manual-mode.

Configuration and Crontab Entry for Realtime QC

First, enable RadialMetric output:

1. Edit line 21 of AnalysisOptions.txt in /Codar/SeaSonde/Configs/RadialConfigs.
2. Restart AnalyzeSpectra

   1           !21 Enable Radial Metric Output: 0(Off), 1(Enable), 2(Enable MaxVel.)

Second, set crontab entry to run qccodar:

1. Make a place to log data, e.g. $ mkdir ~/logs
2. Place entry in crontab to run every 15 minutes and log the output

$ crontab -l
1,11,21,31,41,51 * * * * /Codar/SeaSonde/Users/Scripts/collect/collect.pl
00,15,30,45 * * * * PATH=$PATH:/sbin /Users/codar/miniconda/envs/qccodar/bin/qccodar auto >> /Users/codar/logs/qccodar-auto.log 2>&1

If you get sh: sysctl: command cannot be found in output or log, sysctl might be in another path. qccodar still runs even when this cannot be found. In MacOS -- sysctl is sometimes located in /usr/bin (or /sbin) and may not be in the path under cron. Placing PATH=$PATH:/usr/sbin in the task entry, adds the path.

Background

Notes

QC Function	Description
Threshold Tests	badflag any values that fall below or above a single threshold
Weighted Averaging	average several values with weights based on signal quality parameters

QC Threshold Tests:

1. DOA peak power (MSR1, MDR1, MDR2) < 5 dB default
2. DOA 1/2 power width (3dB down) (MSW1, MDW1, MDW2) > 50 deg default
3. SNR on monopole (MA3S) < 5 dB default
4. SNR on both loops (MA1S and MA2S) < 5 dB

Weighted Averaging:

1. Weighting based on Music Power (MSP1, MDP1, MDP2)
2. Weighting based on SNR on monopole (MA3S)
3. No weight function (None)

System Requirements

• Python 2.7.x
• Numpy 1.9.x
  • https://pypi.python.org/pypi/numpy
  • Data read into memory are stored in the N-dimensional array datatype (ndarray) for indexing and computation.
• geopy 1.11.0
  • https://pypi.python.org/pypi/geopy
  • geopy.distance.vincenty()
  • Used to compute (LAT, LON) based on range and bearing from site origin in generating RadialShorts file
• watchdog 0.8.2
  • Used to monitor a directory for new files and trigger qc and merge process when new RadialMetric file is created
• CODAR SeaSonde RadialSuite 7.x (version 8 does not support RadialMetric output unless requested from CODAR)
  • /Codar/SeaSonde/Apps/Bin/LLUVMerger.app
  • Used to merge spatial and temporal RadialShorts data to final Radial