pywws.Process

Generate hourly, daily & monthly summaries of raw weather station data

usage: python -m pywws.Process [options] data_dir
options are:
 -h or --help     display this help
 -v or --verbose  increase number of informative messages
data_dir is the root directory of the weather data

This module takes raw weather station data (typically sampled every five or ten minutes) and generates hourly, daily and monthly summary data, which is useful when creating tables and graphs.

Before computing the data summaries, raw data is “calibrated” using a user-programmable function. See pywws.calib for details.

The hourly data is derived from all the records in one hour, e.g. from 18:00:00 to 18:59:59, and is given the index of the last complete record in that hour.

The daily data summarises the weather over a 24 hour period typically ending at 2100 or 0900 hours, local (non DST) time, though midnight is another popular convention. It is also indexed by the last complete record in the period. Daytime and nightime, as used when computing maximum and minimum temperatures, are assumed to start at 0900 and 2100 local time, or 1000 and 2200 when DST is in effect, regardless of the meteorological day.

To adjust the meteorological day to your preference, or that used by your local official weather station, edit the “day end hour” line in your weather.ini file, then run Reprocess.py to regenerate the summaries.

Monthly summary data is computed from the daily summary data. If the meteorological day does not end at midnight, then each month may begin and end up to 12 hours before or after midnight.

Wind speed data is averaged over the hour (or day) and the maximum gust speed during the hour (or day) is recorded. The predominant wind direction is calculated using vector arithmetic.

Rainfall is converted from the raw “total since last reset” figure to a more useful total in the last hour, day or month.

Functions

Process(params, raw_data, calib_data, …)	Generate summaries from raw weather station data.
calibrate_data(logger, params, raw_data, …)	‘Calibrate’ raw data, using a user-supplied function.
generate_daily(logger, day_end_hour, …)	Generate daily summaries from calibrated and hourly data.
generate_hourly(logger, calib_data, …)	Generate hourly summaries from calibrated data.
generate_monthly(logger, rain_day_threshold, …)	Generate monthly summaries from daily data.
main([argv])

Classes

Average()	Compute average of multiple data values.
DayAcc()	‘Accumulate’ weather data to produce daily summary.
HourAcc(last_rain)	‘Accumulate’ raw weather data to produce hourly summary.
Maximum()	Compute maximum value and timestamp of multiple data values.
Minimum()	Compute minimum value and timestamp of multiple data values.
MonthAcc(rain_day_threshold)	‘Accumulate’ daily weather data to produce monthly summary.
WindFilter([decay])	Compute average wind speed and direction.

class pywws.Process.Average

Compute average of multiple data values.

add(value)

result()

class pywws.Process.Minimum

Compute minimum value and timestamp of multiple data values.

add(value, time)

result()

class pywws.Process.Maximum

Compute maximum value and timestamp of multiple data values.

add(value, time)

result()

class pywws.Process.WindFilter(decay=1.0)

Compute average wind speed and direction.

The wind speed and direction of each data item is converted to a vector before averaging, so the result reflects the dominant wind direction during the time period covered by the data.

Setting the decay parameter converts the filter from a simple averager to one where the most recent sample carries the highest weight, and earlier samples have a lower weight according to how long ago they were.

This process is an approximation of “exponential smoothing”. See Wikipedia for a detailed discussion.

The parameter decay corresponds to the value (1 - alpha) in the Wikipedia description. Because the weather data being smoothed may not be at regular intervals this parameter is the decay over 5 minutes. Weather data at other intervals will have its weight scaled accordingly.

The return value is a (speed, direction) tuple.

Parameters:decay (float) – filter coefficient decay rate. Return type:(float, float)

add(data)

result()

class pywws.Process.HourAcc(last_rain)

‘Accumulate’ raw weather data to produce hourly summary.

Compute average wind speed and maximum wind gust, find dominant wind direction and compute total rainfall.

reset()

add_raw(data)

result()

class pywws.Process.DayAcc

‘Accumulate’ weather data to produce daily summary.

Compute average wind speed, maximum wind gust and daytime max & nighttime min temperatures, find dominant wind direction and compute total rainfall.

Daytime is assumed to be 0900-2100 and nighttime to be 2100-0900, local time (1000-2200 and 2200-1000 during DST), regardless of the “day end hour” setting.

reset()

add_raw(data)

add_hourly(data)

result()

class pywws.Process.MonthAcc(rain_day_threshold)

‘Accumulate’ daily weather data to produce monthly summary.

Compute daytime max & nighttime min temperatures.

reset()

add_daily(data)

result()

pywws.Process.calibrate_data(logger, params, raw_data, calib_data)

‘Calibrate’ raw data, using a user-supplied function.

pywws.Process.generate_hourly(logger, calib_data, hourly_data, process_from)

Generate hourly summaries from calibrated data.

pywws.Process.generate_daily(logger, day_end_hour, calib_data, hourly_data, daily_data, process_from)

Generate daily summaries from calibrated and hourly data.

pywws.Process.generate_monthly(logger, rain_day_threshold, day_end_hour, daily_data, monthly_data, process_from)

Generate monthly summaries from daily data.

pywws.Process.Process(params, raw_data, calib_data, hourly_data, daily_data, monthly_data)

Generate summaries from raw weather station data.

The meteorological day end (typically 2100 or 0900 local time) is set in the preferences file weather.ini. The default value is 2100 (2200 during DST), following the historical convention for weather station readings.

pywws.Process.main(argv=None)

---

Comments or questions? Please subscribe to the pywws mailing list http://groups.google.com/group/pywws and let us know.