pywws.filedata
Store weather data in easy to access files
Introduction
This module is at the core of pywws file based storage. It stores data on disc, but without the overhead of a full scale database system. I have designed it to run on a small memory machine such as a Raspberry Pi or even a router. To minimise memory usage it only loads one day’s worth of raw data at a time into memory.
From a “user” point of view, the data is accessed as a cross between a
list and a dictionary. Each data record is indexed by a
datetime.datetime object (dictionary behaviour), but
records are stored in order and can be accessed as slices (list
behaviour).
For example, to access the hourly data for Christmas day 2009, one might do the following:
from datetime import datetime
import pywws.filedata
hourly = pywws.filedata.HourlyStore('weather_data')
for data in hourly[datetime(2009, 12, 25):datetime(2009, 12, 26)]:
print(data['idx'], data['temp_out'])
Some more examples of data access:
# get value nearest 9:30 on Christmas day 2008
data[data.nearest(datetime(2008, 12, 25, 9, 30))]
# get entire array, equivalent to data[:]
data[datetime.min:datetime.max]
# get last 12 hours worth of data
data[datetime.utcnow() - timedelta(hours=12):]
Note that the datetime.datetime index is in UTC. You may
need to apply an offset to convert to local time.
The module provides five classes to store different data.
RawStore takes “raw” data from the weather station;
CalibStore, HourlyStore, DailyStore
and MonthlyStore store processed data (see
pywws.process). All are derived from the same CoreStore
class, they only differ in the keys and types of data stored in each
record.
Detailed API
Classes
|
Stores 'calibrated' weather station data. |
|
|
|
Stores daily summary weather station data. |
|
Stores hourly summary weather station data. |
|
Stores monthly summary weather station data. |
|
Stores raw weather station data. |
Comments or questions? Please subscribe to the pywws mailing list http://groups.google.com/group/pywws and let us know.