obspyh5 0.3.0

Installation

Install h5py and obspy. After that install obspyh5 using pip by:

pip install obspyh5

Alternatively you can install obspyh5 by downloading the source code and running:

python setup.py install

Usage

Basic example using the obspy plugin:

>>> from obspy import read
>>> stream = read()  # load example stream
>>> print(stream)
..3 Trace(s) in Stream:
BW.RJOB..EHZ | 2009-08-24T00:20:03.000000Z - 2009-08-24T00:20:32.990000Z | 100.0 Hz, 3000 samples
BW.RJOB..EHN | 2009-08-24T00:20:03.000000Z - 2009-08-24T00:20:32.990000Z | 100.0 Hz, 3000 samples
BW.RJOB..EHE | 2009-08-24T00:20:03.000000Z - 2009-08-24T00:20:32.990000Z | 100.0 Hz, 3000 samples
>>> stream.write('test.h5', 'H5')  # declare 'H5' as format
>>> print(read('test.h5'))  # Order is not preserved!
3 Trace(s) in Stream:
BW.RJOB..EHZ | 2009-08-24T00:20:03.000000Z - 2009-08-24T00:20:32.990000Z | 100.0 Hz, 3000 samples
BW.RJOB..EHE | 2009-08-24T00:20:03.000000Z - 2009-08-24T00:20:32.990000Z | 100.0 Hz, 3000 samples
BW.RJOB..EHN | 2009-08-24T00:20:03.000000Z - 2009-08-24T00:20:32.990000Z | 100.0 Hz, 3000 samples

Example iterating over traces in a huge hdf5 file. After each iteration the trace is not kept in memory and therefore it is possible to process a huge hdf5 file on a PC without problems.

>>> from obspyh5 import iterh5
>>> for trace in iterh5('huge_in.h5')
        trace.do_something()
        trace.write('huge_out.h5', 'H5', mode='a')  # append mode to write into file

Alternative indexing

obspyh5 supports alternative indexing.

>>> from obspy import read
>>> import obspyh5
>>> print(obspyh5._INDEX)  # default index
waveforms/{network}.{station}/{location}.{channel}/{starttime.datetime:%Y-%m-%dT%H:%M:%S}_{endtime.datetime:%Y-%m-%dT%H:%M:%S}

The index gets populated by the stats object when writing a trace, e.g.

>>> stats = read()[0].stats
>>> print(obspyh5._INDEX.format(**stats))
'waveforms/BW.RJOB/.EHZ/2009-08-24T00:20:03_2009-08-24T00:20:32'

To change the index use set_index.

>>> obspyh5.set_index('xcorr')  # xcorr indexing
>>> obspyh5.set_index('waveforms/{network}.{station}/{distance}')  # custom indexing

When using the ‘xcorr’ indexing stats needs the entries ‘network1’, ‘station1’, ‘location1’, ‘channel1’, ‘network2’, ‘station2’, ‘location2’ and ‘channel2’ of the first and second station. An example:

>>> from obspy import read
>>> import obspyh5
>>> obspyh5.set_index('xcorr')  # activate xcorr indexing
>>> stream = read()
>>> for i, tr in enumerate(stream):  # manipulate stats object
        station1, station2 = 'ST1', 'ST%d' % i
        channel1, channel2 = 'HHZ', 'HHN'
        s = tr.stats
        # we manipulate seed id so that important information gets
        # printed by obspy
        s.network, s.station = s.station1, s.channel1 = station1, channel1
        s.location, s.channel = s.station2, s.channel2 = station2, channel2
        s.network1 = s.network2 = 'BW'
        s.location1 = s.location2 = ''
>>> print(stream)
ST1.HHZ.ST0.HHN | 2009-08-24T00:20:03.000000Z - 2009-08-24T00:20:32.990000Z | 100.0 Hz, 3000 samples
ST1.HHZ.ST1.HHN | 2009-08-24T00:20:03.000000Z - 2009-08-24T00:20:32.990000Z | 100.0 Hz, 3000 samples
ST1.HHZ.ST2.HHN | 2009-08-24T00:20:03.000000Z - 2009-08-24T00:20:32.990000Z | 100.0 Hz, 3000 samples
>>> stream.write('test_xcorr.h5', 'H5')
>>> print(read('test_xcorr.h5'))
ST1.HHZ.ST0.HHN | 2009-08-24T00:20:03.000000Z - 2009-08-24T00:20:32.990000Z | 100.0 Hz, 3000 samples
ST1.HHZ.ST1.HHN | 2009-08-24T00:20:03.000000Z - 2009-08-24T00:20:32.990000Z | 100.0 Hz, 3000 samples
ST1.HHZ.ST2.HHN | 2009-08-24T00:20:03.000000Z - 2009-08-24T00:20:32.990000Z | 100.0 Hz, 3000 samples

Note

See also ASDF for a more comprehensive approach.

Use case: Cross-correlation of late Okhotsk coda (notebook).