-
Notifications
You must be signed in to change notification settings - Fork 11
/
CHANGELOG
71 lines (71 loc) · 3.59 KB
/
CHANGELOG
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
v0.7.2:
* improve the documentation and the tutorials
v0.7.1:
* use pyfftw if environment variable YAM_FFTW is set and different from "no" or "false"
(before pyfftw was used if environment variable was set, regardless of its value)
v0.7.0:
* consequently use float16 in the code,
switching to other dtypes may not be completely supported at this point and
the possibility to change the dtype may be removed later
* use --njobs 0 for sequential processing (before it was 1)
and --njobs 1 for processing of each element in a new process
* index for cross-correlation is nested one level deeper now to not have
too many datsets in one group for cross-correlations of 10min data pieces
* add new plottype: velocity change versus time of different station/component combinations
together with joint estimate
* allow to set compression and other options when creating a dataset
* allow to set libver for new HDF5 files (default is 'earliest' for best backwards compatibility)
v0.6.0:
* yam can now be used on Windows, tests are passing, but multiproccesing is not yet working
(see issue #4)
v0.5.0:
* major refactoring of stretch module, some config options for stretching are renamed,
the resulting dictionaries changed in a backwards incompatible way, please (a) use
yam version < 0.5 or (b) adapt config file and run yam stretch again for any existing project
v0.4.3:
* add demean option (before filtering)
* add plugin option for timenorm
* refactor timenorm function, some kwargs are deprecated
* timenorm and spectral_whitening are now working with Trace objects (instead of numpy arrays)
* add dpi option for most plots to set resolution of saved images
* more verbose error message if coordinates of station cannot be found
* fix: guarantee same length of traces which are correlated (necessary if tolerance_shift is set)
v0.4.2:
* adapt to GFZ FDSN webservices and allow only 1 thread when downloading example data
v0.4.1:
* extension of images is now specified without dot, e.g. 'png', 'pdf'
v0.4.0:
* rename plotting options setting x- and y-limits to xlim and ylim
* parallel processing for stretching, reference trace must be used (--reftr)
* allow full stack with '' or 'all' (e.g. for creating a refence trace)
* parallel processing for full stack
* store correlations and sim_mat as half precission float16 by default
* fix various small bugs
* move some functions to new io module
v0.3.0:
* add scan subcommand
* add debug option (--pdb)
* add exhaustive online documentation
* add possibility to correlate "R" and "T" rotated components (experimental)
* add --parallel-inner-loop feasible for correlating many stations
* downsample now performed before preprocessing
* add decimate option (downsampling after preprocessing)
* reduce run time of tests
* fix: timenorm "clip" was not correctly working
* fix: timenorm "mute_envelope" was not working with masked data
v0.2.1:
* refactor part of the imaging module
* rename some kwargs in imaging module
v0.2.0:
* rename read_stretch to read_dicts
* remove unnecessary "root" group in HDF5 files
* change stretch to velocity change inside dictionary returned by read_stretch (sign change)
* change unit of velocity change and str_range to %
* corr_vs_time plot: allow to plot stack at top
* allow strings for length and overlap option in correlate config (e.g. "1h" for one hour)
* fix: correctly plot gaps in corr_vs_time and sim_mat plots
* fix: downsample now defaults to None (previous default: 10)
v0.1.2:
* fix: remove_response option was not working
v0.1.0:
* initial release