-
Notifications
You must be signed in to change notification settings - Fork 2
/
giverny-version.txt
144 lines (142 loc) · 7.34 KB
/
giverny-version.txt
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
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
giverny - stable release version 2.3.11 (16 August 2024)
- package for querying Johns Hopkins Turbulence Database (JHTDB) datasets.
version 2.3.11 updates:
- changed "position" variable calls to require "pchip" as the temporal interpolation method.
- fixed "position" variable calls to pass the specified spatial interpolation method.
version 2.3.10 updates:
- corrected metadata file generation bug.
version 2.3.9 updates:
- tweaked plotting code in getData notebooks.
- fixed input verification for 'option' parameter.
version 2.3.8 updates:
- updated aspect ratio of 3d plots to be proportional.
version 2.3.7 updates:
- added jupyter-core package version.
version 2.3.6 updates:
- fixed package dependencies for new 3d visualizations.
version 2.3.5 updates:
- DEMO notebook tweaks.
version 2.3.4 updates:
- corrected the time array for time series plotting.
version 2.3.3 updates:
- changed 'timepoint' to 'time' in the notebooks.
- moved where 'time' is specified by the user in the notebook.
- updated notebook filenames.
- fixed givernylocal getData output formatting.
version 2.3.2 updates:
- formatted (x, y, z) values in the time series demo plot to round to 2 decimal places.
- changed to getting the SciServer username from the SciServer.Authentication package.
- removed givernylocal getCutout function.
- updated result from getData to start with an array dimension for each queried timepoint.
- added timepoint column to the interpolation tsv file to account for time series queries.
version 2.3.1 updates:
- switched from dask futures to dask delayed for parallel queries in getCutout and getData.
version 2.2.7 updates:
- updated getCutout contour plot variable names to match getData.
version 2.2.6 updates:
- fixed package dependency installations.
- updated time series to include timepoint_end if delta_t lands on it.
- changed default 'option' parameter placeholder values.
version 2.2.5 updates:
- removed tqdm from the givernylocal package dependency list in the README.
version 2.2.4 updates:
- fixed README SciServer hyperlinks.
version 2.2.3 updates:
- added time series queries as one call to the backend code, rather than separate calls for each
timepoint in the series.
version 2.2.2 updates:
- fixed import dependency in turbulence_toolkit.py
- moved dask cluster creation to turbulence_toolkit.py to fix bottleneck with pchip interpolation and
visitor points.
- creation of givernylocal PyPI package.
version 2.2.1 updates:
- creation of giverny PyPI package.
version 2.1.12 updates:
- add package __init__.py files.
version 2.1.11 updates:
- updated giverny package metadata.
version 2.1.10 updates:
- fixed timepoints for the discrete time index datasets processed by pyJHTDB.
- isotropic4096, rotstrat4096, and channel5200
- simplified dask cluster and client code.
version 2.1.9 updates:
- reverted dask cluster and client creation, and shutdown, to when a query is made.
- preserves compute resources.
- fixes issue when running the notebook in a REST container.
version 2.1.8 updates:
- removed SciServer.Authentication package.
- changed local pickled metadata folder path.
version 2.1.7 updates:
- code cleanup.
- updated linear interpolation function names for the sabl datasets.
version 2.1.6 updates:
- simplified error checking for getCutout.
- added documentation differentiating time points and snapshot number for different datasets.
- changed first "timepoint" to 1 for snapshot number datasets in getData to keep consistent with getCutout.
- updated memory limit for the dask localcluster to handle the increased maximum cutout size.
version 2.1.5 updates:
- removed check to shutdown the cluster and client in the dask monitoring function if 1 or more
of the workers closes unexpectedly.
version 2.1.4 updates:
- fixed chunk volume calculation in getData pre-processing.
version 2.1.3 updates:
- removed python packages that no longer needed to be imported in some files.
- small performance improvement to the getData visitor read function.
version 2.1.2 updates:
- fixed dask localcluster and client to close after 5 minutes of idle time in order to free up
allocated compute resources when not in use.
- the dask client will stay running up to 2 hours during processing. all queries should finish
well under 2 hours, but if an exception occurs during processing then the cluster will automatically
close if the python kernel remains alive and no other queries are made within 2 hours.
version 2.1.1 updates:
- built getData and getCutout functions for querying all datasets.
- added stable atmospheric boundary layer ("sabl") datasets and z-axis linear step-down interpolation methods
to handle them.
version 1.2.7 updates:
- added user interface for choosing how to save output files.
- reformatted interpolation/differentiation output to match legacy code.
- added automatic metadata file generation.
- fixed stale variable states when using a dask cluster for concurrent processing.
- simplified example point distributions for interpolation queries.
- added documentation to the user-interface.
version 1.2.6 updates:
- added documentation to the user interface.
- mapped fully overlapped chunk groups for interpolation buckets together.
version 1.2.5 updates:
- added getGradient, getHessian, and getLaplacian functions.
- added .xmf file generation for getCutout.
version 1.2.4 updates:
- simplified the metadata files and switched them to be stored on a fileDB node.
- fixed HTML documentation in the notebook.
version 1.2.3 updates:
- removed h5py package.
- updated HDF5 output file formatting for getCutout.
- updated contour plotting and data retrieval functions compatability with xarray.
- updated filepathing to reflect new distribution of iso8192 zarr data on fileDB.
version 1.2.2 updates:
- removed "import mmap".
version 1.2.1 updates:
- converted the notebook to process zarr files instead of binary files.
- simplified getCutout, getPressure, and getVelocity functions.
version 1.1.4 updates:
- added 'lag4', 'lag6', 'lag8', 'm1q4', and 'm2q8' interpolation methods.
version 1.1.3 updates:
- code simplifications.
- removed worker sorting for dask parallel processing.
- updated the pickled file directory to a folder in "scratch".
version 1.1.2 updates:
- implemented memory map for reading the voxels.
- removed pre-processing step of grouping morton-adjacent voxels that overlap the user-specified box.
- simplified identifying database files the user-specified box is in.
- added creating pickle files for the database filepaths and SQL metadata.
- adjusted the contour plot axes ticks to be centered for each datapoint.
- added load balancing for dask worker assignment.
- fully vectorized the reading of the database files.
version 1.1.1 updates:
- converted output_array to xarray.
- simplied retrieving values for a datapoint (or range of datapoints) by using xarray's built-in functionality.
- added contour plot function.
- moved isotropic cube class instantiation out of the backend code.
- updated dask backend to be fully automated.
- implemented improved vectorization.
- added stride method.