forked from gem/oq-engine
-
Notifications
You must be signed in to change notification settings - Fork 0
/
changelog_risklib
511 lines (447 loc) · 22.4 KB
/
changelog_risklib
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
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
## DO NOT ADD NEW ITEMS: use `debian/changelog` instead ##
python-oq-risklib (0.12.1-0~precise01) precise; urgency=low
[Michele Simionato]
* Improved the validation of NRML files
-- Matteo Nastasi (GEM Foundation) <[email protected]> Thu, 17 Mar 2016 14:29:04 +0100
python-oq-risklib (0.12.0-0~precise01) precise; urgency=low
[Michele Simionato]
* Reduced the data transfer for the average losses by a factor of 2
insured_losses is false in the event based risk calculator
* Fixed the uniform hazard spectra CSV export and used a more consistent
format
* Fixed a 32 bit/64 bit casting error for risk computations with a site model
starting from a non-empty datastore
* Discriminate the case of a missing datastore from the case of having no
permission to access it
* Saved half the space in absence of insured_losses in the event based risk
calculator
* Rewritten the CSV exporters for dmg_by_asset, dmg_by_taxo and dmg_total
* Added CSV exporters for the ground motion fields
* Stored `time_events` and `CostCalculator` in the datastore as arrays
* Made possible to generate the report even for descriptions with
non-ASCII characters
* Added a tag `riskInvestigationTime` to the exported loss maps and curves
* Made customizable the header of CSV files
-- Matteo Nastasi (GEM Foundation) <[email protected]> Wed, 02 Mar 2016 14:10:06 +0100
python-oq-risklib (0.11.0-0~precise01) precise; urgency=low
[Daniele Viganò, Matteo Nastasi]
* Updates Copyrights to 2016 and uniformized headers
[Michele Simionato]
* Raised an error early if there is a PoE=1 for classical damage
calculations
* Introduced a cutoff for minIML=0 in continuous fragility functions
* Fixed the `ses_ratio` when the number of samples is greater than 1
* Changed the order of the arguments in the commands `oq-lite show` and
`oq-lite export`
* Fixed a subtle bug: now the epsilons are correctly associated to the
assets because `asset.idx` is correctly set
* Removed the `assets_by_site` pickled array from the datastore
* Fixed the XML exporters for loss maps/curves mean and quantiles
* Now it is possible to override (most) configuration parameters with the
command `oq-lite run --param NAME1=VALUE1 NAME2=VALUE2 ...`
* Enhanced the reports generated by `oq-lite info --report` and
`oq-lite show fullreport`
* Made the check on the costs more forgiving for damage calculators
* Optimized the source model parsing
* Serialized in HDF5 format the CompositionInfo object
* Fixed the initial accumulator in the EventBasedRuptureCalculator
[Anirudh Rao]
* Removed the extra ScenarioDamage demo
* Removed unused parameters from the risk demo job files
(including `asset_hazard_distance` and `intensity_measure_types`)
[Graeme Weatherill]
* Incorporated ArbitraryMFD and YoungsCoppersmithMFD into the supported
MFDs
[Michele Simionato]
* Added a validation on the IMT for fragility functions in NRML 0.5 format
* Fixed an ordering bug in `get_scenario_from_nrml`
* Saved the site collection and the risk models in HDF5 format
* Renamed `fatalities` to `occupants` to be consistent with the
convention on the file names of the risk models
* Introduced a serialization protocol to save objects in the datastore
[Matteo Nastasi]
* Added framework to run bash scripts tests and collect associated
xunit report
* Used the python-prctl library to automatically kill processes when
the parent dies unexpectedly (linux-only)
[Michele Simionato]
* Unified the calculators classical and classical_tiling
* The epsilon_matrix is not saved anymore
* Source filtering is now on demand, depending on the parameter
`filter_sources`.
* Refactored the source filtering/splitting mechanism and improved
substantially the classical_tiling calculator
* Managed correctly the case of vulnerability functions with a different
number of loss ratios
* Now the flag `individual_curves` is honored for aggregate curves
in the event based risk calculator
* Fixed a bug breaking the compution of statistics of aggregate curves
with insured losses
* Added a validation on the cost type in the exposure
* Added a validation for missing region_grid_spacing
* Added a cross validation for the `time_event` parameter
* Added a validation for empty descriptions in fragility models
* Changed the classical_tiling calculator to split the heavy sources first
* Fixed the GMPETable GSIM and added an end-to-end test for it
* Changed the management of the seeds in the event based calculator:
now we generate a seed per each rupture
* Worked around a bug in h5py.Group.visitems in Ubuntu 12.04
* Reduced the precision of the site mesh to 32 bits
* The source model logic tree parser supports NRML 0.5
* Reduced the precision of the scenario_damage outputs to 32 bits
* Added an XML exporter for the UHS curves
* Rewritten the classical_tiling calculator in oq-lite
* Fixed a bug when exporting the hazard curves in presence of quantiles/means
* Improved the header of the exported CSV files, to make it QGIS-friendly
* Fixed a bug in the avg_losses for the event based risk calculator
* Keep exporting files, even if they are empty
* Exported information about the realizations into a csv file
* Ported the disaggregation calculator to oq-lite
* Added IMT information in the hazard curve XML exporter
* Fixed a small bug in classical_damage for steps_per_interval = 1
-- Matteo Nastasi (GEM Foundation) <[email protected]> Mon, 15 Feb 2016 11:04:31 +0100
python-oq-risklib (0.10.0-0~precise01) precise; urgency=low
[Michele Simionato]
* Ported the classical_bcr calculator to oq-lite
* Added a check on the uniform_hazard_spectra parameter depending on the IMTs
* Fixed a sorting error in the CSV exporters by asset
* Fully ported the classical_risk calculator to oq-lite
* Removed the (unused and experimental) loss fractions functionality from
the classical_risk calculator
* Added a check for the case of no assets within the region
* Fixed the check for missing taxonomies in the risk model
* Added `--profile` flag to oq-lite run
* The performance info has been moved fully in the datastore
* Saved one loss table per realization in the datastore
* Fixed a bug in the oq-lite exporter of the hazard maps
* Removed a spurious comma from the rupture tag
* Removed the epsilon_sampling feature from event_based_risk
-- Matteo Nastasi (GEM Foundation) <[email protected]> Mon, 14 Dec 2015 10:12:22 +0100
python-oq-risklib (0.9.0-0~precise01) precise; urgency=low
[Graeme Weatherill]
* Added fault geometry epistemic uncertainties to the source model logic tree
[Michele Simionato]
* Fixed the `oq-lite show --rlzs` command
* Fixed a minor bug in the source writer for nodalPlane nodes
* Added an attribute `discretization` to `areaGeometry` nodes
* Added a command oq-lite show_attrs to display the attributes of a HDF5
dataset
* Added flags `asset_loss_table` (default False) and `avg_losses`
(default True) to build the event loss table and the average losses
per asset respectively
* Added a check on the cost types read from the exposure
[Matteo Nastasi]
* Demos moved to /usr/share/openquake/risklib
[Michele Simionato]
* Added outputs `agg_curve-rlzs` and `agg_curve-stats` in the HDF5 file
* Changed the storage of the event based risk outputs in the HDF5 file
* Added a consistency check between the risk model keys in the job.ini and
the risk model files
* Scenario Risk and Scenario Damage calculators can work with a single
configuration file now
* When computing insured losses for an exposure without deductible and
insuredLimit, raise an error early
* Improved the export names of the oq-lite calculators
* Updated all the risk demos
* Changed the EventBasedRisk demo to produce loss curves
* Fixed a bug with the error management: now a critical error in the
the pre-calculation stops the whole computation and there is a single
traceback
* Improved the memory consumption and data transfer of the event_based_risk
calculator
* All data of a computation are now stored in a single calc_XXX.hdf5 file
* The ProbabilisticEventBased risk demo works with a single job.ini
* Added support for NRML 0.5 for fragility functions and deprecated NRML 0.4
* Added a ConsequenceModel parser and implemented the calculation of
consequences in the scenario_damage calculator
[Matteo Nastasi]
* Imposed version for python-h5py dependency, Ubuntu 12.04 will use
backported version from GEM repository
[Michele Simionato]
* Added a test for a bug of h5py version 2.0.1
* Added a check for specific_assets not in the exposure
* Added XML exports for the hazard curves
* Enhanced the .rst report with info about the exposure and the source model
* Changed the PerformanceMonitor to write only on the controller
* Introduced HDF5 outputs for the scenario_risk calculator
* Introduced HDF5 outputs for the scenario_damage calculator
* Added the ability to read scenario NRML files
* Silenced an ElementTree warning
* Added a test with multiple assets on the same site
* Improved the error message for invalid exposures
* Fixed a few tests on Windows
-- Matteo Nastasi (GEM Foundation) <[email protected]> Tue, 17 Nov 2015 10:57:39 +0100
python-oq-risklib (0.8.0-0~precise01) precise; urgency=low
[Michele Simionato]
* Added a --pdb flag
* Added validation to the source IDs
* It is now possible to pass arguments to the GSIMs in the logic tree file
* Added data transfer information to the CSV reports
* Fixed several bugs in the ebr calculator
* The composite source model is saved in the HDF5 file
* Added a command 'oq-lite info --report job.ini' to produce a report
on a calculation without running it
* Negative calculations IDs are recognized in the oq-lite commands
* The GMFs are saved in the HDF5 file
* Added vulnerability functions with Probability Mass Function
* oq-lite has now a 'reduce' command to reduce large computations
* The epsilon_matrix is now saved in HDF5 format
* Fixed a bug in apply_reduce when the first argument is a numpy array
* Added a functionality 'write_source_model' to serialize sources in XML
-- Matteo Nastasi (GEM Foundation) <[email protected]> Wed, 23 Sep 2015 15:17:28 +0200
python-oq-risklib (0.7.2-0~precise01) precise; urgency=low
[Matteo Nastasi]
* Packaging system improvement
[Michele Simionato]
* Fixed an ordering bug in the GSIM logic tree
-- Matteo Nastasi (GEM Foundation) <[email protected]> Fri, 05 Jun 2015 11:55:09 +0200
python-oq-risklib (0.7.1-0~precise01) precise; urgency=low
[Michele Simionato]
* Fixed an export bug in the scenario_damage calculator: now the case of
multiple assets on the same location is treated correctly
-- Matteo Nastasi (GEM Foundation) <[email protected]> Wed, 20 May 2015 09:58:07 +0200
python-oq-risklib (0.7.0-1~precise01) precise; urgency=low
[Matteo Nastasi, Daniele Viganò]
* Fixed dependencies version management
-- Matteo Nastasi (GEM Foundation) <[email protected]> Thu, 07 May 2015 13:56:09 +0200
python-oq-risklib (0.7.0-0~precise01) precise; urgency=low
[Matteo Nastasi, Daniele Viganò]
* Add binary package support for both Ubuntu 12.04 (Precise)
and Ubuntu 14.04 (Trusty)
[Michele Simionato]
* Added a check on invalid IMTs when using the JB2009 correlation model
* Made sure that the site collection is ordered by (lon, lat) in all cases
* Extended the scenario calculators so that they can manage multiple GSIMs
at the same time
* Ported the event_based calculator to oq-lite (except disaggregation)
* Introduced a DataStore class to save either generic pickled objects or
numpy arrays; if installed, it can use the h5py library
* Added defaults to the parameters in job.ini
* Extended the 'oq-lite info' command to plot some useful informations
about the source model
* Ported the classical_risk calculator to oq-lite
* Added an 'oq-lite pick_rlzs' command to display the distance of the
hazard curves from the mean, by using the RMSEP distance
* Ported the scenario_risk calculator to oq-lite
* Added a check to forbid exposures with an asset number=0
* In the tests you can compare for equality files containing numbers, possibly
ignoring the last digits
* Increased the number of SES generated by the demo
ProbabilisticEventBased from 10 to 200
* For the event based calculator, print the site index where the discrepancy
with the classical hazard curves is greater
* Now the engine is able to gunzip transparently the logic tree files too
* Added a convergency test for the hazard curves generated by the event
based calculator when the number of samples is high
* Added a script to reduce large source models and large exposure models
* Fixed a bug in the NRML writer for nodes with value=0
* Added a 'plot' command to oq-lite
-- Matteo Nastasi (GEM Foundation) <[email protected]> Thu, 07 May 2015 10:51:21 +0200
python-oq-risklib (0.6.0-1) precise; urgency=low
[Matteo Nastasi]
* Wrong oq-hazardlib version dependency fixed
-- Matteo Nastasi (GEM Foundation) <[email protected]> Fri, 27 Feb 2015 11:54:22 +0100
python-oq-risklib (0.6.0-0) precise; urgency=low
[Michele Simionato]
* Changed the event based risk demo to use fully the specific assets feature
* Implemented export of SESCollections
* Classical tiling calculator in oq-lite
* Raise an error if there are assets missing the number attribute
* Added a prefilter flag
* Moved the management of zipfiles into oq-lite
* Export lon and lat in the hazard curves
* Simplified the management of GMPE logic tree realizations
* Towards supporting Ubuntu 14.04
* Fixed error when minIML|maxIML are missing in the fragility model
* Ported the generation of UHS curves into risklib
* Better error message for duplicated IMLs
* Implemented hazard map export in oq-lite
* Create the export_dir if possible
* Moved the demos inside oq-risklib
* Now sites can be extracted from the site_model.xml file
* Fixed the oversampling bug
* Added a warning in case of oversampling
* Provided a user-friendly string representation of the class RlzsAssoc
* Added classical damage QA tests
* Add a test case for sampling two source models of different weight
* Implemented classical damage calculator
* Moved hazard maps and mean and quantile functions from the engine into risklib
* Added a check for duplicated branchset IDs
* If export_dir is not given, save on $HOME
-- Matteo Nastasi (GEM Foundation) <[email protected]> Wed, 25 Feb 2015 17:04:04 +0100
python-oq-risklib (0.5.1-0) precise; urgency=low
* consistency in version management between debian/ubuntu package and
library from git sources
-- Matteo Nastasi (GEM Foundation) <[email protected]> Thu, 18 Dec 2014 15:42:53 +0100
python-oq-risklib (0.5.0-1) precise; urgency=low
* Fixed version in the documentation
* Removed an incorrect warning
* Scenario damage tests
* Added a DamageWriter class and the expected outputs for the ScenarioDamage
QA tests
* Improve the error message when the user sets the wrong calculation_mode
* Moved get_realizations from the engine into commonlib
* Added more validation to dictionaries
* Added a new parameter complex_fault_mesh_spacing
* Ported the scenario tests to oq-lite
* Added a get_params utility
* Simplified the monitoring
* Add the QA tests data removed from the engine
* Added a forgotten .rst file
* Use shallow-clone to improve CI builds speed
* Fix documentation
* Merged commonlib inside risklib
* Move the calculation of input/output weights into commonlib
* The export_dir is now always given
* Small fixes to commonlib
* Better error message for regions that cannot be discretized
* Fixed precision
* Add a facility to read source models in commonlib
* Add an hard limit to check_mem_usage()
* Fixed test_different_levels_ok
* Removed an excessive validation
* Added a validation is_valid_hazard_curves
* Reverted the check on duplicated vulnerabilitySetID
* Improved the validation of vulnerability models
* Examples, tests and notebook for the feature of oq-lite
* Some refactoring of the classes RiskModel and RiskInput
* Basic support for risk calculators in commonlib
* Building the documentation of risklib and some refactoring
* Build the documentation of commonlib
* Support of the oq-lite command-line tool
* Converting the region_constraint into a WKT polygon in commonlib, not in
the engine
* Fixed a subtle ImportError
* Added a forgotten file
* Support for the simplification of the risk calculators
* Rewritten the RiskModel class and refactored the Workflow classes
* Loss per event per asset
* Dependency check
* Updated version of risklib
* Merged nrmllib inside commonlib
* Added get_exposure and some refactoring
* Changed the RiskModel API
* The investigationTime attribute is optional
* Use the new validation mechanism for vulnerability/fragility functions
* Reflected the change in risklib
* Fixed typo in an error message
* reversed edges to make complex surface comply with Aki & Richards conven...
* Moved the node context manager to the node library
* Better debugging representation of a node
* The maximum_distance must be mandatory in all hazard calculators
* Restored the NrmlHazardlibConverter
* Used PlanarSurface.from_corner_points
* Improved the validation on PMF
* Added a generic SourceConverter class for all sources, including the
NonParametric ones
* Added a lazy function read_nodes
* Added a "node_factory" validation facility to the node library
* Rewritten the openquake validation mechanism; now it is done in commonlib
* Support validation code for the engine
* Added a few validation functions
* Added a validation on the GSIM name
* Added MeanLossTestCase and some comments/docstrings
* Small improvements
* Logging the size of received data, to assess the stress on rabbitmq
* Parallel filtering
* Fixed the branch ordering
* Ordering the sources after splitting-filtering them
* Added an EpsilonProvider class
* Introduced get_epsilons
* Improved the splitting mechanism
* Fix an ordering on the exported GMF
* Moved parse_config from the engine to commonlib
* Given a proper ordering to BranchTuples
* Decoupled the sampling logic from the GsimLogicTree class
* Get more information about the number of realizations in full enumeration
* Small refactoring of the block-splitting routines
* Fixed the sampling in GsimLogicTree
* Small changes to support the refactoring on the engine side
* packager.sh: missing update fixed
* Risk loaders
* Added a property .imts to the RiskModel class
* The rupture seed is now visible in the XML file
* Made explicit the dependence from the getters
* GMPE logic tree fix
* Many improvements to make the SourceCollector more usable from the
command-line
* Fix for the case of empty SES
* Add a debug flag to enable set -x in packager.sh
* Improved the SourceCollector
* Fix gmf duplication
* Removed logictree.enumerate_paths
* Moved modules to manage sources and logic trees from the engine
* Minor refactoring of scientific.scenario_damage
* Reflected the API change in risklib
* Refactoring of risklib needed to solve the problem of the block size
dependence
* Remove CalculationUnit
* Removed some useless code from risklib
* Fix branch var to be compliant within the new CI git plugin
* Updates Copyright to 2014
* rupture_to_element facility
* Ci devtest
* Renamed common->commonlib
* Add setup.py
* Improved validation and some cleanup
* Decouple the gsim logic tree from the SES output and remove dead code
* Moved the tests directory inside the package
* Refactored the conversion library
* More work on the conversion/validation library
* Add loss type to risk outputs
* Pull request for the validation library
* Standard Loss map GeoJSON
* Exporting the GMF in XML should not require keeping all the tree in memory
* No unicode + StringIO in tests
* Fix parsing of isAbsolute
* Insured loss curves statistics
* Csv tools
* Fixed the streaming xml writer
* Hazard Curve GeoJSON parser
* GeoJSON map node values should be floats
* Node library
* Ruptures now have a tag attribute
* Revise insured losses algorithm
* Added an InvalidFile exception
* Add stdDevLoss field
* Compute Curve stats
* Implemented a StreamingXMLWriter
* Minor optimizations by using memoization
* Fix quantile maps computation
* Fix Asset statistics
* NRML parsers/writers should be able to accept either file paths or
file-like objects
* Refactoring: added risk workflows
* Add '__version__' to package init
* Hazard map GeoJSON writer
* GeoJSON LossMap Writer
* Remove 'ndenumerate'
-- Matteo Nastasi (GEM Foundation) <[email protected]> Wed, 10 Dec 2014 11:17:03 +0100
python-oq-risklib (0.3.0-1) precise; urgency=low
* Lxc sandbox - improved CI with sandboxed source tests (LP: #1177319)
* Refactoring: remove curve module (LP: #1174231)
* Update Event Based algorithm (LP: #1168446)
* Fix sampling in lognormaldistribution when mean = 0, covs = 0
(LP: #1167863)
* Strictly increasing vulnerability function in classical calculator
(LP: #1165076)
* Added concurrent.futures to risklib (temporary solution)
* Update average loss formula (LP: #1156557)
* Added AGPL license file
* Refactoring needed to support Structure dependent IMT in scenario damage
(LP: #1154549)
* Implemented scenario_damage and scenario directly in risklib (LP: #1154110)
* Make the risklib able to read csv inputs (LP: #1154110)
* Fix OQ Engine fails working end-to-end when there is a different number of
gmvs per site (LP: #1144388)
* Fix Insured losses computation
* Removed Asset and AssetOutput classes
* Small refactoring of the FragilityFunctions so that it is easier to
instantiate them from the FragilityModelParser in nrml (LP: #1100235)
-- Matteo Nastasi (GEM Foundation) <[email protected]> Mon, 24 Jun 2013 16:31:18 +0200
python-oq-risklib (0.2.0-1) precise; urgency=low
* Rename of the package and namespace refactoring
-- Matteo Nastasi (GEM Foundation) <[email protected]> Sat, 09 Feb 2013 10:18:32 +0100
python-oq-risklib (0.1.0-1) precise; urgency=low
* Upstream release
-- Matteo Nastasi (GEM Foundation) <[email protected]> Wed, 12 Dec 2012 17:06:39 +0100