Releases: lisphilar/covid19-sir
Releases · lisphilar/covid19-sir
CovsirPhy v2.14.0
Data loader
- [New] Reduce memory usage of LinelistData by not saving raw data #449
- [New] data source for tests performed and vaccinations #429
- [New] dataset of Population pyramid #56
- [New] dataset regarding metadata of Japan-specific dataset #475
- [Fix] use minimum number of tests in PCRData class for positive_rate() #441
Dataset complement
- [New] improve the way to calculate recovery period for recovered data full complement #420
Scenario analysis
- [New] Select plot columns and colors in Scenario.records() and .record_diff() #422
- [New] Relationship of OxCGRT index and parameter values (short-term prediction) #280
- [Fix] Scenario.records() create a figure with Infected/Fatal/Recovered but Confirmed included in the returned dataframe
Acknowledgement
@Inglezos, special thanks for your discussions and pull requests!
Related issues and pull requests: https://github.com/lisphilar/covid19-sir/milestone/21?closed=1
CovsirPhy v2.13.3
Data loading
- [Fix] ParserError when reading linelist data dated 18Dec2020 #442
Acknowledgement
@Inglezos, special thanks for your contributions!
Related issues and pull requests: https://github.com/lisphilar/covid19-sir/milestone/24?closed=1
CovsirPhy v2.13.2
Data loading
- [Fix] Spain and UK recovered data need improvement complements #427
- [Fix] Cleaned datasets uses too much memory #430
Acknowledgement
@Inglezos, special thanks for your discussions and a pull request!
Related issues and pull requests: https://github.com/lisphilar/covid19-sir/milestone/23?closed=1
CovsirPhy v2.13.1
Dataset complement
- Fix: Full compliment initial condition (problem with France) #421
SIR-derived models
- Fix: Example SIRFV model errors #417
Acknowledgement
@Inglezos, special thanks for your discussions and pull request review!
Related issues and pull requests: https://github.com/lisphilar/covid19-sir/milestone/22?closed=1
CovsirPhy v2.13.0
Data loader
- New: LinelistData remove incomplete records when cleaning #366
- New: Subset linelist data with outcome by LinelistData.closed() #362
- New: Records for daily tests executed #389
- New: DataLoader.jhu() replaces wrong last day recovered data in Japan internally #383
- Fix: LinelistData.subset(country) does not work when province names were registered in the records #360
Load local dataset
- New: Set province name of CountryData instance without province column #404
- Fix: JHUData created by JHUData.from_dataframe raises IndexError #406
- Docs: Load custom contry data #403
Dataset complement
- New: Calculate recovery period with linelist to complement recovery data in China during initial phase (Jan-Mar) #357
- New: Handling irregular JHU infected/recovered data for Brazil and Germany #339
- New: Show the kind of complement was done with
JHUData.subset_complement()#382 - Fix: Germany analysis - weird infected cases and predictions #274
- Fix: Complemented dataset is not used in PolicyMeasures.estimate() #373
SIR-derived models
- Fix: SIRD example estimator error #402
- Discussion: Other SIR models status - SIRX integration possibility #275
Scenario analysis
- New: Visualization of daily new cases method #367
Error handling
- New: Raise customized exceptions rather than built-in exceptions #359
Acknowledgement
@Inglezos and @baluballa, special thanks for your discussions and pull requests!
Related issues and pull requests: https://github.com/lisphilar/covid19-sir/milestone/20?closed=1
CovsirPhy v2.12.0
Data loader
- New dataset: Linelist of case reports #281
- Fix: ensure_country_name() selects random country - use countries abbreviations #356
Dataset complement
- New: Handling Netherlands missing recovered data #334
- New: UK: non-monotonic increasing of confirmed data #351
- New: UK: complement recovered data when the number of recovered cases is too small than expected #340
Acknowledgement
@Inglezos , thank you for your discussions and pull requests!
Related issues and pull requests: https://github.com/lisphilar/covid19-sir/milestone/19?closed=1
CovsirPhy v2.11.0
Supported Python version
- New: Support Python 3.6 #335
Datasets
- New: DataLoader.jhu(): not show the detailed citation list when verbose=1 #322
- New: Scenario class: handling mechanism for un-updated recovered cases data #269
S-R trend analysis
- FIx: Trends visualization - missing colored lines for Greece #309
- FIx: Trends analysis bug - wrong conditions for selecting best RMSLE #307
Parameter estimation
- New: Scenario class: scoring method of phase setting and parameter estimation #319
- Fix: Scenario.estimate(): low accuracy of parameter estimation with SIR-F model because of short timeout #291
- Fix: Greece weird death cases predictions #271
- Fix: Scenario.trend() and .estimate(): low accuracy in 0th phase #270
Scenario analysis
- New: Sceanario.history(): show actual number of cases #330
- Fix: Scenario.restrospective() does not receive keyword arguments of Scenario.estimate() #328
Acknowledgement
@Inglezos , thank you for your discussions and pull requests!
Related issues and pull requests: https://github.com/lisphilar/covid19-sir/milestone/18?closed=1
CovsirPhy v2.10.0
Datasets for Kaggle users
- New: Improve datasets preparation - setup automation #266
- New: input.py: directory and sub-directrories to save Kaggle datasets #267
- Docs: No updates are available for Kaggle dataset regarding the number of COVID-19 cases #261
Model validation
- New: Tool to confirm a new model can be used for parameter estimation #290
Scenario analysis
- Fix: Scenario.trend() phase visualization problems and runtime error #282
- Fix: Scenario.simulate(): title of the figure should use "simulated" rather than "predicted" #283
- Fix: Scenario.retriospective() raises OverflowError when separation date is near change dates #296
Discussion and documentation
- Wrong number of recovered cases for Greece and USA #268
- Training/Test dataset split - evaluation metrics & accuracy #276
- Documentation of term definition #288
Acknowledgement
@Inglezos , thank you for your discussions and pull requests!
Related issues and pull requests: https://github.com/lisphilar/covid19-sir/milestone/16?closed=1
CovsirPhy v2.9.1
CovsirPhy v2.9.0
Scenario analysis
- New: Comparing Scenario analysis estimate & simulation for the same dates #220
- New: Scenario.history_rate(): show change rates of parameter values in one figure #224
- New: Scenario.describe(): show the number of confirmed cases #250
- Fix: Scenario.add() does not show scenario names with ValueError when the end date is not correct #252
- Fix: Error codes of Scenario.add() is not clear when format of end date is not correct #253
- Fix: Stdout of Scenario.estimate() is not formated when 10th phase is included #255
- Fix: Title of figure for Scenario.history(target="Rt") should be "Reproduction number", not "Rt" #248
Global analysis
- Fix: PolicyMeasures.param_history() should be renamed as PolicyMeasures.history() #257