This is the sixth 1.0 beta release.
Highlights include:
- support for automated reference persistence forecasts
- reports can now be rendered to PDF
- added reference data from NREL PVDAQ
- adjusted quality check thresholds and added new daytime flags
See below for the full list of changes.
- Added support for week-ahead persistence forecasts with :py
solarforecastarbiter.utils.get_data_start_end
and :pysolarforecastarbiter.main.run_persistence
modifications. (55
) (392
) - :py
solarforecastarbiter.pvmodel.calculate_power
and :pysolarforecastarbiter.datamodel.Site
expect atemperature_coefficient
with units of % of DC capacity/C instead of 1/C (350
) (433
) - Add :py
solarforecastarbiter.io.utils.mock_raw_report_endpoints
to mock report API endpoints used, primarily to test report serialization (441
) - Change default :py
solarforecastarbiter.validation.validator.detect_stale_values
and :pysolarforecastarbiter.validation.validator.detect_interpolated_values
window
value from 3 to 6. (124
) - :py
solarforecastarbiter.io.reference_data.common.create_forecasts
now calls :pysolarforecastarbiter.io.reference_data.common.create_persistence_forecasts
and :pysolarforecastarbiter.io.reference_data.common.create_nwp_forecasts
which create the Forecast objects from persistence and NWP templates, respectively. (375
) (455
) - Added :py
solarforecastarbiter.reference_forecasts.main.make_latest_persistence_forecasts
to support periodic generation of persistence reference forecasts. (375
) (455
)
- Reference net load forecasts provided using week-ahead persistence. (
55
) (392
) - Datamodel now supports
'net_load'
as an allowed variable. (55
) (392
) - Posting of daily validation now splits requests to avoid missing periods and limit each request to one week of data (
424
) (435
) - PDF report figures are generated instead of SVG for easy integration into PDF reports (
360
) (437
) - Added support for NREL PVDAQ sites to the reference database functions. (
397
) (438
) - Refactor :py
solarforecastarbiter.metrics.calculator
to accept and calculate for individual ProbabilisticForecast (CDFs) and ProbabilisticForecastConstantValue (single values). (445
) partially (425
) - Reports can now be rendered into PDF files via pdflatex (
319
) (439
) - Report generation via the CLI can now run through mock API endpoints to test report serialization (
405
) (441
) - Add derived quality flag masks to :py
solarforecastarbiter.validation.quality_mapping.convert_mask_into_dataframe
output. These currently include'DAYTIME', 'DAYTIME STALE VALUES', 'DAYTIME INTERPOLATED VALUES'
. (124
) - A public Docker image is available, and Github Actions use this image for continuous integration tests (
446
) - Add
LIMITS EXCEEDED
quality checks for AC and DC power, including :pysolarforecastarbiter.validation.validator.check_ac_power_limits
and :pysolarforecastarbiter.validation.validator.check_dc_power_limits
functions. (452
) (458
) - Adjusted University of Oregon Solar Radiation Measurment Laboratory (UO SRML) to include available PV power data. (
442
)
- Fix incorrect ordering of months and weekdays in metrics plots. (
428
) (430
) - Ensure data is sorted from reference data sources before slicing and posting to the API (
435
) - Fix bug with metrics.preprocessing for CDF ProbabilisticForecast reference forecasts. (
445
) - Remove UO SMRL sites from reference dataset that have not reported recent data. (
436
) (442
) - Fixed timezones for DOE RTC sites (
469
)
- Will Holmgren (
wholmgren
) - Leland Boeman (
lboeman
) - Cliff Hansen (
cwhanse
) - Tony Lorenzo (
alorenzo175
) - Justin Sharp (
MrWindAndSolar
) - Aidan Tuohy
- Adam Wigington (
awig
) - David Larson (
dplarson
)