Changelog
Back to All

Yes Energy vintage data, client library delete, and more!

over 2 years ago by Ellery Berk

Myst Platform Release (2022-08-25)

Hello Myst Platform users,

We’ve made changes that make available Yes Energy vintage and bid close datatypes, HPO trial predictions and actuals, and quality of life improvements to you this week!

Please update your myst-alpha package at your earliest convenience to ensure you have access to this week’s updates (instructions here).

💡New Yes Energy datatypes:

  • You can now use vintage data from Yes Energy! The connector now accepts an optional parameter forecast_vintage_offset representing the vintage offset with which to query a forecast (vintage enabled) datatype. Its format is DD:HH:MM. This will query forecasts published at or before HH:MM on DD days before the natural time of the forecasts. Vintage names of pre-defined market times (e.g. BID_CLOSE) are also supported. For more information, see https://yesenergy.knowledgeowl.com/help/forecast-services.
  • You can now work with the following bid close datatypes: BIDCLOSE_LOAD_FORECAST, BIDCLOSE_SOLAR_FORECAST, DAWF - BIDCLOSE, SOLAR_COPHSL_BIDCLOSE, SOLAR_PVGRPP_BIDCLOSE, SOLAR_STPPF_BIDCLOSE, WIND_COPHSL_BIDCLOSE, WIND_FORECAST_BIDCLOSE, WIND_STWPF_BIDCLOSE, WIND_WGRPP_BIDCLOSE

🚧

Use Yes Energy vintage data instead of bid close for now

We have noticed some data discrepancies between Yes Energy’s Time Series API and what is returned by the DataSignals Cloud implementation of the Forecast Vintage API. Specifically, data requested with ‘BID_CLOSE’ forecast vintage does not always match the equivalent BID_CLOSE datatype available in the Time Series API for solar forecasts in ERCOT. Yes Energy suggests that, instead of querying with forecast vintage “BID_CLOSE”, one should query using the day and hour offset. For example, the equivalent of “BID_CLOSE” in ERCOT would be the vintage offset 01:10:00, indicating a single day prior at or before 10:00 AM.

⚡️Enhancements:

  • Once you’ve run an HPO, you can now retrieve the results (predictions and actuals) of all trials through the client library! If you’d like to retrieve the backtest results for your HPO’s best trial, you can use hpo_result = hpo.get_result() and hpo_result.best_trial.to_pandas_dataframe(). If you’d like the first trial, you can use hpo_result.trials[0].to_pandas_data_frame() – and update the index to select other trials in your HPO.
  • To smooth out the graph development process, you can now deploy graphs without policies. This is useful if you would like to fit or run your graph ad hoc (just once) without having to specify policies first. For more information on the ad hoc run and fit feature, see our previous release notes.
  • You can now delete projects using the client library by using project.delete()! You can also filter projects by title by using myst.Project.filter(title=”My Project”). Using these together allows you to delete projects by title – without having to look up project UUIDs.
    Note: There is a confirmation for delete request that requires you to input your project title. If you’d like to skip this confirmation, you can set need_confirmation to False.
  • The Jobs table on the Monitor page in the web application should load much faster due to some performance improvements we’ve made in our backend.

As always, please do not hesitate to reach out via the chat embedded in the Myst Platform or email us at [email protected] with questions or feedback.

Thank you!
Ellery and the Myst team