v5.0.0b5

✨ IMPROVEMENTS

• The check on matching no_weight_points and datapoints in source data is now done before the data is processed any further

🐛 BUG FIXES

• lowers=false now also works when split=true is set
• fix split data if there is only one datapoint in the selected values
• fix split data if there are multiple hours with the same lowest or highest price

What's Changed

• Split fixes by @TheFes in #92

Full Changelog: v5.0.0b4...v5.0.0b5

v5.0.0b4

✨ IMPROVEMENTS

• New error messages based on the start and end time are added. There will now be a clear error if the start is after the end and if the hours setting is higher than the difference between start and end

🐛 BUG FIXES

• when a datetime or datetime string was used the microseconds were not removed from start and end which made the selection not working properly

What's Changed

• Add additional error messages by @TheFes in #89
• Remove microseconds from start and end by @TheFes in #90

Full Changelog: v5.0.0b3...v5.0.0b4

v5.0.0b3

✨ IMPROVEMENTS

• Add version to debug output

🐛 BUG FIXES

• Fix attribute finder

What's Changed

• Fix attr find by @TheFes in #88

Full Changelog: v5.0.0b2...v5.0.0b3

v50.0b2

🐛 BUG FIXES

• Fix key finder issue caused because day variable was set too late.

What's Changed

• Fix key finder by @TheFes in #87

Full Changelog: v5.0.0b1...v5.0.0b2

v5.0.0b1

✨ IMPROVEMENTS

• hours based on weight setting can also have decimal points now
• datapoints per hour are aligned (based on end, start and hours input) and the same setting is used to recalculate all 3.

🐛 BUG FIXES

• fix check if input for program was correct

What's Changed

• several updates by @TheFes in #86

Full Changelog: v5.0.0b0...v5.0.0b1

v5.0.0b0

BETA RELEASE

Version 4.2 needed a lot of bug fixes, sorry about that. In the last days a made a lot of changes based on the feedback of users. So many thanks for that. Keep sending in issues and questions, it will help to improve the macro even more!

Initially I've created the macro with only full hours in mind. The price data I worked with (Nordpool) only had hourly data, and the macro only did not really take minutes into account for the setting for the start and end period of the prices to check. In v3.0.0 support for sensors with more than one price per hour was added, and in 4.0.0 support for decimals in the hour setting was added.
Version brings the same support to the start and end parameters. You always could enter data which wasn't a whole hour, but now the data will be converted to actually use that input.

🚨 BREAKING CHANGES

• Refactor of look_ahead
  • The way look_ahead works has been reworked. look_ahead_minutes is no longer used, and therefor marked as depreciated in the macro. For now the parameter will be left in, so it will not break you current templates. In v5.0.0 the current time is checked, and converted to a start time shortly before it. To support this, the data in your source sensor is split into smaller parts (minimum of 5 minutes, so 12 parts per hour). So if you enable look_ahead and the current time is eg 11:06, the start time for the price selection will be 11:05 and there will be 12 data points per hour. If the current time is 12:18, the start time will be 12:15 and there will be 4 data points per hour.
  • One of the big advantages of this, is that it will correctly work with mode='is_now'. In previous version that would only be true in the minutes set in look_ahead_minutes (so by default in the first 5 minutes of the hour), after that the current hour would no longer be taken into account for the price data, so it would no longer be one of the cheapest hours.

✨ IMPROVEMENTS

• Both start and end will be used to split the data, and start and end will be adjusted accordingly. start will be set to the closed time before it, and end to the closed time after. In the end the data will be split according to the highest data points per hour value calculated for start, end and hours
• Refactor of look_ahead as described above.
• mode='split is now a parameter split which can be set to true or false. So you can now eg use mode='is_now', split=true to immediately see if the current time is in one of the cheapest blocks. Not all modes work the same as for split=false
  • all: the same data as in the previous version. This will also still be returned if you set mode='split'
  • start: the start time of the first block
  • end: the end time of the last block
  • min and max: the lowest and highest price in the prices found
  • time_min and time_max: the start time of the block with the lowest or highest price
  • average and weighted_average: both give the average of all prices found, as weights are not taken into account for the split setting, this will be the same for both modes
• a debug parameter is aded, if set to true it will return a dictionary with a lot of data which can be helpful for debugging, including the output based on the other parameters.
• if no input for the sensor is provided, the macro will try to find a sensor with the attributes as set for attr_today and attr_tomorrow. So basically if you use the Nordpool integration, or an integration with the same attributes it wil give the start time of the cheapest hour today if you only set use cheapes_energy_hours(). This can be helpful if for some reason the entity_id changes. Do note that it's most resource friendly to set the sensor data yourself if it not matches the defaults, so the macro doesn't have to search for it.

Again, if there are any bugs, questions, feature request, please let me know!

What's Changed

• Refactor start and end conversion to datetime by @TheFes in #77
• Add debug setting by @TheFes in #78
• Add sensor finder by @TheFes in #80
• Refactor time calculations and look_ahead by @TheFes in #81
• Refactor mode split to separate setting by @TheFes in #82
• Depreciate datapoints per hour by @TheFes in #84

Full Changelog: v4.2.3...v5.0.0b0

v4.2.3

🐛 BUG FIXES

• Fix a bug where include_tomorrow was not working when input for end was 00:00 or when there was no input at all
• Fix bug in look_ahead_minutes

v4.2.2

🐛 BUG FIXES

• Fix a bug where hours would be set to 1 in case a the program parameter was used

v4.2.1

🐛 BUG FIXES

• Fixed a bug where an additional day would be added to the end time

What's Changed

• Fix end time by @TheFes in #71

Full Changelog: v4.2.0...v4.2.1

v4.2.0

✨ IMPROVEMENTS

• added an additional parameter look_ahead_minutes to determine how many minutes can be passed in the current hour to still take it into account. This defaults to 5. So eg at 13:04 the current hour will be still taken into account, but not anymore after 13:05. Set to 60 to have it the same as it currently was, set 0 to never take the current hour into account.
• Added an additional parameter datapoints_per_hour. This is only needed when your source sensor provides more than one price per hour (eg prices per 30 minutes) and the input for hours is not a whole number (eg 2.4). If not set in such cases, it could be that the macro will give an error about a mismatch between the datapoints per hour and the weight points per hour.
• Added some important notes to the README

🐛 BUG FIXES

• Fxed a bug in how the hours were split when the input for hours wasn't a whole number
• Fixed a bug where the last datapoitn would not be taken into account when the input for hours was not a whole number.

What's Changed

• v4.2.0 by @TheFes in #70

Full Changelog: v4.1.0...v4.2.0