[energidataservice] Define grid tariff filters in YAML

[jlaur]

This should make grid tariff filters more readable and therefore easier to maintain.

The current draft is a baseline providing only the same functionality as with the filters provided by code. The next step is to add support for periodization.

With the introduction of #16656, historic tariffs can be persisted. When grid company make changes to the way they publish tariffs, filters need to be adapted. This can include moving a cutoff date and querying new charge code types and notes. Without periodization of the filters, it's not possible to persist historic tariffs. Additionally, it's not possible to take into account upcoming filter changes.

[jlaur]

I'm marking this ready for review now, even though it's only the "middle" step/baseline. Tariff definitions are now in YAML, but periodization is not yet supported.

It's a bit more complicated than I initially thought. For example, when a console command requests persistence of period 01.01.2024-31.01.2024 and the tariff filter is changed per 16.01.2024, we need to calculate intersecting periods and handle them individually:

• 01.01.2024-15.01.2024
• 16.01.2024-31.01.2024

This needs to be rewritten:

openhab-addons/bundles/org.openhab.binding.energidataservice/src/main/java/org/openhab/binding/energidataservice/internal/handler/EnergiDataServiceHandler.java

Lines 483 to 498 in f37f39c

	DatahubTariffFilter filter = getDatahubTariffFilter(datahubTariff);
	DateQueryParameter start = filter.getStart();
	DateQueryParameterType filterStartDateType = start.getDateType();
	LocalDate filterStartDate = start.getDate();
	if (filterStartDateType != null) {
	// For filters with date relative to current date, override with provided parameters.
	filter = new DatahubTariffFilter(filter, DateQueryParameter.of(startDate), DateQueryParameter.of(endDate));
	} else if (filterStartDate != null && startDate.isBefore(filterStartDate)) {
	throw new IllegalArgumentException("Start date before " + start.getDate() + " is not supported");
	}
	Collection<DatahubPricelistRecord> datahubRecords = downloadPriceLists(globalLocationNumber, filter);
	ZoneId zoneId = timeZoneProvider.getTimeZone();
	Instant firstHourStart = startDate.atStartOfDay(zoneId).toInstant();
	Instant lastHourStart = endDate.plusDays(1).atStartOfDay(zoneId).toInstant();
	Map<Instant, BigDecimal> tariffMap = new PriceListParser().toHourly(datahubRecords, firstHourStart,
	lastHourStart);

And this also needs to be more sophisticated:

openhab-addons/bundles/org.openhab.binding.energidataservice/src/main/java/org/openhab/binding/energidataservice/internal/provider/ElectricityPriceProvider.java

Lines 302 to 316 in f37f39c

	private boolean downloadTariffs(DatahubPriceSubscription subscription, boolean force)
	throws InterruptedException, DataServiceException {
	GlobalLocationNumber globalLocationNumber = subscription.getGlobalLocationNumber();
	if (globalLocationNumber.isEmpty()) {
	return false;
	}
	DatahubPriceSubscriptionCache cache = getDatahubPriceSubscriptionDataCache(subscription);
	if (!force && cache.areTariffsValidTomorrow()) {
	logger.debug("Cached tariffs of type {} still valid, skipping download.", subscription.getDatahubTariff());
	cache.update();
	return false;
	} else {
	return cache.put(downloadPriceLists(subscription));
	}
	}

All in all, it requires more work and testing.

I still think using YAML even without periodization is an improvement in its own right.

[jlaur]

@lsiepel - any chance of giving this some mileage with M3? 😉

[lsiepel]

@lsiepel - any chance of giving this some mileage with M3? 😉

Would have been nice, but Kai can start the build any moment, so better to not frustrate the build proces.

[lsiepel]

Thanks LGTM