Files
MeterVault/src/Infrastructure/Normalization/AnalysisDataWriter.cs
T
Florian Schmidt 8940ef25c3
ci / build-test (push) Successful in 2m31s
Analysis: one selected period, one set of numbers, on every page
The dashboards told several stories at once. Overview asked for full
calendar years, meter detail for a fixed 12-month window that was really
13, Trends for 24 months with an Apply button, and the energy pages for
60. Each page derived "today" from UTC, so the first hours of a local day
belonged to yesterday. A missing tariff, a month nobody measured and a
genuine zero all rendered as 0. And a virtual meter -- the one thing the
spreadsheet leans on hardest -- was excluded from analysis outright:
MeterPeriodService returned null for it and the page offered a flow
diagram instead.

docs/DASHBOARD_ANALYSIS_CHANGE_BRIEF.md is the work order. Every choice it
left open is settled in docs/ANALYSIS_IMPLEMENTATION_NOTE.md as D-01..D-58
plus amendments A-01..A-30; code, tests and release notes cite those ids.

The analysis layer

Core/Analysis holds the pure rules: period presets resolved once in the
instance zone into a local date range and a half-open UTC range, bucket
plans, calendar-unit comparisons, coverage runs with a resolution class,
normalized quantities and units, the totals policy, the virtual formula
parser/validator/evaluator, and the cost calculator. "Now" comes from
TimeProvider; services never read the clock.

Normalization now writes, in the same transaction as consumption and by
diff, per-meter rollups by local day and month plus coverage runs and a
rollup state (AnalysisDataWriter). AnalysisReader answers a request from
those tables -- month rollups for month and year buckets, day rollups
otherwise, at most two partial edge days from consumption -- and
CostReader prices the result month by month. Pages, /api/v1 and the CSV
export read nothing else. The unused continuous aggregates are dropped.

The reader's statement count per request is constant whether it covers one
meter or a thousand. On a synthetic 1,000-meter, ten-year instance the
brief's target request (100 meters, ten years, monthly) takes 374 ms
against a two-second target, and the Overview went from 48,244 SQL
statements per load to 205.

Missing is not zero

Every bucket carries a status -- available, partial, missing, unresolved,
invalid, pending -- derived from coverage, never from the amount, with
provenance and a reason code beside it. A true zero is a number and a bar
on the baseline; an unknown bucket is a gap that says why; a month whose
data only exists monthly says so instead of inventing daily detail; a
scope with no tariff says "not priced" instead of 0. Rows whose interval
closes after now are reported separately rather than counted.

Virtual meters are analysis subjects

A virtual meter stores a canonical definition -- expression over m<id>
references, result kind, unit and cost rule -- validated on save and on
read for syntax, unknown or self references, loops and unit/kind rules.
It is evaluated on read from its sources' rollups over their joint
coverage: a missing source makes the bucket missing, an observed zero is
a valid input, a non-finite result is invalid with its dependency path,
and the page lists each source's contribution. Topology links are
topology only and never rewrite a saved calculation; expression-less
meters from older installs are converted once at startup. The editor has
Sum, Difference and Advanced modes with a live preview.

Totals and the bill

Per energy type the totals policy separates use, grid import, export,
generation and runtime, marks breakdown meters as breakdowns and virtual
meters as views, and never adds across units. The bill follows it: grid
import where there is one, separately priced subsections at their own
price, feed-in only on export meters, standing charges once per scope per
local day, manual costs once on their start day, categories as
non-overlapping covers whose composition reconciles to the bill. The
seeded demo's yearly totals now match the spreadsheet.

Pages and navigation

The period lives in the URL and every page reads the same contract, so a
link, a reload and the browser's Back button keep it. Shared components
carry it: page header with breadcrumbs, period toolbar, theme-aware chart
with an accessible table beside it, metric cards, comparison and
availability states, attention items that each link to the one action
that fixes them. Meter detail leads with an Analysis tab and resolves its
tabs by key; the energy page has Overview, History, Flow and Meters; the
old cost-only Trends page is a general Analysis page over portfolio, type,
category, meter or a meter comparison. Records tabs are paged server-side
instead of showing the latest 200. Everything is English and German,
light and dark, down to 360px.

Some figures change on purpose; docs/RELEASE_NOTES.md lists each one and
what the first start after the update does (it rebuilds all analysis data
before the web server listens). docs/SDD.md and CLAUDE.md describe the
system as it now is.

Tests: 1,733 Core and 746 integration, all green, plus an opt-in
performance suite with a synthetic 1,000-meter generator.
2026-09-20 10:29:13 +02:00

214 lines
8.9 KiB
C#

using MeterVault.Core.Analysis;
using MeterVault.Core.Analysis.Rollups;
using MeterVault.Infrastructure.Persistence;
using MeterVault.Infrastructure.Persistence.Analysis;
using Microsoft.EntityFrameworkCore;
using Microsoft.EntityFrameworkCore.ChangeTracking;
namespace MeterVault.Infrastructure.Normalization;
/// <summary>
/// Writes one meter's analysis tables (D-12) — day and month rollups, coverage runs and rollup state — by diff
/// against what is stored: a row that is still right is not touched, a changed one is updated in place, a new
/// one added and a vanished one removed. Stages the changes; the caller's <c>SaveChanges</c> writes them in its
/// transaction, together with the meter's consumption.
/// </summary>
/// <remarks>
/// <para>
/// A recompute runs on every live reading, and almost all of a meter's history comes out of it unchanged. A
/// diff keeps that to the rows the new reading actually moved — usually the current day and month and the
/// last coverage run — instead of rewriting years of rows each time.
/// </para>
/// <para>
/// The change tracker, not only the database, is the truth here: one context may recompute a meter several
/// times before it saves (an import touching a meter twice, a worker ingesting two readings in one scope).
/// Rows staged by an earlier pass are still tracked — added, modified or deleted — and the diff builds on
/// them rather than tripping over their keys.
/// </para>
/// </remarks>
internal sealed class AnalysisDataWriter(MeterVaultDbContext db)
{
/// <summary>Every analysis table is keyed by meter first, under this property.</summary>
private const string MeterIdProperty = nameof(MeterRollupState.MeterId);
private readonly MeterVaultDbContext _db = db;
/// <summary>
/// Stages the meter's rollups and coverage, and its state row. <see cref="MeterRollupState.BuiltAt"/> moves to
/// <paramref name="now"/> only when something about the meter's analysis data changed.
/// </summary>
/// <returns>True when any row was added, changed or removed.</returns>
public async Task<bool> WriteAsync(
int meterId,
MeterRollups rollups,
IReadOnlyList<CoverageRun> coverage,
MeterRollupState state,
DateTimeOffset now,
CancellationToken cancellationToken)
{
ArgumentNullException.ThrowIfNull(rollups);
ArgumentNullException.ThrowIfNull(coverage);
ArgumentNullException.ThrowIfNull(state);
// Listing tracked entries detects changes across the whole context by default — every consumption row
// the recompute just staged, several times over. Nothing here depends on it: the diff compares values
// itself, and SaveChanges detects the edits it makes.
var autoDetect = _db.ChangeTracker.AutoDetectChangesEnabled;
_db.ChangeTracker.AutoDetectChangesEnabled = false;
try
{
return await WriteCoreAsync(meterId, rollups, coverage, state, now, cancellationToken).ConfigureAwait(false);
}
finally
{
_db.ChangeTracker.AutoDetectChangesEnabled = autoDetect;
}
}
private async Task<bool> WriteCoreAsync(
int meterId,
MeterRollups rollups,
IReadOnlyList<CoverageRun> coverage,
MeterRollupState state,
DateTimeOffset now,
CancellationToken cancellationToken)
{
var changed = false;
var days = await CurrentAsync(_db.ConsumptionRollups, meterId, cancellationToken).ConfigureAwait(false);
changed |= Sync(
days, rollups.Days, d => (d.Day, d.Kind), b => (b.Start, b.Kind),
(row, bucket) => row.Assign(bucket), bucket => ConsumptionRollup.From(meterId, bucket));
var months = await CurrentAsync(_db.ConsumptionRollupMonths, meterId, cancellationToken).ConfigureAwait(false);
changed |= Sync(
months, rollups.Months, m => (m.Month, m.Kind), b => (b.Start, b.Kind),
(row, bucket) => row.Assign(bucket), bucket => ConsumptionRollupMonth.From(meterId, bucket));
var runs = await CurrentAsync(_db.MeterCoverage, meterId, cancellationToken).ConfigureAwait(false);
changed |= Sync(
runs, coverage, r => r.SpanFrom, r => r.From.ToUniversalTime(),
(row, run) => row.Assign(run), run => MeterCoverageRun.From(meterId, run));
var states = await CurrentAsync(_db.MeterRollupStates, meterId, cancellationToken).ConfigureAwait(false);
var entry = states.SingleOrDefault();
if (entry is null)
{
_db.MeterRollupStates.Add(new MeterRollupState
{
MeterId = meterId,
Revision = state.Revision,
Zone = state.Zone,
NormalizedUnit = state.NormalizedUnit,
Kind = state.Kind,
BuiltAt = now.ToUniversalTime(),
});
return true;
}
var stored = entry.Entity;
var resurrected = entry.State == EntityState.Deleted;
if (resurrected)
{
entry.State = EntityState.Modified;
}
var stateChanged = resurrected
|| stored.Revision != state.Revision
|| !string.Equals(stored.Zone, state.Zone, StringComparison.Ordinal)
|| !string.Equals(stored.NormalizedUnit, state.NormalizedUnit, StringComparison.Ordinal)
|| stored.Kind != state.Kind;
if (changed || stateChanged)
{
stored.Revision = state.Revision;
stored.Zone = state.Zone;
stored.NormalizedUnit = state.NormalizedUnit;
stored.Kind = state.Kind;
stored.BuiltAt = now.ToUniversalTime();
}
return changed || stateChanged;
}
/// <summary>
/// The meter's rows as they stand for this context: the stored rows (tracked, so identity resolution hands
/// back anything an earlier pass already holds) plus rows an earlier pass added and has not saved, minus the
/// ones it removed, which keep their entry (state Deleted) so they can be brought back.
/// </summary>
/// <remarks>
/// A tracked row the database no longer has — removed behind the tracker's back, as <c>ExecuteDelete</c> or a
/// cascade does — is detached, as the consumption rows are: updating or deleting it would fail the save.
/// </remarks>
private async Task<List<EntityEntry<T>>> CurrentAsync<T>(DbSet<T> set, int meterId, CancellationToken cancellationToken)
where T : class
{
var stored = await set.Where(e => EF.Property<int>(e, MeterIdProperty) == meterId)
.ToListAsync(cancellationToken).ConfigureAwait(false);
var inDatabase = new HashSet<T>(stored, ReferenceEqualityComparer.Instance);
var entries = _db.ChangeTracker.Entries<T>()
.Where(e => e.State != EntityState.Detached && (int)e.Property(MeterIdProperty).CurrentValue! == meterId)
.ToList();
foreach (var stale in entries.Where(e => e.State != EntityState.Added && !inDatabase.Contains(e.Entity)))
{
stale.State = EntityState.Detached;
}
return [.. entries.Where(e => e.State != EntityState.Detached)];
}
/// <summary>Reconciles the tracked rows of one table with the desired values; true when anything changed.</summary>
private bool Sync<TEntity, TDesired, TKey>(
List<EntityEntry<TEntity>> current,
IEnumerable<TDesired> desired,
Func<TEntity, TKey> entityKey,
Func<TDesired, TKey> desiredKey,
Func<TEntity, TDesired, bool> assign,
Func<TDesired, TEntity> create)
where TEntity : class
where TKey : notnull
{
var changed = false;
var byKey = current.ToDictionary(e => entityKey(e.Entity));
var wanted = new HashSet<TKey>();
foreach (var item in desired)
{
var key = desiredKey(item);
wanted.Add(key);
if (!byKey.TryGetValue(key, out var entry))
{
_db.Add(create(item));
changed = true;
continue;
}
if (entry.State == EntityState.Deleted)
{
// Removed by an earlier pass that has not been saved: the row is wanted again, so it stays
// and is updated with whatever it holds now.
assign(entry.Entity, item);
entry.State = EntityState.Modified;
changed = true;
continue;
}
changed |= assign(entry.Entity, item);
}
foreach (var (key, entry) in byKey)
{
if (wanted.Contains(key) || entry.State == EntityState.Deleted)
{
continue;
}
// An added row is simply dropped from tracking; a stored one is deleted.
_db.Remove(entry.Entity);
changed = true;
}
return changed;
}
}