awattar-backtesting.github.io

A tool to visualize your electricity usage with aWATTar

Run Locally

Just start index.html, but make sure to allow CORS from local machine. Therefore you either run your own webserver or just start chrome with disabled web security: chrome.exe --user-data-dir="C://Chrome dev session" --disable-web-security

Calc export tariffs

Export tariffs are automatically detected if the data is provided accordingly by the network operator.

Netz OÖ

Change the line

"Date";"kWh";"kW";"Status"; to "Date";"Feed-in kWh";"kW";"Status"; to use the feed-in data.

Tests

The calculation pipeline runs in Node as well as in the browser, which lets us test it without a headless browser.

npm ci
npm test          # one-shot
npm run test:watch

CI runs the same suite via .github/workflows/test.yml.

Layout

• docs/calc/pipeline.js — runPipeline({ bytes, h0Sheet, marketdata, onWarning }) is the single entry point: bytes in, aggregated daily / monthly buckets out. Pure orchestration; the only side-effecting dependency is marketdata.addDay, which is injected by the caller.
• docs/marketdata.js — Marketdata takes a pluggable fetcher. The browser uses createBrowserFetcher (HTTP /cache60/... then aWATTar API fallback).
• tests/lib/node-fetcher.js — Node-side fetcher; reads docs/cache60/<ts> from the filesystem. Tests never hit the live aWATTar API.
• tests/lib/runtime.js — shared fixtures: loadH0Sheet(), newMarketdata(), readSample(name), plus the samplesDir / cacheDir paths.

Test suites

tests/smoke.test.js — runs every file in samples/ through runPipeline and asserts each one resolves with ok: true and produces at least one day of aggregated output. Samples that currently fail are tracked in the knownFailures map; entries there must continue to fail with the documented message — if a known failure starts succeeding, the test fails so the entry can be removed.

tests/golden.test.js — locks down byte-for-byte aggregation output for a handful of representative samples covering distinct preprocessing paths (provider preamble stripping, plain CSV, TINETZ date normalization, einspeisung / feedin, binary .xls, .xlsx with stripXls fixups). Decimal values are formatted to fixed precision so the snapshots are stable across platforms. Refresh after intentional changes with npx vitest run -u.

Adding a new sample

1. Drop the file into samples/.
2. npm test — the smoke suite picks it up automatically.
3. If it fails for a real reason, fix the parser; if it exposes a pre-existing bug you don't want to fix yet, add it to knownFailures in smoke.test.js with a substring of the error/reason message.
4. Optional: add it to goldenSamples in golden.test.js if it covers a code path no other golden touches, and run npx vitest run -u.