Skip to content

v1.14.5 — APRA data-correctness patch

Latest
Latest

Choose a tag to compare

View all tags
@AnthonyPuggs AnthonyPuggs released this 02 Sep 07:28
· 3 commits to main since this release
v1.14.5
c53deea

An APRA data-correctness patch. The response contract {metadata, series, observations} and all tool signatures are unchanged.

Fixed

  • APRA ratios are returned as percentages, with units — APRA stores ratios as fractions displayed with an Excel percent format, and the Key Stats sheet mixes them with ($m) dollar rows under one table with no declared unit. The parser now reads each cell's number format: percent-formatted cells come back as displayed with unit Per cent and decimals from the format, and ($m) rows get $ million. adi_capital_ratio reads 20.3 rather than 0.203 and adi_liquidity_coverage_ratio 137.8 rather than 1.378, matching the published tables. Non-percent values are untouched.
  • phi_premium_revenue and phi_claims_expense parse again — APRA's June 2026 private health insurance performance workbook renamed its data sheet from Database to Database (trailing space) with the layout otherwise unchanged. Sheet lookup now matches on stripped names.
  • APRA URL seeds refreshed — all ten bundled seed URLs had moved to newer editions and a new path, so seed resolution was returning 404 and every APRA fetch fell through to landing-page scraping. The PHI performance fallback URL is updated to the June 2026 file.

Changed

  • The nightly live suite now sweeps every curated APRA concept and fails if any returns no observation newer than 200 days, so the next upstream layout change surfaces immediately instead of being certified by a green run.

Full changelog: v1.14.4...v1.14.5

Assets 2
Loading