Repository navigation
docs: correct the validation tolerance, document n_Storey as inert, clean up register - #48
Merged
Merged
Conversation
State the validation tolerance as the 2.5% the validator actually enforces. Two pages said 2% while cmd/validate/main.go has always used 2.5, and data/README.md already said 2.5%. Record a fresh measurement of the Spanish variants alongside the November 2025 table, which stays as the dated record of that run. The 72 Spanish variants now pass 48 of 72, and the deviation is a uniform -2.32% offset that sits just inside the 2.5% gate, so the pass count moves without the underlying problem changing. The prose no longer quotes the stale 22.2% figure. Document n_Storey as accepted but inert. It reaches the result only through the envelope-area estimation path, gated on Code_TypeIntake_EnvelopeArea being Estimation, which no shipped TABULA variant selects; its remaining consumers are plausibility ratios ending in Check_EnvArea_ExactToEstim, which nothing reads. Align the README on PostgreSQL 15 to 17, matching getting-started.md. Remove second person, flag emoji and tick marks, and add audience front matter to the four MkDocs pages. Replace em dashes in data/README.md and five OpenAPI descriptions.
The comment on tolerancePercent claimed 2% next to a 2.5 constant, and the comment above the pass check repeated it. The constant name already states what the check does, so the second comment is removed rather than corrected.
The comment said n_Storey has less effect once real surfaces are given. It has none. Both paths it feeds are inert: the envelope-area estimation branch is gated on a code no shipped variant carries, and the plausibility ratios end in a value nothing reads.
Codecov Report✅ All modified and coverable lines are covered by tests. 📢 Thoughts on this report? Let us know! |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Documentation pass over the MkDocs pages, the arc42 chapters and the OpenAPI descriptions, plus two code comments the pass showed to be wrong.
Factual corrections
Validation tolerance.
docs/validation.mdanddocs/getting-started.mdsaid ±2%.cmd/validate/main.goenforces 2.5, anddata/README.mdalready said ±2.5%. Both pages now say 2.5%, and the two stale comments incmd/validate/main.goare corrected in their own commit.Spain. The November 2025 results table is unchanged and stays as the dated record of that run. A new "Measured 2026-09-23" subsection carries a fresh measurement of all 72 Spanish variants: 48 of 72 passing, signed mean error -2.32%, 71 of 72 computing below the reference and none above, failures between 2.50% and 3.57%, the offset flat across magnitude. Germany, Italy and France show no comparable bias on a 25-row sample each. A warning box explains why the pass count is unstable: a 2.3% offset inside a 2.5% gate, with references rounded to one decimal worth up to 2.9% at the lowest magnitudes. The prose no longer quotes the stale 22.2% figure. Tracked in #47.
n_Storeyis accepted but inert. It reaches the result only through the envelope-area estimation path, gated onCode_TypeIntake_EnvelopeArea == "Estimation"(calc_level_05.go,calc_level_06.go,calc_level_07.go), and no shipped TABULA variant carries that code. Its remaining consumers are the envelope plausibility ratios, which converge onCheck_EnvArea_ExactToEstim(calc_level_10.go:40); nothing reads that value, since level 11 takes onlyHTransmissionThermalBridgingfrom level 10. The override table indocs/api.mdno longer pairs it withh_room, which is a real lever, and the OpenAPI description and the handler doc comment now state both reasons.PostgreSQL. README said
>= 15wheredocs/getting-started.mdsays 15 to 17. Both now say 15 to 17.Register
Second person removed,
audience: developerfront matter added to the four MkDocs pages, flag emoji and tick marks removed from the country table, em dashes replaced indata/README.mdand five OpenAPI descriptions. One clause of session narrative removed from ADR-002. ADR-004, ADR-005 and ADR-006 are untouched, as isdocumentation/content/07-deployment-view.md.Licence, copyright and acknowledgement wording is unchanged throughout.
Checks
mkdocs build --strictexits 0. The OpenAPI file still parses as YAML.go build ./...,go vet ./...andgofmtare clean, and the calc and api test packages pass.