12 Commits

Author	SHA1	Message	Date
Edward Langley	c2bc939608	improvise-x1ns (2nd): the preflight counts a date measure's parts ... P1 from ChatGPT Codex on #97, live at f128087 — in the change that fixed the first instance of this same defect. `preprocess` infers Numeric -> Date -> Categorical, and `encoded_width` had no date arm at all. A date string does not parse as f64, so a measure column holding dates fell through to the categorical count. `build_date` emits epoch_day/year/month/day/dow, and seven parts once any cell carries a time, so a column with ONE distinct timestamp promised 1 and materialised 5. Against a budget of 64 that is an under-count of four on a single column, in the direction that spends the allocation the preflight exists to prevent. The fix is structural rather than another remembered branch, because two instances one branch apart is the tell. `inferred_kind` reproduces `preprocess`'s own match arm for arm and returns a `Forced`; `encoded_width` matches that with NO wildcard, so a fourth variant is a compile error here instead of silently taking the categorical arm — which is exactly how Date was missed. The .missing flag moves out of the arms, since every branch owes it on the same terms. The fixture carries a time case because of the RIVAL FIX, not the bug: "a date column is 5" and "5, or 7 when any cell has a time" both repair the plain-date failure and differ only on a column containing a time. Injected, `date_parts` hard-coded to 5, the plain-date case still PASSES and the time case fails at "promised 8 ... materialised 10". Injecting the bug proves the test sees the bug; injecting the rival proves it sees the rule. Both arms compare against `data.feature_names.len()`, never against a width computed a second way. Also recorded in gp.md: the "must never under-count" line was written as a proposed kb.pl invariant while the tree already held this counter-example. An invariant generalised from the defect just fixed generalises the fix, not the code; the question that would have caught it is "what else does `preprocess` do that I have not mirrored", which is about the other function and not the bug. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01FC2jVn8Au7E9JPkw6bwXXH	2026-09-21 10:35:55 -07:00
Edward Langley	f128087999	review: correct six prose claims, four of them false ... An adversarial review of this branch found that the prose was worse than the diff — including, twice, in the exact genre the branch was congratulating itself for catching. Recorded rather than quietly rewritten, because that is the shape. Must-fix, all verified before correcting: - `fit.rs`'s comment on the exact check still read "this is the check that enforces the bound in both directions of a drift" — verbatim the sentence gp.md §14 now spends 25 lines refuting. Gotcha 48's third clause exactly: a fact stated twice is changed by grep, not by edit. I refuted the copy in the prose and left the copy in the code. - "false for eight months" was fabricated. `encoded_width` landed 2026-09-19, the sentence 2026-09-20; the repository itself is six months old. A paragraph whose point is that unchecked cost claims are checkable closed on one. - "NumericCategory was the one artefact family with no coverage" is false: it was one of three, and `ValueOfAMeasure` and `AbsentCoordinate` still have none. Corrected in both files that said it, and the remaining gap is named. The decline's stated reason was wrong, which matters most because a decline is the artefact nothing checks. It claimed one-hot is the only encoding of a category whose splits are spellable; `is_expr` is a `primary`, so an ordinal split IS spellable as a weighted indicator sum. Verified by running `parse_formula`, not by re-reading the grammar, and now pinned by `an_ordinal_split_on_a_category_is_spellable_as_an_indicator_sum` so the paragraph cannot go quietly wrong again. e8no stays declined on grounds that survive: a category's items carry insertion order rather than magnitude, and the emitted formula spells one IS per item either way, so the width is not saved. Also weakened two overclaims — "marking it a Label yields the same fit" (false: the feature count moves the RNG draw sequence, gotcha 37b) and "asserted rather than argued" (the dichotomy is argued from spell()'s exhaustive match; the test asserts one instance) — narrowed "the only coordinate predicate is is_expr" to what is true of it, corrected an uncheckable test count, named which two of the three consumers of parse_number_format are actually guarded, and moved an evidence block that had been inserted under the wrong module heading. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01FC2jVn8Au7E9JPkw6bwXXH	2026-09-21 10:26:27 -07:00
Edward Langley	e045f9871b	improvise-x1ns: the preflight counts the (missing) one-hot; improvise-e8no declined ... improvise-x1ns (found by ChatGPT Codex on PR #90, live on main at ce895ff). `encoded_width` computed `has_missing` and used it in the numeric arm only, while `build_categorical` maps every absent cell to MISSING_CAT and includes it in `distinct(&mapped)`. A forced category with N distinct present values and one record lacking that coordinate promised N features and materialised N+1. That under-reports in the direction that costs. The preflight runs before `preprocess`; the exact check on `feature_names.len()` runs after, so it bounds the error message and not the allocation the preflight exists to prevent. gp.md §14 claimed drift was "bounded in both directions by the exact check" — wrong, and wrong against the argument two paragraphs above it in the same section. That sentence is corrected rather than softened; the previous wave read this shape and reasoned it away with it. The clamp stays outside the addition, so an entirely-missing column — which `preprocess` skips — is still counted as one and the fix moves only its own case. improvise-e8no is DECLINED, on evidence rather than on appetite. `formula.pest`'s only coordinate predicate is `is_expr`, a single-value equality: no IN, no ordering over items, and a coordinate reaches an `Expr` only through IS. So the predicates a tree may state about a category are exactly its items, one at a time, and one-hot is the only encoding whose splits are spellable — the width is the size of the vocabulary, not an implementation's appetite. Ordinal encoding is already the classified refusal `Artefact::NumericCategory`, and hashing has no document name at all. The pre-zv8t "fit that succeeded" therefore either never split on the column, in which case Label is exactly equivalent and in budget, or was refused at storage time: no fit is lost that was both storable and used the column. Asserted by a_split_on_a_numeric_category_read_as_a_number_cannot_be_stored, which is also the first coverage NumericCategory has had. Both refusal messages said "a category with a distinct value per record becomes one column per record", which diagnoses the runaway shape and not e8no's. They now say a category becomes one column per distinct item, whether its items read as numbers or not — that clause being the whole of what zv8t changed. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01FC2jVn8Au7E9JPkw6bwXXH	2026-09-21 10:18:38 -07:00
Edward Langley	e50473df71	docs(gp): fitted_formula takes Features, not Overrides ... Second self-review finding in the same paragraph, and the same class: §11 said "`crate::dataset` and `fitted_formula` take the caller's `Overrides`". Only the first does. `fitted_formula` takes a `Features`, which `Features::of` read off whatever `Dataset` it was handed — the Overrides are two steps upstream. The conclusion was right and the mechanism was wrong, which is the kind of sentence that survives review because the reader checks the conclusion. Rewritten to name the actual chain. Also verified while checking it, and worth recording because it looked like a drift and is not: the two test call sites still passing `Overrides::default()` to `crate::dataset` describe the SAME dataset production now builds. In `preprocess` the `present.is_empty()` drop happens BEFORE `ov.force` is consulted, so forcing an entirely-missing column categorical changes nothing — `an_empty_category_before_a_measure_does_not_break_the_role_check`'s precondition is as strong as it reads. The other site's fixture has only Label categories, so `category_columns` is 0 and the two Overrides are literally equal. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01FC2jVn8Au7E9JPkw6bwXXH	2026-09-21 03:58:43 -07:00
Edward Langley	e6e39a03ff	review(gp): the zv8t prose claimed a cost it has, and the cost is measurable ... Self-review of 99cfb7f, of the prose rather than the diff. §11 said the forcing cost nothing — "no document that fitted before stops fitting" — reasoning that a tree splitting on a numeric category was refused as an artefact anyway. That is false. The refusal is conditioned on `used_features`, so a fit that simply never split on the numeric category SUCCEEDED, and the wide categories are exactly the ones a fit has least reason to split on. Measured with a throwaway #[ignore] probe on both sides of the fix — 100 records, a 100-item numeric `Year` beside a `Cost` that predicts `Score` on its own: pre-fix: FIT LANDS -> SUM(IF(Cost < 48.0150032043457, 970.3.., 3684.7..)) post-fix: REFUSED -> would encode 101 features from 100 records So the sentence one paragraph below improvise-wd6n's corrected "permanent refusal" was the same failure mode: a claim nothing in the suite could check, green either way. Noting it because the pattern, not the sentence, is the finding — `doctrine(prose_that_calls_something_permanent_forecloses_the_ticket)` has a sibling in prose that calls something free. The trade is still right, for a reason that is NOT "both refuse": the old success was accidental, since `Year` was encoded as a continuous number over 1900–1999 that the fit could only have used by producing a formula it would then be refused for. What is genuinely lost is the case where the user wanted neither — a wide numeric category left alone while the fit uses the measures. Filed as improvise-e8no (P3), with the acceptance criterion that the fix is a budget that can decline A COLUMN rather than the whole table, and an explicit "do not close this by making the forcing conditional on cardinality", which would be inference by another name. `a_wide_numeric_category_now_refuses_a_fit_that_used_to_ignore_it` pins both halves. Its second half is what keeps it from being a bare characterization: it marks `Year` a Label and asserts the fit then lands on `Cost`, because advice in an error message is a promise and this is the half worth asserting. Note for the record: 99cfb7f's own message carries the false claim and cannot be corrected. Per design-principles §6, the correction lives in context/ and this says so — the message is superseded by §11 and by the test-evidence entry. 1463 passing, up from 1462. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01FC2jVn8Au7E9JPkw6bwXXH	2026-09-21 03:56:47 -07:00
Edward Langley	99cfb7fb41	fix(gp): a numeric- or date-valued category is spellable (improvise-zv8t) ... `fit_measure` handed `preprocess` an `Overrides::default()`, so it never said which header columns are document categories — although the very next line told `Features::of`. `preprocess` then inferred Numeric -> Date -> Categorical from the values, so a `Year` category with items 2024/2025 became a numeric column and a category of ISO dates became a date column, and both were refused as encoding artefacts with `Year IS 2024` spellable all along. `fit_measure` now builds its `Overrides` from the same `header[..category_columns]` slice it already hands `Features::of`, forcing each to `Forced::Categorical`, so the one place the document's own answer was known is now the only place either caller reads it from. `Forced::Categorical` already existed; no formula-language change was needed. Three things follow, and the last is a cost. The date decline in gp.md §11 now means what it says. A `DatePart` could arrive over a *category* column by exactly this accident, so that decline was narrower than it read; it now covers a genuine date measure and nothing else. The `Artefact` arms stay and stay correct. `crate::dataset` and `fitted_formula` take the caller's `Overrides`, so a caller that does not force can still produce a `Numeric` role over a category column. The arm is no longer reachable from `fit_measure`, which is the only production entry point — it is not dead code, and deleting it would derive the classification from one caller's habits. Forcing widens the encoding, deliberately: a two-item `Year` is now two one-hots rather than one numeric feature, and a two-hundred-item one is refused by `MAX_FEATURES` where before it was refused one stage later as an artefact. Both refuse, so nothing that fitted before stops fitting; what changes is which guard speaks, and the width guard's message names the repair. `encoded_width` had to learn the same slice for that to hold, or it counted such a column as one feature and handed the refusal to the exact check *after* materialising the matrix — the side of the cost the preflight exists to stay on (§14). Evidence. The two spelling tests needed no fault injected: on the unfixed tree `fit_measure` refuses with the two real messages the ticket named. Each fixture gives the model one regular category, no other feature column, and a target that category alone determines — because the refusal is conditioned on `used_features`, so a fit whose champion happened to be a bare leaf would be accepted and prove nothing. Asserting `formula.raw` contains `Year IS` rather than merely `is_ok()` closes that in both directions. The preflight test did need one: dropping the `col >= category_columns` clause makes it fail by name, with the *exact* check speaking where the preflight should. The prose this corrects was itself the ticket's subject. improvise-wd6n's §11 called `Artefact::NumericCategory` a permanent refusal that "no vocabulary closes", which is false and is what filed improvise-zv8t. §11 and the gp test-evidence entry are both updated to say it is closed rather than open. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01FC2jVn8Au7E9JPkw6bwXXH	2026-09-21 03:52:42 -07:00
Edward Langley	630b84066c	fix(gp): NumericCategory is a filed gap, not a permanent refusal ... Review findings on this branch, all three in the gp commit. The prose was the worst of them: §11 called Artefact::NumericCategory a permanent refusal that "no vocabulary closes". It is not. fit_measure hands preprocess an Overrides::default(), so it never says which header columns are categories — although the next line tells Features::of — and preprocess then infers Numeric -> Date -> Categorical from the values. A `Year` category with items 2024/2025 becomes a numeric column and is refused, with `Year IS 2024` spellable all along. Forced::Categorical is the existing knob. Filed as improvise-zv8t. Prose that calls something permanent is what stops the ticket being filed, which makes this the branch's own defect class one level up. MissingFlag's words said "that measure", but its arm is not conditioned on is_category and dataset.rs writes "" for an absent coordinate, so a .missing flag over a category is reachable — the same "declined in words naming something they are not" the commit was written to fix. Now "that column". ValueOfAMeasure's tail likewise covered only the present case. The test module claimed "each artefact family" while its fixture reaches three of five; named for what it reaches. Re-injected the collapsed-reason fault after the reword: still fails, so the negative clauses still bite. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01FC2jVn8Au7E9JPkw6bwXXH	2026-09-20 22:06:34 -07:00
Edward Langley	173d9820a6	feat(gp): decline the artefact families separately, in their own words ... improvise-wd6n. Spelling::Artefact was one variant standing for five distinct families, the last five reached through a `_ =>` wildcard in spell(). The one refusal they fed said the tree "splits on a col.missing flag or a date part", so three of the five were declined in words naming something they are not. Artefact now names which family, spell() is exhaustive (a sixth FeatureRole is a compile error rather than a sixth thing that message misdescribes), and the refusal gives one clause per family. The two families the ticket asks about are declined in writing in context/code/gp.md §11, separately, because each needs different vocabulary: col.missing needs a word for absence that reaches apply_agg and eval_agg, and the date parts need a date type improvise does not have anywhere. Nothing new is emitted, so improvise-4zl7's route-then-aggregate rule is untouched. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01FC2jVn8Au7E9JPkw6bwXXH	2026-09-20 21:51:17 -07:00
Edward Langley	022e12f9b4	fit: route per record and aggregate, never route on the aggregate ... improvise-4zl7. A tree is fitted on individual records; an improvise formula is evaluated at a CellKey, and an aggregating cell used to aggregate each Expr::Ref first and evaluate the IF once on the totals. That is the same answer for a linear formula and a different one for anything that routes, so every fitted tree meant one thing on a record row and something else one row up. The user ruled: aggregate the rule over the members. The rule needed no new aggregate vocabulary. SUM(e) already means "fold e over the members" and simply did not: eval_agg specialised Expr::Ref and, for every other inner expression, summed the raw matching cells and ignored the expression entirely. It now folds — one entry per member record, the expression evaluated at that record's own coordinates. eval_agg is shared by eval_expr and eval_expr_cached, so the fold has exactly one home; each evaluator passes its own recursion in. Expr::Is(Filter) is the new arm in both evaluators: `Cat IS Item` is 1.0 when the cell's own Cat coordinate is Item, keeping the evaluator f64-only and reusing Formula::filter's own context.get/unwrap_or(false). A grammar production in primary, before ref_expr, with both operands through improvise_formula::quote's one home; IS joins the keyword guard. The emitter reads feature index -> role off preprocess's plan rather than reconstructing it: gp_tree::data::feature_roles is the one statement of the expansion order and encode_row now drives off it too. improvise_gp::Features carries one Spelling per feature — Measure, Coordinate or Artefact — decided by role and column, never by the name's shape. A one-hot over a real category is emitted as `Region IS North` and accepted; col.missing and the date parts stay Artefact and stay refused (improvise-wd6n). fit-measure on a document a category explains now lands instead of refusing, and the refusal no longer depends on which features the search happened to pick. Evidence: eight hand-computed tests in improvise-core over a four-record fixture whose Cost straddles the threshold inside each region, so route-then-aggregate and aggregate-then-route differ at every level. Two faults injected and watched to fail — reverting the fold fails 7 of 8, and a constant-false eval_is fails exactly the 3 coordinate tests. Reordering feature_roles fails both it and encode_row_replays_the_training_expansion. The fold is O(members) per evaluation per fixed-point pass and unbenchmarked: improvise-9iyv. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>	2026-09-20 11:52:27 -07:00
Edward Langley	5629f4c295	docs(gp): drop the two ```text fences; only a runnable one earns the exemption ... rustdoc does not execute a ```text fence, so `improvise-gp/src/lib.rs`'s sample formula and `fit_csv.rs`'s invocation block were prose wearing a fence and were claiming the checker's doctest exemption. Both move verbatim into context/code/gp.md (§10 and §17) and the cap now applies to both headers. What remains fenced in this unit is the two blocks promoted in 15774af, which rustdoc compiles and runs. Also takes scripts/check-comment-runs.py from fix/checker-silent-success: a file argument used to walk nothing and print success, and --check-evidence accepted any `fn`. Re-ran both per directory against the fixed checker — clean, and 49 evidence headings all name a `#[test]`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>	2026-09-20 10:05:50 -07:00
Edward Langley	15774afd7f	docs(gp-tree): cap the comment runs, and make two sketches real doctests ... Every run in crates/gp-tree is now at most two lines; the arguments are in context/code/gp.md and the bug narratives in context/test-evidence/gp.md. Two `ignore` fences were the only code blocks in this crate, and both had drifted: the `fit` sketch passed `(&ds.rows, &y, &spec, ...)`, a signature improvise-dkqn removed, and the `breed` sketch called a `rank` that has never existed. `ignore` is why nothing caught either. Both are now compiled and run doctests (+2 tests), on `fit::fit` and on `breed` respectively. `real_unit_mse` carried `run_fit`'s first two lines of doc, left behind when the two functions were reordered. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>	2026-09-20 09:51:39 -07:00
Edward Langley	2c30ef0138	docs(gp): move the gp-tree / improvise-gp reasoning into context/ ... The arguments in these two crates' comments, written out where a comment run cap cannot reach them: context/code/gp.md for the design arguments and context/test-evidence/gp.md for the bug narratives, keyed by test name. Nothing in crates/ changes yet. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>	2026-09-20 09:49:00 -07:00