Wave 2: PanelProfile outcome/dose shape extensions + autonomous-guide worked examples by igerber · Pull Request #366 · igerber/diff-diff

Add PanelProfile outcome_shape + treatment_dose extensions and autonomous-guide worked examples (Wave 2)

igerber

and

claude

committed

Address local AI review (Wave 2): exact dose-invariance, tolerance-aware integer detection

igerber

and

claude

committed

Address PR #366 CI review round 1: scope TreatmentDoseShape as descriptive-only; fix WooldridgeDiD method kwarg

igerber

and

claude

committed

Address PR #366 CI review round 2 (2 P3): align ROADMAP wording with corrected scope; cover new exports in import-surface test

igerber

and

claude

committed

Address PR #366 CI review round 3 (1 P3): add duplicate-row gate to ContinuousDiD prerequisite summaries

igerber

and

claude

committed

Address PR #366 CI review round 4 (2 P1 + 1 P2): WooldridgeDiD Poisson estimand wording + is_count_like non-negativity guard

igerber

and

claude

committed

Address PR #366 CI review round 5 (1 P1 + 1 P2): add strictly-positive-dose gate to ContinuousDiD prerequisite summaries

igerber

and

claude

committed

Address PR #366 CI review round 6 (1 P1 + 1 P2 + 1 P3): reframe ContinuousDiD prerequisite list as profile-side screening + add first_treat caveat

igerber

and

claude

committed

Address PR #366 CI review round 7 (1 P3): include non-negativity clause in CHANGELOG is_count_like description

igerber

and

claude

committed

Address PR #366 CI review round 8 (1 P3): reconcile §4.7 ContinuousDiD opening bullet with the five-check screen below it

igerber

and

claude

committed

Address PR #366 CI review round 9 (1 P2 + 1 P3): rewrite linear-OLS rationale in functional-form terms; rename §5.2 + clarify dose constancy

igerber

and

claude

committed

Address PR #366 CI review round 10 (1 P2): finish converting count-outcome rationale to functional-form framing

igerber

and

claude

committed

Address PR #366 CI review round 11 (1 P1): reframe ContinuousDiD preflight checks as standard-workflow predictions, not estimator gates

igerber

and

claude

committed

Address PR #366 CI review round 12 (1 P1 + 1 P2): remove relabeling-as-fallback wording; correct duplicate-row "fit-time stop" claim

igerber

and

claude

committed

Address PR #366 CI review round 13 (1 P1 + 1 P2 + 1 P3): drop "derive first_treat from dose" framing; add PanelProfile backward-compat defaults; fix test_continuous_did docstring

igerber

and

claude

committed

Address PR #366 CI review round 14 (1 P3): drop "standard workflow / derive first_treat from dose" wording from two test docstrings

igerber

and

claude

committed

Address PR #366 CI review round 15 (1 P1): drop "registry-documented fixes" overclaim for ContinuousDiD recoding

igerber

and

claude

committed

Address PR #366 CI review round 16 (1 P1): fix has_never_treated implementation to match documented contract on continuous panels with negative doses

igerber

and

claude

committed

Address PR #366 CI review round 17 (1 P1): split "no never-treated" vs "negative dose" branches; HAD only valid on the former

igerber

and

claude

committed

Drop autonomous-vs-full size invariant from test_full_is_largest

igerber

and

claude

committed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Wave 2: PanelProfile outcome/dose shape extensions + autonomous-guide worked examples#366

Wave 2: PanelProfile outcome/dose shape extensions + autonomous-guide worked examples#366
igerber merged 20 commits intomainfrom
agent-profile-shape-extensions

Commits on Apr 25, 2026

Add PanelProfile outcome_shape + treatment_dose extensions and autonomous-guide worked examples (Wave 2)

Address local AI review (Wave 2): exact dose-invariance, tolerance-aware integer detection

Address PR #366 CI review round 1: scope TreatmentDoseShape as descriptive-only; fix WooldridgeDiD method kwarg

Address PR #366 CI review round 2 (2 P3): align ROADMAP wording with corrected scope; cover new exports in import-surface test

Address PR #366 CI review round 3 (1 P3): add duplicate-row gate to ContinuousDiD prerequisite summaries

Address PR #366 CI review round 4 (2 P1 + 1 P2): WooldridgeDiD Poisson estimand wording + is_count_like non-negativity guard

Address PR #366 CI review round 5 (1 P1 + 1 P2): add strictly-positive-dose gate to ContinuousDiD prerequisite summaries

Address PR #366 CI review round 6 (1 P1 + 1 P2 + 1 P3): reframe ContinuousDiD prerequisite list as profile-side screening + add first_treat caveat

Address PR #366 CI review round 7 (1 P3): include non-negativity clause in CHANGELOG is_count_like description

Address PR #366 CI review round 8 (1 P3): reconcile §4.7 ContinuousDiD opening bullet with the five-check screen below it

Address PR #366 CI review round 9 (1 P2 + 1 P3): rewrite linear-OLS rationale in functional-form terms; rename §5.2 + clarify dose constancy

Address PR #366 CI review round 10 (1 P2): finish converting count-outcome rationale to functional-form framing

Address PR #366 CI review round 11 (1 P1): reframe ContinuousDiD preflight checks as standard-workflow predictions, not estimator gates

Address PR #366 CI review round 12 (1 P1 + 1 P2): remove relabeling-as-fallback wording; correct duplicate-row "fit-time stop" claim

Address PR #366 CI review round 13 (1 P1 + 1 P2 + 1 P3): drop "derive first_treat from dose" framing; add PanelProfile backward-compat defaults; fix test_continuous_did docstring

Address PR #366 CI review round 14 (1 P3): drop "standard workflow / derive first_treat from dose" wording from two test docstrings

Address PR #366 CI review round 15 (1 P1): drop "registry-documented fixes" overclaim for ContinuousDiD recoding

Address PR #366 CI review round 16 (1 P1): fix has_never_treated implementation to match documented contract on continuous panels with negative doses

Address PR #366 CI review round 17 (1 P1): split "no never-treated" vs "negative dose" branches; HAD only valid on the former

Drop autonomous-vs-full size invariant from test_full_is_largest