Commit f381ed5
Address PR #353 CI review round 6 (1 P3)
P3 - stute_joint_pretest docstring drift:
The Raises block still listed `G < _MIN_G_STUTE` as a ValueError
condition, but R5 converted that branch to a UserWarning + full-NaN
StuteJointResult return to match single-horizon stute_test and
keep the event-study workflow from crashing on staggered-filtered
small panels.
Fix: rewrote the Returns and Raises docstring blocks to describe
the actual contract. Returns now enumerates the three NaN-result
branches (small G, constant dose, any-NaN residuals / fitted) with
their warning behavior. Raises is narrowed to the genuinely-raising
conditions: empty input, key-mismatch, str-label collision, shape
mismatch, negative doses, too-few bootstrap replicates, invalid
alpha. Explicitly notes that small-G does NOT raise.
No code changes; docstring-only edit. 126 tests still pass;
black/ruff/mypy clean.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>1 parent e3f7450 commit f381ed5
1 file changed
Lines changed: 15 additions & 3 deletions
| Original file line number | Diff line number | Diff line change | |
|---|---|---|---|
| |||
1961 | 1961 | | |
1962 | 1962 | | |
1963 | 1963 | | |
| 1964 | + | |
| 1965 | + | |
| 1966 | + | |
| 1967 | + | |
| 1968 | + | |
| 1969 | + | |
| 1970 | + | |
| 1971 | + | |
| 1972 | + | |
| 1973 | + | |
| 1974 | + | |
1964 | 1975 | | |
1965 | 1976 | | |
1966 | 1977 | | |
1967 | 1978 | | |
1968 | | - | |
1969 | | - | |
1970 | | - | |
| 1979 | + | |
| 1980 | + | |
| 1981 | + | |
| 1982 | + | |
1971 | 1983 | | |
1972 | 1984 | | |
1973 | 1985 | | |
| |||
0 commit comments