scipy/special-complement
| Rule | special-complement |
|---|---|
| Pack | SciPy |
| Severity | warning |
| Fix | unsafe, with xpd check --fix --unsafe-fixes |
| Library versions | every version |
1 - special.f(x) drops the digits of a small result and gives 0 once the function rounds to 1. Call the complement function that the message names.
cdf = 0.5 * (1 + special.erf(x / np.sqrt(2)))cdf = 0.5 * special.erfc(-x / np.sqrt(2))What the rule reports
Section titled “What the rule reports”The rule reports 1 plus or minus a scipy.special function, where SciPy has a function that
keeps the digits of a small result. The old form lost 3 or more digits in the following ranges:
| Old form | Digits lost for | Write instead |
|---|---|---|
1 - special.erf(x) |
x from 2.75 to 26.5 | special.erfc(x) |
1 - special.erfc(x) |
x below 6e-5 | special.erf(x) |
1 - special.gammainc(a, x), 1 - special.gammaincc(a, x) |
the tails, where the result falls below about 1e-4 | special.gammaincc(a, x), special.gammainc(a, x) |
1 - special.betainc(a, b, x) |
x from 0.32 to 0.95, with a = 2 and b = 30 | special.betaincc(a, b, x) |
1 + special.erf(x) |
x below -2.75 | special.erfc(-x) |
1 - special.ndtr(x) |
x above 4.25 | special.ndtr(-x) |
1 - special.expit(x) |
x above 8 | special.expit(-x) |
How to fix it
Section titled “How to fix it”xpd check --fix --unsafe-fixes applies the fixes of this rule. Check the result, because an unsafe fix can change what the code does.
The unsafe fix writes special.erfc(x), special.erf(x), special.gammaincc(a, x), special.gammainc(a, x), special.betaincc(a, b, x), special.erfc(-x), special.ndtr(-x), special.expit(-x). betaincc needs SciPy 1.11. erf, erfc, gammainc, gammaincc and betainc imported by name are fixed only in a file that imports the new name from scipy.special.
Message
Section titled “Message”xpd prints this message with the issue. A name that starts with $ stands for code from your file.
Subtracting $SP.$F from 1 drops the digits of a small result and gives 0 once $F rounds to 1. In the runs that lost up to all 16 digits, and $SP.$NEW was never a digit worse. Write $SP.$NEW($$$A).Browse the other rules of the SciPy pack.