Assuming:
- there is a "hypothetical Gregorian calendar" with same notations of the realistic (but no week days);
- therefore each date of the hypotetical calendar has the format "month day, year" like realist;
- the hypothetical calendar is equal to the realist, same conventions, except all years be leap years (ruler is "February of all years have 29 days");
- therefore each date of realistic is a valid date of the hypothetical;
- January 1, 1BC is the realist and the hypothetical Gregorian calendar epoch;
- therefore each date of realist and hypothetical can be mapped to a integer that is "days from epoch date" (not same result).
Given a date $d$ (valid date of realistic) and assuming $d_r$ that is "days lasted from epoch date to $d$ in realistic Gregorian calendar" and $d_h$ that is "days lasted from epoch date to $d$ in hypothetical Gregorian calendar", find the formula(s) that converts $d_r$ to $d_h$. Can be like
$t_1 = f_1(d_r)$
$t_2 = f_2(d_r,t_1)$
$t_3 = f_3(d_r,t_1,t_2)$
$...$
$t_n = f_n(d_r,t_1,t_2,...,t_{n-1})$
$d_h = f_{n+1}(d_r,t_1,t_2,...,t_{n-1},t_n)$
setting $f_1,f_2,...,f_n,f_{n+1}$. Don't use components of $d$ to calculate. I ask to avoid modular arithmetic and use Euclidean division notations like $\lfloor\frac{n}{d}\rfloor$ or $q(n,d)$ or $q(n/d)$ and $n-d*\lfloor\frac{n}{d}\rfloor$ or $r(n,d)$ or $r(n/d)$.
If you do more:
- converts $d_h$ to $d_r$;
- proof that each date of realistic is a valid date of the hypothetical;
- proof that convertions works.