Use _helper_ functions to encapsulate the model fitting logic

[aphalo]

Model fitting statistics are in pairs of 'line' and 'eq' versions. The model fitting should be identical, while the quering and further processing after model fitting differs. Except for stat_distrmix_eq() and stat_distrmix_line() which use a helper function, the code is duplicated across pairs of statistics. The same applies to stat_fit_residuals(), stat_fit_fitted()andstat_fit_deviations()`.

• [ ] Helper for stat_fit_residuals(), stat_fit_fitted()andstat_fit_deviations()`
• [x] Helper for stat_distrmix_eq() and `stat_distrmix_line()
• [ ] Helper for stat_poly_eq() and stat_poly_line() (possibly also stat_multcomp())
• [x] Helper for stat_quant_eq(), stat_quant_line() and stat_quant_band() See #25
• [ ] Helper for stat_ma_eq() and stat_ma_line()