Fix_lpr drop some households categories to 0 sample for hh_transitions

[tianxie1995]

No description provided.

[tianxie1995]

Issue: fix_lpr step will replace some households with a set mappings in order to reduce/increase workers, this process will potentially dropped some households categories to 0 which will prevent hh_transition

How to reproduce the issue:

• in year 2021, hh categories contains households

hh.query('(age_of_head < 35.0)&(cars < 2.0)&(persons >= 5.0)&(children >= 3.0)&(age_of_head >= 25.0)&(workers >= 1.0)&(workers < 2.0)&(persons < 6.0)&(cars >= 1.0)&(children < inf)&(large_area_id == 5.0)&(income < 31705.0)&(income >= 0.0)&(race_id == 4.0)')

• but after fix_lpr step, its count drop to 0

[tianxie1995]

Implemented a fix to household_transition which will append hh_seeds and person_seeds at the start of the transition so sampling will at least have baseyear seed samples.

[tianxie1995]

Year 2050 controls zero sample category:

large_area_id                       5.0
race_id                             2.0
age_of_head_min                    18.0
age_of_head_max                    25.0
persons_min                         2.0
persons_max                         3.0
children_min                        0.0
children_max                        1.0
cars_min                            1.0
cars_max                            2.0
workers_min                         2.0
workers_max                         3.0
income_min                     113184.0
income_max                    1534133.0
total_number_of_households          7.0
ob                                  0.0
Name: 2050, dtype: float64