Phetk

- Added lxml as a required dependency during installation as it might not be preinstalled in some platforms.

***

- Updated default _All of Us_ common variant matrix table (ACAF) file path used by `.by_genotype()` method in Cohort module
with _All of Us_ updated path.
- Updated README to add some minimum VM configuration suggestions.

***

- Removed SNOMED codes from SQL query (_ehr_dx_code_query_) generating _ehr_length_, _dx_code_occurrence_count_,
_dx_condition_count_, and _age_at_last_event_ covariates in `.add_covariates()` method in Cohort module.
- This was to make it consistent with ICD event query for phecode mapping,
i.e., only ICD9CM and ICD10CM would be used as vocabulary_id for these queries.
- For _All of Us_ users, this change should affect less than 2% of covariate data previously generated
by `.add_covariates()` method from version 0.1.36 or earlier, and should not significantly change previous analysis results.