Load genetic data into coyote. This can be achieved in two ways.
- Using flags via the load command. This is similar to the legacy perl-script
- Using YAML-file via the yaml command.
This new CLI is written in python and combines all various data types that 4 different perl-scripts handled before. It also handles both DNA and RNA samples. Additionally improved case_id incrementation that was crudely handled by perl-scripts before and adds the option to update cases with new information or overwriting old information. This could be both meta-information or types of genetic variation.
Handles data types:
- CSV (gzipped)
- YAML
- JSON
- VCF
import_coyote_sample.py --help
import_coyote_sample.py load --help
import_coyote_sample.py yaml --help
- make sure update flag works for yaml import
- documentation
- better testing examples
- (SNV-indels that are not all filtered)
- fewer of other data types
- create pip requirements.txt