Use identical help text for shared options

[victorlin]

from #1591 (comment)

Scope

• --metadata-id-columns
• --metadata-delimiters
• ?

Context

Help text for these arguments differ across subcommands. This is the --metadata-id-columns help text for augur merge:

  --metadata-id-columns COLUMN [COLUMN ...]
                        Possible metadata column names containing identifiers, considered in the order given. Columns will be considered for all metadata tables. Only one ID column will be inferred for each
                        table. (default: strain name)

for augur filter:

  --metadata-id-columns METADATA_ID_COLUMNS [METADATA_ID_COLUMNS ...]
                        names of possible metadata columns containing identifier information, ordered by priority. Only one ID column will be inferred. (default: ('strain', 'name'))

The augur filter version is used ~everywhere else. The augur merge version is an improvement of this from 0552f5f.

It has also propagated outside this repo to set_final_strain_name.py but I wouldn't worry about that as it's slated for removal.

Possible solutions

1. Store help text in a shared variable.

2. Use a shared parser config to add these options. Example:

   def add_input_metadata_options(parser):
       parser.add_argument("--metadata-id-columns", ...)
       parser.add_argument("--metadata-delimiters", ...)
       ...

   This might be more complicated to get right across all but helps reduce redundancy of not just the help text but also other parameters to add_argument including the option name.

For any solution, there should be a way to add an additional subcommand-specific snippet such as "Columns will be considered for all metadata tables" which is only applicable to augur merge.