Fim new model version field

[shawncrawley]

Refs #824

After merging PR #949 and manually deploying the code, a few issues were still found that needed to be cleaned up:

• When testing for apocalyptic flooding, it turns out that there may not be any "zero stage" records. So we now check if df_zero_stage_records is empty or not before operating on it rather than assuming it's not empty
• This enforces the dropping and recreating of the "{db_fim_table}*" tables. That shouldn't add much overhead at all and ensures that any tweaks to those schemas gets rolled out to the tables (as is the case here with the model_version column being added)
• Reorders the fim_version and model_version columns in a couple of places where they are inserted so they don't get mixed up

[RobHanna-NOAA]

I forget which tables we are dropping the "fim_version" field. Is it all of them?

[shawncrawley]

I forget which tables we are dropping the "fim_version" field. Is it all of them?

Yes, all of them on the ingest side. The ras2fim.geocurves* tables and handfim_cache.hydrotable* tables do not need to be aware of the fim_version they belong to.