only write to s3 if data has changed

[smnorris]

Updated requirement - it is not necessary to retain all downloads indefinitely - only write download to s3 if something has happened.

A file structure something like this should work:

/MVRD/surrey/<layer>.gdb.zip                                                            
/MVRD/surrey/<layer>_changes/<layer>.gdb.zip               
/MVRD/surrey/<layer>_changes/change_report.xlsx      
/MVRD/surrey/<layer>_changes/change_report.gdb.zip

[smnorris]

Definitions of data in the changes folder:

1. Spatial layers representing the diff

• additions
• deletions
• modifications - geometry only
• modifications - attribute only (with old/new attribute diff)
• modifications - geometry + attribute (with old/new attribute diff)

All modification layers to use the latest geometries.

2 Report summarizing changes with columns:

• jurisdiction
• layer
• record_count_old
• record_count_new
• record_count_difference
• record_count_difference_pct
• n_unchanged
• n_deletions
• n_additions
• n_modified
• n_modified_spatial_only
• n_modified_spatial_attributes
• n_modified_attributes_only