I propose a high-level "reports" subpackage which will massage the sqlite data to produce reports of consumption, trends, etc. The following would be useful:

• A style profile of the current inventory, to see what styles are lacking
• Most consumed drinks/styles within a date range
• Fastest depleting drinks/styles ("first to go") within a date range

To keep things organized, the low-level API calls should continue to stay in abv/model, while high-level data processing and report generation would reside in abv/reports.