Demo: September 18

Agenda


Import big picture

What are we doing?
  1. We are importing current client records and their associated property records, from the Access database that runs PAMS, who have an active appeal on a property, for 2020 appeals, who have not submitted the newly required RAVA.
  2. Piotr will create scripts that will generate the necessary CSVs for import in Nov/Dec, and import other counties in bulk later. We want to do as many as possible in December, when 95% of the records are created.
  3. Rought estimated 3.2k cient records (per Piotr).
  4. We’re building a system that can regularly receive an import of new records via CSV, as well as “one offs” (completed).
Why are we doing this?
  1. There are too many records to create them all by hand.
  2. A future application-to-application inteface will reuse the import logic anyway.
  3. Risk reduction: clients attest to these facts that SC builds its case upon.
What is the big picture workflow?
  1. Piotr runs queries and generates CSVs.
  2. P. uploads client, porperty CSVs to the web app.
  3. the webapp parses the CSVs, checks for errors, rejects if invalid
  4. Show actions to be taken? Have actions approved by importer?
  5. User accounts, property accounts created “as needed”
  6. Emails — perhaps a pending queue, that can send over time?
Discuss: what does it mean to import a record?

Launch version 1: “UPSERT”

  1. If the client record does not yet exist (on the basis of a primary key) create it;
  2. If the client record does exist… update it?

Future version 2:

What if the user updates—their name, email? What’s the single source of truth?

Demo: import client records

Deep dive: technical Q&A session

Decisions made:

Post meeting notes