Change Postgres to store data in JSON

[neongreen]

Tables (categories, items, traits) should be compressed into a single categories table:

    CREATE TABLE categories (
      uid text PRIMARY KEY,           -- Unique category ID
      data jsonb NOT NULL
    );

• Change the schema. (No need to write a migration, just change the v0 schema.)
• Implement selectCategories, selectCategory, insertCategory, updateCategory.
• Benchmark how long it takes to execute selectCategories with code compiled with -O on the official database.
• Benchmark how long it takes to do updateCategory id on the biggest category with code compiled with -O on the official database.

Based on the latter two points we can decide whether we need any fine-grained select/update functions or whether we can get away with simply reading and writing everything all the time.

[neongreen]

You can use gauge for benchmarking. Using benchmark suites is inconvenient so just add a new command to the CLI.

[neongreen]

The new command should find the biggest category in the current database and then perform the benchmarks.