chore: Add a basic cli for generate_data.py

[dangotbanned]

Description

A pretty modest start, but adds the ability to control the scale factor and prevents data generation happening on module import.

I've wanted to be able to control the (previously global) parameter since (#805 (comment)) 😅

$ python generate_data.py --help
usage: generate_data.py [-h] [-sf SCALE_FACTOR]

Generate the data required to run TPCH queries.

options:
  -h, --help            show this help message and exit
  -sf, --scale-factor SCALE_FACTOR
                        Scale the database by this factor (default: 0.1)

                        ┌──────────────┬───────────────┐
                        │ Scale factor ┆ Database (MB) │
                        ╞══════════════╪═══════════════╡
                        │ 0.1          ┆ 25            │
                        │ 1.0          ┆ 250           │
                        │ 3.0          ┆ 754           │
                        │ 100.0        ┆ 26624         │
                        └──────────────┴───────────────┘

Related issues

• Child of chore(typing): Improve tpch typing #3420
• Started from reviewing chore(typing): Improve tpch typing #3420

[dangotbanned]

I kinda wanna add logging for the file writes.
There isn't any feedback on if any part of the script was successful.

Would be especially nice for a larger --scale-factor, where the runtime can be MUCH LONGER

[FBruzzesi]

Thanks @dangotbanned - left a few comments! It would be nice if we would finally be able to run anything regarding #805

[FBruzzesi]

You can add type=float to parse it as a float, otherwise scale_factor will be a string in main (which does not really matter as we always use it inside string formatting, but it's not what the annotation in main is declaring)

[dangotbanned]

You can add type=float to parse it as a float, otherwise scale_factor will be a string in main

Woops - I did have that but lost it when copying over somehow?

I only learned about type and default requiring the default to be a string yesterday as well 😭

[FBruzzesi]

Is there any real benefit of not having the imports on the top level in this case?

[dangotbanned]

If those imports are at the top-level, then the module will import them eagerly.

By moving them to where they are used - the script has them when needed - but importing the module (like in __init__.py) doesn't