Add functions to generate multi-row partitions with flexible size

[vponomaryov]

First step is to define following function in the prepare section of a rune script:

  pub async fn prepare(db) {
    ...
    db.init_partition_row_distribution_preset(
      "foo", ROW_COUNT, ROWS_PER_PARTITION, "70:1,20:2.5,10:3.5").await?;
    ...
  }

With this function we pre-create a preset with the foo name
and instruct it to calculate number of partitions and their rows-sizes like following:

• 70% of partitions will be of the ROWS_PER_PARTITION size
• 20% of 2.5*ROWS_PER_PARTITION
• 10% of the 3.5*ROWS_PER_PARTITION.

Then, in the target functions we can reuse it like following:

  pub async fn insert(db, i) {
    let idx = i % ROW_COUNT + OFFSET;
    let partition_idx = db.get_partition_idx("foo", idx).await? + OFFSET;
    ...
  }

As a result we will be able to get multi-row partitions in a requested size proportions.

This feature is optional to use.
Number of presets is unlimited. Any rune script may use multiple different presets for different tables.

[fruch]

I'm missing some docs on how init_partition_row_distribution_preset is used and what are the parameters arguments and such.

and an example how it gonna be used within a rune script, it's hard to see it here.

Edit: PR description should go into the code, maybe into README.md or into /docs/

[fruch]

I think you missing the case the input is above 100%

[vponomaryov]

It is 90+9 = 99%.

[fruch]

I didn't see a case like 90+15, as a negative case I've meant

I assume from what I've read, that it would fail, right ?

[vponomaryov]

Right, it will fail.

[vponomaryov]

Added one more test.

[vponomaryov]

I'm missing some docs on how init_partition_row_distribution_preset is used and what are the parameters arguments and such.

and an example how it gonna be used within a rune script, it's hard to see it here.

Edit: PR description should go into the code, maybe into README.md or into /docs/

You want me to update the README.md file in scope of this PR?
Even if we have information in the commit message?

[fruch]

I'm missing some docs on how init_partition_row_distribution_preset is used and what are the parameters arguments and such.
and an example how it gonna be used within a rune script, it's hard to see it here.
Edit: PR description should go into the code, maybe into README.md or into /docs/

You want me to update the README.md file in scope of this PR? Even if we have information in the commit message?

yes, I don't expect user to read a commit message to understand how to use something, do you ?

[vponomaryov]

I'm missing some docs on how init_partition_row_distribution_preset is used and what are the parameters arguments and such.
and an example how it gonna be used within a rune script, it's hard to see it here.
Edit: PR description should go into the code, maybe into README.md or into /docs/

You want me to update the README.md file in scope of this PR? Even if we have information in the commit message?

yes, I don't expect user to read a commit message to understand how to use something, do you ?

Done

[fruch]

LGTM

[fruch]

@vladzcloudius @gcarmin FYI,

as part of the work we need for some cases, @vponomaryov implement this new feature for latte
I might correlate with other needs you might have, and you input about it would be appreciated