Deployment debug tab on add order

[findolor]

Closes #867

Motivation

See issue: #867

Solution

• Decoupled the debug logic from run scenario button. It now fetches the correct data after each keystroke using a debounced function
• Added a visual component similar to order quotes

Checks

By submitting this for review, I'm confirming I've done the following:

• made this PR as small as possible
• unit-tested any new functionality
• linked any relevant issues or PRs

[hardyjosh]

So @findolor this looks great on the happy path but i found a couple of issues:

1. Simple one - if there's only one thing on the calculate-io stack, the component breaks. I assume because it's looking for [0] and [1] and only [0] is there. If this is the case it should just say there isn't enough on the stack to get both a max amount and ratio.
2. Another simple one - if ratio is 0 i get this error

3. For these tokens I'm getting "UNKNOWN/UNKNOWN" as both pairs?

tokens:
  polygon-usdt:
    network: polygon
    address: 0xc2132D05D31c914a87C6611C10748AEb04B58e8F
    decimals: 6
  polygon-ioen:
    network: polygon
    address: 0xd0e9c8f5fae381459cf07ec506c1d2896e8b5df6
    decimals: 18

orders:
  polygon:
    orderbook: polygon
    inputs:
      - token: polygon-ioen
      - token: polygon-usdt
    outputs:
      - token: polygon-ioen
      - token: polygon-usdt

4. If there's any error at all with the rainlang or yaml the spinner spins indefinitely and no error is shown. We need to show the error, here's a super rough example

Buuuut, only some pairs may have an error. For example this one should only show an error for one of the pairs, but an error for any pair means the whole call fails. The other thing is we need to show partial traces, so if there is an error we still need the trace up until the error, as we do with quotes.

[hardyjosh]

If I have a really long value (in this case it'smax-value()), then the hex version gets pushed off the side and i can't scroll.

might need an overflow-x-scroll wrapper on the table.

[hardyjosh]

something is weird where the older error isn't being removed when new information comes through

[hardyjosh]

still getting UNKNOWN/UNKNOWN for the pairs

[hardyjosh]

See comments for manual testing feedback

[hardyjosh]

@findolor i wonder if it would be good to use tanstack for this.

at the moment if you go to another tab, then back to the debug tab you have to wait for it to load again, even if you haven't changed any of the text. using tanstack would get us caching so it would load instantly.

[hardyjosh]

@findolor we're still missing partial stacks - if you get a revert for some quote you can still load up whatever stack was produced before the revert. You can see how i implemented that, i thought it would be the same here.

So for example, in the below image i'd expect to be able to load up the stack and see 14, because that was produced before the revert.

[hardyjosh]

@findolor can we have the block paused by default? when the block is constantly updating it has to get a new fork every time which makes it quite slow to produce the debug. but if it's paused it's much faster (same fork every time so it's all local).

[hardyjosh]

@findolor also - i know we didn't want to upgrade anything to do with charting in this PR, but let's at least rename the "Run all scenarios" button to "Generate charts"