Large nums as strings

[matthew1001]

Relates to hyperledger/firefly#1568

With large JSON numbers supported in FF (see hyperledger/firefly#1569 and hyperledger/firefly#1581) the FireFly UI received large numbers correctly but then lost precision when rendering them.

For example the following Data viewer:

which was rendered from the following JSON payload from FF:

...
  "datatype": {
    "name": "mattdata",
    "version": "1.0.0"
  },
  "value": {
    "id": 1000000000000000001
  }
...

This PR uses the MIT-licenses package lossless-json to replace the JSON.parse step in all fetch(...) calls. It stringifies numbers that are too large to be represented as a Javascript number e.g.

which is from the same JSON payload as above.

The parser leaves all other data untouched (i.e. it returns the same JSON as JSON.parse except for the case where a number is determined to be a number that cannot be safely represented as a Javascript number).

Here is an example of a small number, which is represented exactly as it would be before this PR:

Here is an example of a string, also represented exactly as it would be before this PR:

I think representing numbers in this way as strings is definitely preferable to showing incorrect data, but it might be useful to FF users to show some indication in UI panels that data has been manipulated to render it correctly. I'm considering that work out of scope for this fix but mentioning here for discussion.

[EnriqueL8]

Great stuff - I think just documenting that long number might appear as strings in the docs is fine for now

[EnriqueL8]

actually based on this example https://github.com/josdejong/lossless-json?tab=readme-ov-file#lossless-json couldn't we just call parse() and that's it?

[EnriqueL8]

Or is it because lower down in the react component we are actually doing some manipulation?

[matthew1001]

The problem with just calling parse() is that small numbers all end up as LosslessNumbers as well and are therefore stringified in the UI. Which I don't think we want. So the extra number parser allows us to retain Javascript numbers for smaller numbers.

[EnriqueL8]

But they don't look like stringified here

// LosslessJSON.parse will preserve all numbers and even the formatting:
console.log(stringify(parse(text)))
// '{"decimal":2.370,"long":9123372036854000123,"big":2.3e+500}'

[matthew1001]

So it's because we want the large numbers as a string (instead of a BigInt which would render in the UI as 1000000000000001n). In their example they are using their own stringify to render it as a large number without a string, but we would need to refactor a lot of the codebase to use that instead of our existing JSON serialisation logic.

[EnriqueL8]

Got caught out that stringify is actually their own function