feat: added option to load multiple api keys selected at random order

[Aviksaikat]

What I did

fixes: #87

How I did it

• Added option to load API Keys in random order. Have to define API keys as an array of items.

Example

# bash/zsh/sh
export WEB3_INFURA_PROJECT_ID="cxxxxxxxxxxxxx, f4xxxxxxxxxxxxx, 6xxxxxxxxxxxxx, axxxxxxxxxxxxx"

# fish
set -gx WEB3_INFURA_PROJECT_IDS "cxxxxxxxxxxxxx, f4xxxxxxxxxxxxx, 6xxxxxxxxxxxxx, axxxxxxxxxxxxx"

How to verify it

• Trust me bro ;)

Checklist

• Passes all linting checks (pre-commit and CI jobs)
• New test cases have been added and are passing
• Documentation has been updated
• PR title follows Conventional Commit standard (will be automatically included in the changelog)

[Aviksaikat]

Have to add API keys as repo secret in a variable called WEB3_INFURA_PROJECT_IDS to pass the CIs

[Aviksaikat]

Maybe we can work on the randomness better with seed or something else

[Aviksaikat]

Keeping the load_api_keys method public which will help refresh the API key.

[antazoey]

Looks pretty good! I added some small feedback if you want try and address.
Do you have any screenshots or demos showing it working?

[Aviksaikat]

[Aviksaikat]

There is an issue. It's only loading a random API key & not constructing the URI to make the API calls.

Line: provider.py#L60-61 may be the cause.

        if (ecosystem_name, network_name) in self.network_uris:
            return self.network_uris[(ecosystem_name, network_name)]

[Aviksaikat]

There is an issue. It's only loading a random API key & not constructing the URI to make the API calls.

Line: provider.py#L60-61 may be the cause.

        if (ecosystem_name, network_name) in self.network_uris:
            return self.network_uris[(ecosystem_name, network_name)]

Added a method get_new_uri to generate a new URI while keeping it backward compatible.

[antazoey]

To change the API key after connection, you'd have to reconnect

[Aviksaikat]

Do you think this requires any more corrections?

[antazoey]

this is a nice feature and we appreciate it.
I think it needs a tad more work however!

[antazoey]

LGTM
(1 typo)

[Aviksaikat]

Let me know if it's ok or not.

[antazoey]

Running pytest leads to failing tests:

[Aviksaikat]

Running pytest leads to failing tests:

Ahh maybe I missed the assertion. It must be > 0. As you need to have more than I key setup in order to be > 1

[Aviksaikat]

@antazoey to pass the tests you have to set multiple API keys like this

[antazoey]

@antazoey to pass the tests you have to set multiple API keys like this

Can you make tests add fake key for testing this feature? I don't want to have to create multiple keys to run the tests.

[Aviksaikat]

I think the project has already env variable setup. When we call reconnect it's trying to connect to the API endpoint. If you setup the env here it's all passing

With mock API keys

[antazoey]

@Aviksaikat We should be able to test the key selection logic alone without making requests to infura

[Aviksaikat]

The provider fixture is trying to make the request

[antazoey]

The provider fixture is trying to make the request

are you able to refactor?

[Aviksaikat]

The provider fixture is trying to make the request

are you able to refactor?

Got rid of the last test

[antazoey]

Looks good!, just delete the commented out part

[antazoey]

Also please address the flake8 issues!

[Aviksaikat]

Done. But to pass the existing first 2 tests the variables need to be set

[antazoey]

tests pass! but flake8 is still failing.
Also i am bit unsure of the reconnect method for a few reasons

1. it does not reconnect
2. it might not be needed - can jus use disconnect and have that clear the keys

[antazoey]

lgtm