use alloy instead of web3 whenever possible

[ppca]

This PR replaces the use of web3 crate with alloy in rpc.rs. Now we will be making respond() call on eth with alloy.
Also replaced web3 with alloy in most places except web3::signing::recover as I did not find an equal function in alloy yet.

[Copilot]

Pull Request Overview

This PR refactors the code to use alloy’s primitives instead of web3 for keccak256 hashing, except for web3::signing::recover. Key changes include:

• Replacing web3::signing::keccak256 calls with alloy::primitives::keccak256 across multiple modules.
• Updating Cargo.toml dependencies to favor alloy and related packages.
• Refactoring contract instantiation in the Ethereum client code to use alloy-based types.

Reviewed Changes

Copilot reviewed 10 out of 10 changed files in this pull request and generated no comments.

Show a summary per file

File	Description
integration-tests/src/commands.rs	Substituted web3 keccak256 call with alloy version, though without dereference operator.
integration-tests/src/actions/sign.rs	Updated keccak256 call with alloy and added dereference operator.
integration-tests/src/actions/mod.rs	Replaced keccak256 usages consistently with alloy calls using dereference.
integration-tests/Cargo.toml	Added alloy dependency.
chain-signatures/node/src/rpc.rs	Refactored Ethereum client setup to use alloy contract instance.
chain-signatures/node/Cargo.toml	Removed web3 dependency and added alloy-signer-local.
chain-signatures/crypto/src/kdf.rs	Updated hashing and address conversion using alloy.
chain-signatures/crypto/Cargo.toml	Replaced web3 with alloy workspace dependency.
Cargo.toml	Added alloy dependency to workspace with necessary features.

Comments suppressed due to low confidence (1)

integration-tests/src/commands.rs:41

• For consistency with other changes where alloy::primitives::keccak256 is dereferenced, consider using '*alloy::primitives::keccak256(PAYLOAD)' to ensure the correct type is assigned.

let payload_hashed = alloy::primitives::keccak256(PAYLOAD);

[jakmeier]

Nice, I think alloy will be a better dependency for us long-term.

LGTM

[jakmeier]

nit: I don't see many other cases of more-than-one levels of import group nesting. Not sure if anybody cares, I think we are already quite inconsistent with imports, which is fine by me. But you might want to check if you are using rust analyzer settings in your editor that generates this. I think if you set imports.granularity.group to module, instead of crate, it will be more in line with existing code.

https://rust-analyzer.github.io/book/features.html?highlight=import#import-granularity

[jakmeier]

Counter contract? Is this maybe a copy-paste leftover?

[ppca]

ah yeah. I'll remove it.

[ailisp]

lgtm

[volovyks]

Would it be hard to get rid of web3 completely? It's a heavy crate, and our Rust analyzer is already slow.

[ppca]

It's just that I can't find the equivalent of web3::signing::recover and don't want to spend time on writing code with alloy that does the same thing...