Subgraph Composition : Fix graph init for composed subgraphs

[incrypto32]

This PR enables graph init for composed subgraphs. It creates a new ipfs client to validate the source subgraph manifest

[changeset-bot]

🦋 Changeset detected

Latest commit: 36cf7e3

The changes in this PR will be included in the next version bump.

This PR includes changesets to release 1 package

Name	Type
@graphprotocol/graph-cli	Minor

Not sure what this means? Click here to learn what changesets are.

Click here if you're a maintainer who wants to add another changeset to this PR

[github-actions]

🚀 Snapshot Release (alpha)

The latest changes of this PR are available as alpha on npm (based on the declared changesets):

Package	Version	Info
@graphprotocol/graph-cli	0.96.0-alpha-20250210082839-dbaea17ad26cadc1adaaff578e4fcdf588367424	npm ↗︎ unpkg ↗︎

[cloudflare-workers-and-pages]

Deploying graph-tooling with    Cloudflare Pages

Latest commit:	36cf7e3
Status:	✅  Deploy successful!
Preview URL:	https://c2beb640.graph-tooling.pages.dev
Branch Preview URL:	https://krishna-comp-fixes.graph-tooling.pages.dev

View logs

[YaroShkvorets]

@incrypto32 could you fix the linter error?

[YaroShkvorets]

I think there are some issues with running github actions in your branch after #1919

Rebasing on top of the latest main should help.

[incrypto32]

@YaroShkvorets i just rebase on top of main and the issue persists. I cant figure out what the lint issue is, also im not able to run the lint locally. Getting these error

Oops! Something went wrong! :(

ESLint: 9.19.0

Error: Cannot read config file: /Users/krishna/dev/work/edgeandnode/thegraph/graph-tooling/node_modules/.pnpm/@theguild+eslint-config@0.13.2_eslint@9.16.0_jiti@2.4.1__typescript@5.7.2/node_modules/@theguild/eslint-config/src/index.js
Error: Failed to patch ESLint because the calling module was not recognized.
If you are using a newer ESLint version that may be unsupported, please create a GitHub issue:
https://github.com/microsoft/rushstack/issues
Referenced from: 
    at Object.<anonymous> (/Users/krishna/dev/work/edgeandnode/thegraph/graph-tooling/node_modules/.pnpm/@rushstack+eslint-patch@1.10.4/node_modules/@rushstack/eslint-patch/lib/_patch-base.js:167:19)
    at Module._compile (node:internal/modules/cjs/loader:1469:14)
    at Module._extensions..js (node:internal/modules/cjs/loader:1548:10)
    at Module.load (node:internal/modules/cjs/loader:1288:32)
    at Module._load (node:internal/modules/cjs/loader:1104:12)
    at Module.require (node:internal/modules/cjs/loader:1311:19)
    at require (node:internal/modules/helpers:179:18)
    at Object.<anonymous> (/Users/krishna/dev/work/edgeandnode/thegraph/graph-tooling/node_modules/.pnpm/@rushstack+eslint-patch@1.10.4/node_modules/@rushstack/eslint-patch/lib/modern-module-resolution.js:11:23)
    at Module._compile (node:internal/modules/cjs/loader:1469:14)
    at Module._extensions..js (node:internal/modules/cjs/loader:1548:10)
    
    ```

[YaroShkvorets]

Looks like it's passing. Good for review now?

[incrypto32]

@YaroShkvorets yes, ready for review

[YaroShkvorets]

• Remove hidden: true from --ipfs flag on line 117
• There doesn't seem to be a way to scaffold a subgraph-based subgraph in non-interactive way from command line. Is it possible?
• this command does nothing and exits with code 0 without any error: graph init test test --from-contract=0x0 --protocol=subgraph --network=mainnet

[incrypto32]

@YaroShkvorets just addressed the comments, the CI is failing but i think its not related to the code.

[YaroShkvorets]

@incrypto32 A couple of things:

• Remove hidden: true from --ipfs flag on line 117
• It always defaults to startBlock=0. I think it would be good DevEx if it defaulted to minimum startBlock from the source subgraph data sources.
• There doesn't seem to be a way to scaffold a subgraph-based subgraph in non-interactive way from command line. Is it possible?

[incrypto32]

@YaroShkvorets I just added the support for non interactive mode and addressed the other comments too.

[YaroShkvorets]

1. --from-source-subgraph is fine by me, but wouldn't --from-subgraph be more concise? "source" seems redundant.

2. If you don't specify --start-block in non-interactive mode, it goes with undefined:

❯ graph init usdt usdt --from-source-subgraph QmaNwrB6x9oiB2NZkbPsSjkCWhMT7gzH3NKYR8ESYpcyNW --protocol subgraph --network mainnet
  Generate subgraph
  Write subgraph to directory
✔ Create subgraph scaffold
✔ Initialize subgraph repository
✔ Install dependencies with yarn
✖ Failed to generate code from ABI and GraphQL schema: Command failed: yarn codegen
- Apply migrations
  Skip migration: Bump mapping apiVersion from 0.0.1 to 0.0.2
- Apply migrations
  Skip migration: Bump mapping apiVersion from 0.0.2 to 0.0.3
- Apply migrations
  Skip migration: Bump mapping apiVersion from 0.0.3 to 0.0.4
- Apply migrations
  Skip migration: Bump mapping apiVersion from 0.0.4 to 0.0.5
- Apply migrations
  Skip migration: Bump mapping apiVersion from 0.0.5 to 0.0.6
- Apply migrations
  Skip migration: Bump manifest specVersion from 0.0.1 to 0.0.2
- Apply migrations
  Skip migration: Bump manifest specVersion from 0.0.2 to 0.0.4
- Apply migrations
✔ Apply migrations
- Load subgraph from subgraph.yaml
✖ Failed to load subgraph from subgraph.yaml: Error in subgraph.yaml:

  Path: dataSources > 0 > source > startBlock
  Expected BigInt, found string:
  undefined
error Command failed with exit code 1.

[YaroShkvorets]

Pushed a commit to fix protocol validation. The rest looks good. Flag validation is getting out of hand for this command. Created an issue to address it separately: #1952

One last thing, let's add a test to init.test.ts

[YaroShkvorets]

Wonder why CI workflows are not running. Works for other PRs.