Adding terraform generation to openbench.

[kishorealliiita]

Summary

Adds a Terraform Generation benchmark to openbench: the model gets natural-language prompts and must produce Terraform (HCL) for two tasks (VPC + 3 subnets + 3 EC2, and S3 bucket + bucket policy). The scorer extracts .tf code blocks from the model output, runs terraform fmt, terraform init -backend=false, and terraform validate, and scores 1.0 only if init and validate succeed (no plan/apply or LocalStack). Includes dataset, eval, scorer, config entry, registry import, unit tests for extraction/dataset/scorer, and an integration test for bench eval terraform_generation.

What are you adding?

• Bug fix (non-breaking change which fixes an issue)
• New benchmark/evaluation
• New model provider
• CLI enhancement
• Performance improvement
• Documentation update
• API/SDK feature
• Integration (CI/CD, tools)
• Export/import functionality
• Code refactoring
• Breaking change
• Other

Changes Made

• Dataset (src/openbench/datasets/terraform_generation.py): load_dataset() returns a MemoryDataset of 2 samples (VPC+3 subnets+3 EC2, S3 bucket+bucket policy) with prompt text, target="pass", and task_id in metadata.
• Scorer (src/openbench/scorers/terraform_generation.py): terraform_generation_scorer() parses .tf code blocks from the last assistant message, writes them to a temp dir, runs terraform fmt, terraform init -backend=false, and terraform validate; returns Score(1.0) only if init and validate succeed, otherwise Score(0.0) (no plan/apply or LocalStack).
• Eval (src/openbench/evals/terraform_generation.py): @task terraform_generation() builds a Task with the above dataset, solver [generate()], and the custom scorer.
• Config (src/openbench/config.py): Added terraform_generation to _BUILTIN_BENCHMARKS with BenchmarkMetadata (name, description, category, tags, module_path, function_name).
• Registry (src/openbench/_registry.py): Import and re-export of terraform_generation from openbench.evals.terraform_generation.
• Tests (tests/test_terraform_generation.py): Unit tests for _extract_tf_blocks, dataset loading, and scorer (no output, no blocks, invalid Terraform, valid minimal Terraform).
  Integration (tests/integration/test_cli.py): test_basic_terraform_generation() runs bench eval terraform_generation --limit 1 with a Groq model.

Testing

• I have run the existing test suite (pytest)
• I have added tests for my changes
• I have tested with multiple model providers (if applicable)
• I have run pre-commit hooks (pre-commit run --all-files)

Checklist

• My code follows the project's style guidelines
• I have performed a self-review of my own code
• I have commented my code, particularly in hard-to-understand areas
• I have made corresponding changes to the documentation (if applicable)
• My changes generate no new warnings
• I have added tests that prove my fix is effective or that my feature works
• New and existing unit tests pass locally with my changes