Allow excluding directories from Databricks bundle YAML schema association in VS Code extension

[magnusasperud]

Description

When working in a Databricks Asset Bundle project, the VS Code extension automatically associates the databricks-asset-bundles.json schema with YAML files in the workspace. This default behavior is reasonable because many YAML files in a bundle repository belong to the bundle configuration.

However, in some projects the repository also contains other YAML-based configuration files that are unrelated to Databricks bundles.

Currently there does not appear to be a way to exclude specific directories from the Databricks schema association, which prevents using custom YAML schemas for those files.

---

Use Case

In our repository we maintain ingestion metadata in YAML files with a custom schema. These files are not related to Databricks bundle configuration.

Example structure:

project-root/
└─ .vscode/
│   └─ settings.json
│
├── dab-project/
│   │
│   ├── databricks.yml
│   ├── resources/
│   └── src/
│       └── metadata/
│           ├── company_code.yml
│           └── company_code_text.yml/
└── schemas/
    └── databricks-ingestion.schema.json

These metadata files should use a custom YAML schema configured via the YAML extension:

//settings.json
{
  "yaml.schemas": {
    "schemas/databricks-ingestion.schema.json": [
      "dab-project/src/metadata/*.yml",
    ]
  }
}

---

Current Workaround

The only reliable workaround is adding a schema directive to every file:

# yaml-language-server: $schema=../../../../schemas/databricks-ingestion.schema.json

This works but is not practical when managing many metadata files.

---

Proposed Improvement

Provide a way to exclude directories from the Databricks schema association, for example via a workspace setting.

Example concept:

{
  "databricks.bundle.yamlSchemaExclude": [
    "src/metadata/**"
  ]
}

This would allow the YAML extension to apply custom schemas to those directories while keeping the default Databricks schema behavior for bundle configuration files.

---

Expected Outcome

Users should be able to:

• keep the default bundle schema behavior
• define custom YAML schemas for other configuration files in the same repository
• exclude specific directories from Databricks schema association