feat: add deterministic naming for Qdrant collections #7943

roomote · 2025-09-12T22:06:04Z

This PR attempts to address Issue #7940 by implementing deterministic naming for Qdrant collections.

Problem

RooCode's Codebase Indexing currently creates Qdrant collections based on absolute file paths, causing:

Different collections for Git worktrees (each worktree has a different path)
Inability to share indexes between developers working on the same repository
Duplicated indexing work and storage

Solution

This PR implements a three-tier naming strategy for Qdrant collections:

Custom collection names via .roo/codebase-index.json configuration file
Git repository-based naming using the remote URL for deterministic naming across worktrees and developers
Workspace path fallback for non-git repositories (original behavior)

Key Changes

Added support for custom collection names through .roo/codebase-index.json
Modified QdrantVectorStore to use git repository URL for deterministic collection naming
Implemented URL normalization to ensure consistent hashing across different git URL formats (SSH, HTTPS, etc.)
Added comprehensive test coverage with 11 new test cases
Improved error handling and added graceful fallbacks

Benefits

✅ Same collection name across Git worktrees
✅ Consistent indexing between developers working on the same repository
✅ Reduced storage usage and indexing time
✅ Backward compatible with existing workspace-based collections
✅ Flexible configuration options for custom naming

Testing

All existing tests pass
Added comprehensive test coverage for new functionality
Tested with different git URL formats and edge cases

Fixes #7940

Feedback and guidance are welcome!

Important

Introduces deterministic naming for Qdrant collections using custom names, git URLs, and workspace paths, with comprehensive testing and improved error handling.

Behavior:
- Implements deterministic naming for Qdrant collections in qdrant-client.ts using a three-tier strategy: custom names from .roo/codebase-index.json, git repository URLs, and workspace paths.
- Normalizes git URLs for consistent hashing across formats (SSH, HTTPS).
- Adds error handling and fallbacks for non-git repositories.
Testing:
- Adds 11 new test cases in qdrant-client.spec.ts to cover new naming logic and edge cases.
- Tests include scenarios for custom names, git URL normalization, and workspace path fallbacks.
Misc:
- Updates QdrantVectorStore to handle new naming logic and improve error handling.

^{This description was created by}^{for 426e703. You can customize this summary. It will automatically update as commits are pushed.}

- Add support for custom collection names via .roo/codebase-index.json - Use git repository URL for deterministic collection naming across worktrees - Fall back to workspace path hash when no git repo is available - Normalize git URLs for consistent hashing - Add comprehensive tests for new naming strategies Fixes #7940

- Add better error handling in normalizeGitUrl method - Remove unused import of getGitRepositoryInfo - Improve URL credential removal logic

roomote

Reviewing my own code because apparently I trust no one, not even myself.

roomote · 2025-09-12T22:10:38Z

src/services/code-index/vector-store/qdrant-client.ts

+	private normalizeGitUrl(url: string): string {
+		try {
+			// Remove credentials from HTTPS URLs
+			let normalized = url


The credential removal logic here is good, but consider adding a comment explaining why this is important for security (preventing credential leakage in collection names). This helps future maintainers understand the security implications.

roomote · 2025-09-12T22:10:38Z

src/services/code-index/vector-store/qdrant-client.ts

+		try {
+			const configPath = path.join(workspacePath, ".roo", "codebase-index.json")
+			if (fs.existsSync(configPath)) {
+				const config = JSON.parse(fs.readFileSync(configPath, "utf8"))


Currently this silently falls back when JSON parsing fails. Should we consider logging malformed JSON at a higher level than console.warn to help users debug configuration issues? Or perhaps add validation for the expected schema?

roomote · 2025-09-12T22:10:39Z

src/services/code-index/vector-store/qdrant-client.ts

+				const configContent = fs.readFileSync(configPath, "utf8")
+
+				// Extract remote URL
+				const urlMatch = configContent.match(/url\s*=\s*(.+?)(?:\r?\n|$)/m)


This regex pattern might not handle all edge cases in git config files (e.g., URLs with spaces or special characters). Have you considered using a more robust git config parser library, or should we add more comprehensive error handling for edge cases?

roomote · 2025-09-12T22:10:39Z

src/services/code-index/vector-store/qdrant-client.ts

+	 * @param workspacePath Path to the workspace
+	 * @returns Collection name
+	 */
+	private generateCollectionName(workspacePath: string): string {


Consider adding JSDoc comments to these new private methods explaining the fallback hierarchy (custom name → git URL → workspace path). This would make the code more self-documenting for future contributors.

roomote · 2025-09-12T22:10:39Z

src/services/code-index/vector-store/__tests__/qdrant-client.spec.ts

 		expect((vectorStore as any).vectorSize).toBe(mockVectorSize)
 	})
+
+	it("should use custom collection name from .roo/codebase-index.json if available", () => {


Great test coverage! Consider adding a few more edge case tests:

Malformed JSON in .roo/codebase-index.json

Git repositories with multiple remotes

Collection name length edge cases (approaching the 255 character limit)

These would help ensure robustness across different user configurations.

roomote · 2025-09-12T22:10:39Z

src/services/code-index/vector-store/qdrant-client.ts

+	private loadCustomCollectionName(workspacePath: string): string | undefined {
+		try {
+			const configPath = path.join(workspacePath, ".roo", "codebase-index.json")
+			if (fs.existsSync(configPath)) {


Minor performance note: These synchronous file operations could potentially block the event loop. Since this only happens during initialization it's probably fine, but worth noting for future optimization if initialization performance becomes a concern.

daniel-lxs · 2025-09-17T19:54:42Z

A possible solution is to let the user create a .rooid file, either in the workspace root or inside the .roo folder. The file would just contain one line: the collection ID.

We’d need to confirm a few things first, like how to handle simultaneous writes to the index file, so it doesn’t break.

roomote added 2 commits September 12, 2025 22:00

refactor: improve git URL normalization and error handling

426e703

- Add better error handling in normalizeGitUrl method - Remove unused import of getGitRepositoryInfo - Improve URL credential removal logic

roomote bot requested review from cte, jr and mrubens as code owners September 12, 2025 22:06

github-project-automation bot added this to Roo Code Roadmap and Roo Code Roadmap Sep 12, 2025

github-project-automation bot moved this to Triage in Roo Code Roadmap Sep 12, 2025

github-project-automation bot moved this to New in Roo Code Roadmap Sep 12, 2025

dosubot bot added size:L This PR changes 100-499 lines, ignoring generated files. enhancement New feature or request labels Sep 12, 2025

roomote bot mentioned this pull request Sep 12, 2025

Deterministic naming for Codebase Indexing Qdrant collections #7940

Open

4 tasks

roomote bot commented Sep 12, 2025

View reviewed changes

hannesrudolph added the Issue/PR - Triage New issue. Needs quick review to confirm validity and assign labels. label Sep 12, 2025

daniel-lxs moved this from Triage to PR [Needs Prelim Review] in Roo Code Roadmap Sep 15, 2025

hannesrudolph added PR - Needs Preliminary Review and removed Issue/PR - Triage New issue. Needs quick review to confirm validity and assign labels. labels Sep 15, 2025

daniel-lxs closed this Sep 17, 2025

github-project-automation bot moved this from New to Done in Roo Code Roadmap Sep 17, 2025

github-project-automation bot moved this from PR [Needs Prelim Review] to Done in Roo Code Roadmap Sep 17, 2025

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

feat: add deterministic naming for Qdrant collections #7943

feat: add deterministic naming for Qdrant collections #7943

Uh oh!

roomote bot commented Sep 12, 2025 •

edited by ellipsis-dev bot

Loading

Uh oh!

roomote bot left a comment

Uh oh!

roomote bot Sep 12, 2025

Uh oh!

roomote bot Sep 12, 2025

Uh oh!

roomote bot Sep 12, 2025

Uh oh!

roomote bot Sep 12, 2025

Uh oh!

roomote bot Sep 12, 2025

Uh oh!

roomote bot Sep 12, 2025

Uh oh!

daniel-lxs commented Sep 17, 2025

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

feat: add deterministic naming for Qdrant collections #7943

feat: add deterministic naming for Qdrant collections #7943

Uh oh!

Conversation

roomote bot commented Sep 12, 2025 • edited by ellipsis-dev bot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Problem

Solution

Key Changes

Benefits

Testing

Uh oh!

roomote bot left a comment

Choose a reason for hiding this comment

Uh oh!

roomote bot Sep 12, 2025

Choose a reason for hiding this comment

Uh oh!

roomote bot Sep 12, 2025

Choose a reason for hiding this comment

Uh oh!

roomote bot Sep 12, 2025

Choose a reason for hiding this comment

Uh oh!

roomote bot Sep 12, 2025

Choose a reason for hiding this comment

Uh oh!

roomote bot Sep 12, 2025

Choose a reason for hiding this comment

Uh oh!

roomote bot Sep 12, 2025

Choose a reason for hiding this comment

Uh oh!

daniel-lxs commented Sep 17, 2025

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

roomote bot commented Sep 12, 2025 •

edited by ellipsis-dev bot

Loading