docs: add instruction for supporting vector search in DEVELOPER.md#3137
docs: add instruction for supporting vector search in DEVELOPER.md#3137duwenxin99 wants to merge 1 commit into
Conversation
![gemini-code-assist[bot]](https://avatars.githubusercontent.com/in/956858?s=60&v=4){kind=small}
There was a problem hiding this comment.
Code Review
This pull request updates DEVELOPER.md to include new guidelines for implementing vector search, covering aspects like embedding format validation, formatter reuse, and documentation. The review highlights two issues: an incorrect markdown link syntax for formatters that uses a hardcoded, potentially stale URL, and an inaccurate link text for the BigQuery example, which should be updated for consistency with repository terminology.
| * **Implement Unit Tests** in a file named `newdbtool_test.go`. | ||
| * **Implement Vector Search** if your new tool supports it. You must: | ||
| 1. Validate that the vector embedding format can be injected successfully into your Tool's statement. If not, update `Tool.EmbedParams()` to pass in a vector formatter into `parameters.EmbedParams`. | ||
| 1. Feel free to reuse existing vector [formatters][https://github.com/googleapis/mcp-toolbox/blob/0917e5bf75ca4d29261bf630a7ce4cdab3d9949b/internal/embeddingmodels/embeddingmodels.go#L38] or create new ones. |
There was a problem hiding this comment.
The markdown link syntax is incorrect ([text][url]) and will not render as a clickable link. Additionally, using a hardcoded URL with a specific commit hash is discouraged as it will become stale. It is better to use a relative path to the file within the repository.
| 1. Feel free to reuse existing vector [formatters][https://github.com/googleapis/mcp-toolbox/blob/0917e5bf75ca4d29261bf630a7ce4cdab3d9949b/internal/embeddingmodels/embeddingmodels.go#L38] or create new ones. | |
| 1. Feel free to reuse existing vector [formatters](internal/embeddingmodels/embeddingmodels.go) or create new ones. |
| * **Implement Vector Search** if your new tool supports it. You must: | ||
| 1. Validate that the vector embedding format can be injected successfully into your Tool's statement. If not, update `Tool.EmbedParams()` to pass in a vector formatter into `parameters.EmbedParams`. | ||
| 1. Feel free to reuse existing vector [formatters][https://github.com/googleapis/mcp-toolbox/blob/0917e5bf75ca4d29261bf630a7ce4cdab3d9949b/internal/embeddingmodels/embeddingmodels.go#L38] or create new ones. | ||
| 1. Add tests and documentation for vector injection and vector search. See the [BigQuery Source](docs/en/integrations/bigquery/tools/bigquery-sql.md) for an example. |
There was a problem hiding this comment.
The link text "BigQuery Source" is inaccurate in this context because the link points to the documentation for the bigquery_sql tool, not the source itself. Using "BigQuery SQL tool" ensures consistency with the repository's terminology.
| 1. Add tests and documentation for vector injection and vector search. See the [BigQuery Source](docs/en/integrations/bigquery/tools/bigquery-sql.md) for an example. | |
| 1. Add tests and documentation for vector injection and vector search. See the [BigQuery SQL tool](docs/en/integrations/bigquery/tools/bigquery-sql.md) for an example. |
References
- Tool names must be in snake_case.
2 participants
No description provided.