Data Engineer mode with interactions with the active Notebook

[complyue]

Context

Adding a [Data Engineer] mode here, for Roo to work on the active Notebook together with the user.

Implementation

Tool functions are added to read/edit/exec cells in the active notebook.

Update 2:

Myself roughly satisfied with what's done so far, rebased & squashed for ease of review. Now it automatically execute new/modified cells upon notebook edit.

A flaw I encounter is that even though I wrote the prompts trying to allow multiple cells inserted/replaced in a single tool call, currently the framework wouldn't allow multiple <cell_content/> args to be passed around. It works well so long as the agent does insert/modify/replace a single cell at a time.

What's your suggestions?

Update:

Sonnet MAX saved the day! Now the UI works as expected 😄

Yet I fail to make the tool usage approving UI show up correctly, tho blindly approval works, neither auto-approval works. Also added 2 new settings to truncate huge nb cell outputs, as well as config nb exec timeout, failed to make them show up in settings UI either.

Further enhancement may include allowing modify + exec the notebook in a single round, I'm not sure the best approach.

I'm open to ideas around this feature.

Please review!

Screenshots

How to Test

Switch to [Data Engineer] mode, ask Roo to read/edit/exec the active notebook with any goal.

Get in Touch

I prefer GH messages.

---

Important

Introduces Data Engineer mode for notebook interaction, with settings for output size and execution timeout, and updates localization and state management.

• Behavior:
  • Adds Data Engineer mode to interact with active notebooks, including reading, editing, and executing cells.
  • Supports auto-approval for notebook actions in ChatView.tsx.
• Settings:
  • Adds NotebookSettings.tsx for configuring output size limit and execution timeout.
  • Updates SettingsView.tsx to include notebook settings.
• Context and State:
  • Updates ExtensionStateContext.tsx to manage new notebook-related states.
  • Adds tests in ExtensionStateContext.test.tsx for new state management.
• Localization:
  • Updates localization files for new notebook-related strings in English, Chinese (Simplified and Traditional).

^{This description was created by for db8f194c3a634d2f89e2d719dab6e6c043624b75. It will automatically update as commits are pushed.}

[changeset-bot]

⚠️ No Changeset found

Latest commit: 5409742

Merging this PR will not cause a version bump for any packages. If these changes should not result in a new version, you're good to go. If these changes should result in a version bump, you need to add a changeset.

This PR includes no changesets

When changesets are added to this PR, you'll see the packages that this PR includes changesets for and the associated semver types

Click here to learn what changesets are, and how to add one.

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

[ellipsis-dev]

This pull request introduces a significant number of changes across multiple files, primarily related to notebook tool support, descriptions, and settings, as well as a new NotebookService class and various configurations for notebook operations. Given the size and scope of the changes, it might be beneficial to split the pull request into smaller, logical sections to improve reviewability.

Consider splitting the changes into:

1. Notebook Tool Support: Changes related to adding support for notebook read, edit, and execute actions.
2. Notebook Settings and Descriptions: Changes related to adding descriptions and settings for notebook tools.
3. NotebookService Implementation: Changes related to the implementation of the NotebookService class and related operations.

This approach could help reviewers focus on specific aspects of the feature and ensure a more thorough review process.

[ellipsis-dev]

The error message on line 332 mentions 'cell_index' rather than 'cellIndex'. For clarity and consistency, consider updating the message.

Suggested change

	throw new Error("cell_index is required for executing a cell.")
	throw new Error("cellIndex is required for executing a cell.")

[samhvw8]

cool

[complyue]

thanks to Sonnet MAX! now rebased on 3.9, and UI solved 😃, with i18n (en+zh)

set to optimize some wrt workflow

[complyue]

Myself roughly satisfied with what's done so far, rebased & squashed for ease of review. Now it automatically execute new/modified cells upon notebook edit.

A flaw I encounter is that even though I wrote the prompts trying to allow multiple cells inserted/replaced in a single tool call, currently the framework wouldn't allow multiple <cell_content/> args to be passed around. It works well so long as the agent does insert/modify/replace a single cell at a time.

What's your suggestions?

[ellipsis-dev]

Typo on line 5: 'unless the <noexec/> parameter presents.' should be corrected to 'unless the <noexec/> parameter is present.'

Suggested change

	Description: Edit the active notebook in the editor. This tool allows you to insert new cells or replace existing cells. Note that new/modified code cells will be executed immediately, by default, unless the <noexec/> parameter presents.
	Description: Edit the active notebook in the editor. This tool allows you to insert new cells or replace existing cells. Note that new/modified code cells will be executed immediately, by default, unless the <noexec/> parameter is present.

[ellipsis-dev]

Naming Convention Issue: In the ClineSayTool interface, properties such as cell_content, cell_index, start_index, end_index, cell_type, and language_id are written in snake_case, whereas most other properties in the file use camelCase. Consider renaming them to cellContent, cellIndex, startIndex, endIndex, cellType, and languageId for consistency.

[ellipsis-dev]

The pull request is quite large, with 29 files changed and over 1800 lines added. It includes various changes such as adding notebook functionalities, tool interfaces, descriptions, configuration options, and localization strings.

To improve the review process, it might be beneficial to split this pull request into smaller, more focused ones. Here are some suggestions on how it could be split:

1. Core Functionalities: Changes related to the core functionalities of notebook read, edit, and execute in src/core.
2. Webview Components: Updates to webview components and settings in webview-ui/src/components.
3. Localization: All localization-related changes in webview-ui/src/i18n.

This will help reviewers focus on specific areas and ensure a more thorough review process. Please consider splitting the pull request if these changes are not tightly coupled.

[samhvw8]

@complyue cool, is it ready to review

[complyue]

@samhvw8 yes, please review!

[complyue]

Roo doing MATLAB!

[daniel-lxs]

Hey @complyue.
Thanks for your contribution. We genuinely apologize for letting your PR go stale and not taking the proper time to review it.

This seems like a big change and we would like to have an issue opened first with an use-case for these changes to properly discuss it. We are closing this PR but feel free to create an issue to revisit this.

We've now shifted to a clearer issue-first workflow to avoid this kind of situation going forward. Please create an issue first for any future contributions, as outlined there.

We truly appreciate your patience and would be happy to have you continue contributing.