Dedicated VSCode Extension Tutorial

약 779 단어약 3 분

The extension provides a series of convenient development capabilities for MaaAssistantArknights/MaaFramework, including but not limited to the following features:

tasks.json support, including template preview, next jump, task reference, etc.
Screenshot / Crop image

For details, please visit Extension Store or Repository.

Installation

Searching Maa and installing it from VSCode extensions list is recommended.

팁

When used for the first time, the extension will automatically download the preset version of resources. Search command Maa: select fetch registry to switch the download source (npm / cnpm).

Features

Control Panel

A dedicated control panel is added to the left, with icon

The major features of the extension are based on the interface.json configuration. Select the activated interface.json on the top of the control panel.

The extension has a Maa compatiable mode. If src/MaaCore folder exists inside the opened workspace, the mode will be enabled automatically.

Semantic resource analysis

Select expected resource via the Resource select inside the control panel. The extension will index and diagnose resource via corresponding paths.

If you find that the editing json isn't hinted by the extension, please check if the activated resource contains it.

The term definitions of task / task defs refers to key props of task objects.
The term references of task / task refs refers to values containing task name (e.g. in next) in other task objects.

Query task defs / task refs

The extension supports Jump to task defs, Jump to task refs and View task defs.

When enabling the Maa compatible mode, the extension will be able to parse template task, supporting querying task defs and task refs in conjunction with base tasks. Images that have the same name as the task will be shown when hovering the task defs.

Use Ctrl+T to quickly query and jump to a task definition.

Query / open images

The extension supports open images.

When enabling the Maa compatible mode, the extension will be able to recursively search for the image.

Task completion

The extension supports autocompletion according to all known tasks.

When enabling the Maa compatible mode, typing @ will trigger completion.

Image path completion

The extension supports autocompletion according to the path of all known images.

When enabling the Maa compatible mode, the extension will be able to recursively search for the image.

Check tasks / image paths

The extension supports scheduled scanning and diagnosing all tasks.

Check if contains task defs with same names.
Check if contains unknown task refs.
Check if contains unknown image refs.
Check if contains duplicated task refs in a single task.

Multiple paths resource support

The extension supports resource with multiple paths. The extension will perform logical overlapping according to specified order, thus content loaded later can reference content loaded earlier.

Evaluate Task / Tsak List Expression (Only Maa)

Evaluating the expanded task object and the sources of each properties, and the result of the task list expression via the control panel.

MaaPiCli feature (Only MaaFramework projects)

Scanning and selecting controllers, selecting resource, adding and manipulating tasks, and launching tasks can be done via the control panel.

Take screenshots and crop images

Searching and launching Maa: open crop tool inside VSCode command panel can open Screenshots / Crop panel.

Use Ctrl+Shift+P (Command+Shift+P on MacOS) to open command panel

After selecting and connecting to the controller, use Screencap button to obtain screenshots
Use Upload button to manually upload images.
Hold Ctrl key and select cropping area
Use wheels to zoom
After finishing cropping, use Download button to save the cropping result to the folder of the topest layer of the activated resource
Use Copy button to copy the ROI as an array to the clipboard.
Click Tool button to open the recognition tool panel, where you can directly test recognition on the current image.

경고

If OCR recognition returns an empty result, please check whether the OCR model is configured correctly.