DOCSEARCHbyAlgolia
Submit Docs

coScene Documentation

website·Indexed 31 days ago

https://docs.coscene.cn

Documentation for coScene's visualization tools, including message schemas for 3D vectors, poses, and primitives, as well as configuration for panels such as Log, Plot, Image, and Service Call. It also covers the development of custom visualization extensions and message converters to support Foxglove schemas.

Tokens
153K
Snippets
641
Records
1.4K
Agent score
51%

What's inside coScene

  1. Overview of coScene Workflow System

    coScene provides a task-based workflow system integrated with its data engine. It is designed to automate data cleaning, preprocessing, training, testing, and deployment, allowing AI engineers to focus on model training rather than manual data preparation. The system allows users to define search and filter criteria within the data engine and use those standards as inputs for automated workflows to retrieve and process data.

  2. Overview of cocli record commands

    The cocli record command group provides a suite of tools for managing coScene records and their associated files. Use cocli record -h for a general overview or cocli record [command] --help for specific command details.
    # Available commands include:
# copy, create, delete, describe, download, file, list, moment, move, update, upload, view

cocli record -h

  3. Overview of coScene Multimodal Data Platform

    coScene is a cloud-native operational and development platform designed for managing complex multimodal device and behavior data. It provides a unified solution for handling large-scale, varied-format data through APIs, web applications, and desktop software, specifically targeting industries like autonomous driving and robotics.

  4. Overview of User Scripts

    User scripts are custom in-app scripts written in TypeScript that transform messages from one or more input topics into a new output topic. They support two types of data:

    • Playback data: Messages streaming frame-by-frame (e.g., for Raw Messages or 3D panels).
    • Range-loaded data: Messages for the entire data range (e.g., for Plot or State Transitions panels).

    When transforming range-loaded data, Foxglove runs two instances of the script: one for the full data range and one for the current playback frame. User scripts are local to a specific layout; for global transformations across all layouts, use a message converter.

  5. Overview of Automation Triggers

    Triggers in Coscene define the conditions required to execute automated actions. When a specified triggering condition is met, the system performs a configuration check and executes the associated actions.

  6. Overview of supported data formats for coScene visualization

    To utilize coScene's built-in visualization features, incoming messages should follow specific structures. The platform primarily supports Foxglove schemas. Supported formats include:

    • Protobuf
    • JSON Schema
    • ROS 1
    • ROS 2
    • TypeScript
    • FlatBuffers

    If custom messages are used, they can be converted to supported schemas using a message converter extension.

  8. Overview of coScene Organization Management

    An organization in coScene is a logical entity representing a team or company. It serves as a central hub to manage the following resources:

    • Projects: Units of data management.
    • Members: Users and their associated permissions.
    • Devices: Physical device connections managed centrally.
    • Images: A professional private image registry for Docker images used in automation and cloud testing.
    • Usage & Billing: Tracking of storage, active device counts, and computation time.

  9. Overview of creating records in Coscene

    Records are collections of files used to describe similar scenarios, allowing users to view or process data generated by devices over a specific period. To create a record, the user must have a role with permissions of 'Project Member' or higher.

  10. Overview of coScene Multimodal Data Platform

    coScene is a cloud-native operation and development platform designed for managing complex multimodal scene data from devices and behaviors. It provides a unified solution for handling large-scale data (e.g., high-volume data from intelligent robots) via web applications, APIs, CLIs, and edge software.

  11. Overview of Coscene Organization Management

    An Organization in Coscene is a logical entity representing a team or company. It serves as a container for multiple projects, members, devices, and images. The Organization Management interface allows administrators to manage the following components:

    • Projects: Units for data management.
    • Members: Users within the organization and their respective permissions.
    • Devices: Centralized management of hardware devices used for data collection and remote connection.
    • Images: A private image repository for automation and batch testing.
    • Usage & Billing: Tracking of storage volume, number of devices, and computation duration.

  12. Overview of Automatic Data Collection and Diagnosis

    Automatic Data Collection and Diagnosis is a remote operational service for device data. It uses predefined rules to monitor device logs in real-time, collect target data, and perform automatic diagnosis to analyze device failures and detect potential risks early.