APIJSON-Demo Documentation

CVAuto is a zero-code, zero-annotation automated testing platform for Computer Vision (CV) AI models. It allows users to test the functionality, effect, and performance of AI image recognition algorithms without manual bounding box or label annotation.

Key Workflow:

1. Upload Test Images: Upload images to be tested.
2. API Call: The platform calls the algorithm model via HTTP API to get a JSON response.
3. Automatic Rendering: It automatically draws labels, bounding boxes, lines, points, or polygons on the image based on the JSON response.
4. Automated Assertion & Correction: Users can verify if boxes are correct. If a false positive occurs, users can click near the border to change a checkmark (✓) to an 'X'.
5. Metric Calculation: The platform automatically calculates metrics such as Correct count, False Positives, False Negatives, Recall, Precision, and F1 Score, and allows for comparison between different versions.

AutoUI is a zero-code, zero-annotation Computer Vision (CV) AI automated testing tool. It is designed for algorithm engineers, researchers, and testers to quickly validate the functionality, effectiveness, and performance of AI models (such as YOLO detection, pose estimation, and segmentation) without manual ground truth labeling.

Key Capabilities:

• Automated Inference Visualization: Upload images or videos, call HTTP APIs to get JSON responses, and AutoUI automatically draws labels, bounding boxes, lines, points, and polygons on the media.
• Automated Assertion & Correction: Automatically checks if bounding boxes are correct. Users can manually correct errors (marking a box as ✓ or X) by clicking near the border, which instantly updates performance metrics.
• Metric Statistics: Automatically calculates accuracy, false positives, false negatives, recall, precision, and F1 Score.
• Data Export: One-click download of test reports (including summary metrics and per-image details) and export of training/testing datasets to accelerate model iteration.

The APIJSON-Java-Server directory contains several demo projects tailored to different database and connection pool requirements:

JDBC Demos

• APIJSONDemo: Simple testing with SpringBoot + MySQL/PostgreSQL. Configure via DemoSQLConfig.
• APIJSONBoot: Complex production-ready demo with SpringBoot + MySQL + PostgreSQL + SQLServer + TDengine. Configure via DemoSQLConfig.
• APIJSONFinal: Complex production-ready demo with SpringBoot + MySQL + PostgreSQL. Configure via DemoSQLConfig.

Connection Pool Demos

• APIJSONDemo-Druid: Testing with SpringBoot + Druid + MySQL/PostgreSQL. Configure via application.yml.
• APIJSONDemo-HikariCP: Testing with SpringBoot + HikariCP + MySQL/PostgreSQL. Configure via application.yml.
• APIJSONBoot-MultiDataSource: Complex production-ready demo with SpringBoot + Druid + HikariCP + PostgreSQL + SQLServer + TDengine + Dameng. Configure via application.yml.

Sharding & BigData Demos

• APIJSONDemo-ShardingSphere: Testing with SpringBoot + ShardingSphere + MySQL/PostgreSQL. Configure via application.yml and application-sharding-databases.properties.
• APIJSONBoot-BigData: Complex production-ready demo with SpringBoot + Presto + Trino + Elasticsearch.

APIJSON supports multiple datasources by allowing you to specify a @datasource key at different levels of the JSON request. The resolution follows this precedence order (from highest to lowest):

1. Object level: Specified directly within a specific object (e.g., "User:aa": { "@datasource": "db2" }).
2. Global level: Specified at the root of the JSON request (e.g., { "@datasource": "db2", ... }).
3. Application default: The default datasource configured in the application settings.

This allows you to mix data from different sources (like a relational database and Elasticsearch) in a single request, using cross-datasource references via the @ syntax (e.g., "user_id@": "User:aa/id").

{
  // Global datasource
  "@datasource": "db2",
  "@post": {
    "User:aa": {
       "@datasource": "db2" // Object level datasource
    },
    "ES_blog:a": {
       "@datasource": "elasticSearch" // Overriding to use Elasticsearch
    }
  }
}