Documentation

How to use the framework

Everything an analyst needs to prepare a dataset, run a detection and read the output — plus the API surface the frontend expects.

Supported file format

  • • Extension: .csv
  • • Encoding: UTF-8, comma delimited
  • • Header row required
  • • Maximum size: 25 MB
  • • Column order must match the model feature schema

Upload instructions

  1. 1. Open a detection module from the navigation.
  2. 2. Drag your CSV onto the drop zone, or use Browse file.
  3. 3. Confirm the filename and size shown below the drop zone.
  4. 4. Press Analyze and wait for the service response.

Processing workflow

  1. 1. Client-side validation of type, size and emptiness.
  2. 2. Multipart upload to the FastAPI detection endpoint.
  3. 3. Feature extraction and model inference server-side.
  4. 4. JSON response rendered into metrics, charts and tables.
Result interpretation

Reading the output

Prediction

The winning class for the batch — an attack category or a malware family.

Confidence score

Probability assigned to the winning class, expressed as a percentage.

Risk / severity level

Derived triage priority: Low, Medium, High or Critical.

Recommendation

Suggested containment or mitigation action returned with the prediction.

Distribution charts

Share and absolute count of each predicted class across the batch.

Record table

Per-record predictions so individual flows or samples can be inspected.

API guide

Endpoints the frontend expects

Connected to API Backend

These routes describe the live contract that the dashboard uses to interact with the FastAPI machine learning service.
MethodEndpointPurposeStatus
POST/predict/networkBatch intrusion detection from a flow CSVLive
POST/predict/malwareMalware family classification from a feature CSVLive
GET/docsAPI interactive documentationLive
GET/Service health probeLive
FAQ

Frequently asked questions