Developer Documentation

API & Mathematical Reference

Technical specifications, coordinate system formulas, and programmatic integration guides.

POSThttps://api.aigraphreader.com/v1/digitize

Programmatically submit a graph image or SVG URL to extract calibrated data vectors.

Example Request (cURL)

curl -X POST https://api.aigraphreader.com/v1/digitize \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -F "file=@figure_1_experiment.png" \
  -F "graph_type=xy_cartesian" \
  -F "x_range=[0, 100]" \
  -F "y_range=[0, 500]" \
  -F "scale_x=linear" \
  -F "scale_y=log10"

Response JSON Schema

{
  "status": "success",
  "data": {
    "series": [
      {
        "name": "Series 1",
        "points": [
          {"x": 10.2, "y": 142.5},
          {"x": 20.5, "y": 230.1},
          {"x": 30.1, "y": 385.0}
        ],
        "metrics": {
          "mean_x": 20.26,
          "max_y": 385.0,
          "confidence_score": 0.994
        }
      }
    ]
  }
}

Coordinate Mapping Mathematics

1. Linear Cartesian Transform

For calibrated pixel coordinates $(p_x, p_y)$ within bounding box $[x_0, y_0, x_1, y_1]$:

X = X_min + ((p_x - x_0) / (x_1 - x_0)) * (X_max - X_min)

2. Logarithmic Axis Mapping

For log10 scaled axes spanning multiple orders of magnitude:

Y = 10^( log10(Y_min) + ((p_y - y_0) / (y_1 - y_0)) * (log10(Y_max) - log10(Y_min)) )

3. Ternary Barycentric Coordinates

For three-component systems with vertices $V_A, V_B, V_C$ satisfying $A + B + C = 100\%$:

(A, B, C) = solve_barycentric(p_x, p_y, V_A, V_B, V_C)

Open Interactive Digitizer Studio