[TESTING_ENVIRONMENT] Initial pre-recorded data algorithm positioning analyzer

[mieshki]

• Common JSON format with validation
• CLI usage for automatic testing
• --visual-debug option to see estimation errors interactively
• Handy new algorithms/grids registration
• Allows to test many datasets at once

Current algorithms and datasets are only for demo purposes, proper ones to be developed/collected later on

Initial thresholds:
    <0.15mm (good)
    0.15-0.35mm (acceptable)
    0.35-0.7mm (too big)
    >0.7mm (terrible)

More details in README.md

Example output:

> python analyzer.py --dataset datasets/hexgrid --print-metric-guide

Skipped (grid type incompatibility):
  • ideal_algorithm × test_centers.json: Algorithm supports SquareGrid, dataset uses HexagonalGrid
  • ideal_algorithm × test_long.json: Algorithm supports SquareGrid, dataset uses HexagonalGrid


Metric guide:
N       → Number of samples.
MAE     → Average position error. Lower is better.
Median  → Typical error (robust to outliers).
RMS     → Penalizes large errors strongly.
P95     → 95% of errors are below this value.
P99     → Worst realistic case.
Max     → Absolute worst case.
Fail%   → % of points above failure threshold.
BiasX/Y → Systematic shift on axis (should be ~0).
Score   → Overall quality score (0-100, higher is better).

Good signs:
• Median ≈ MAE
• RMS slightly above MAE
• P95 not >> MAE
• Bias near 0
• Low Fail%
• High Score (>70)

+--------------+------------------------------------+-----+------+------+--------------+------+------+-------+-------+-------+-------+--------+-------+
|  Algorithm   |                Data                |  N  | MAE  | RMS  | Median (P50) | P90  | P95  |  P99  |  Max  | BiasX | BiasY | Fail%  | Score |
+--------------+------------------------------------+-----+------+------+--------------+------+------+-------+-------+-------+-------+--------+-------+
| gaussian_fit | datasets/hexgrid/test_centers.json |  13 | 0.10 | 0.11 |     0.09     | 0.16 | 0.19 |  0.22 |  0.23 |  0.03 | -0.00 | 23.1%  |  85.5 |
| gaussian_fit |  datasets/hexgrid/test_long.json   | 153 | 2.46 | 3.80 |     2.17     | 3.52 | 4.01 | 16.02 | 26.11 |  0.57 |  0.06 | 100.0% |  0.0  |
+--------------+------------------------------------+-----+------+------+--------------+------+------+-------+-------+-------+-------+--------+-------+


> python analyzer.py --help
usage: analyzer.py [-h] [--list-algorithms] [--list-grids] [--list-algorithms-per-grid GRID_TYPE] [--dataset DATASET] [--algorithm ALGORITHM] [--detailed] [--print-metric-guide] [--visual-debug]

Position Algorithm Analyzer - Evaluate magnetic position estimation algorithms

options:
  -h, --help            show this help message and exit
  --list-algorithms     Print available algorithms and exit
  --list-grids          Print supported grid types and exit
  --list-algorithms-per-grid GRID_TYPE
                        List algorithms that support a specific grid type
  --dataset DATASET     Path to a single JSON file or folder containing JSON files
  --algorithm ALGORITHM
                        Algorithm(s) to run (use 'all' to run all algorithms, or comma-separated list like 'alg1,alg2', default: all)
  --detailed            Enable detailed analysis output
  --print-metric-guide  Print metric guide before results (default: False)
  --visual-debug        Enable visual debugging

[socket-security]

Review the following changes in direct dependencies. Learn more about Socket for GitHub.

Diff	Package	Supply Chain Security	Vulnerability	Quality	Maintenance	License
	pypi/​pygame@​2.6.1
	pypi/​numpy@​2.4.2
	pypi/​prettytable@​3.17.0

View full report

[Ross0907]

Rudimentary simulation script : here. Feel free to provide suggestions.