TLS-Data-Paper

Data and dataset creation scripts accompanying the manuscript:

Spectroscopy and Coherent Control of Two-Level System Defect Ensembles Using a Broadband 3D Waveguide

---

Repository structure

src/
  dataset_creation_scripts/
    pulse_width_sweep/
      experiment_2_dataset_creation.py   ← builds pulse width sweep dataset from raw .npy files
      README.md
      column_info.json
      metadata.json
      experiment_2/
        README.md                        ← dataset documentation
    amplitude_control/
      emperiment_3_dataset_creation.py   ← builds amplitude control dataset from raw .npy files
      README.md
      column_info.json
      metadata.json
      experiment_3/
        README.md                        ← dataset documentation
        figure3_dataset_long.npz         ← full dataset (1,800,000 rows × 8 columns)
        figure3_long_preview.csv         ← first + last 5 traces for quick inspection

---

Experiments

Experiment	Description	Dataset rows	Key swept variables
Pulse Width Sweep	Single-pulse width sweep, silicon SiOx sample	21,000,000	pulse width (42 values), frequency (500 values)
Amplitude Control	Two-pulse coherent control amplitude sweep, 3 samples	1,800,000	amplitude (100 values), pulse spacing (6 values)

---

Generating datasets from raw data

Each script accepts the path to the raw .npy files and an output directory:

# Pulse Width Sweep
python src/dataset_creation_scripts/pulse_width_sweep/experiment_2_dataset_creation.py \
    --data_dir  /path/to/pulse_width_sweep/raw/npy \
    --output_dir /path/to/output

# Amplitude Control
python src/dataset_creation_scripts/amplitude_control/emperiment_3_dataset_creation.py \
    --data_dir  /path/to/amplitude_control/raw/npy \
    --output_dir /path/to/output

Both scripts run sanity checks automatically after building and print a pass/fail report.

---

Dependencies

numpy
pandas
tqdm