After a successful Run, SigGer writes one file in the Destination
folder. The path is
Destination / SignalName . ext
where the extension depends on the target-format tumbler.
I/O is performed by the project module ScyIO_UnB.
| Tumbler | Extension | Typical contents | Spec / docs |
|---|---|---|---|
| HDF5 | .h5 |
Hierarchical datasets; waveform plus fs |
HDF5 documentation |
| Parquet | .parquet |
Columnar table (Apache Arrow / Parquet) | Parquet documentation |
| Matlab | .mat |
MATLAB Level-compatible MAT-file | MAT-file versions |
Every writer attaches fs so a reader does not have
to guess the rate from the file name. After the ScyIO write,
generate_signal._attach_sampling_frequency stores
fs as:
fs and sampling_frequency, dataset fs, and variables/fs;
When you later open the record in an analyser, treat
fs as the official rate in hertz used in (3)–(5).
| Signal type | After synthesis | On disk |
|---|---|---|
| Floating-point | float64 working buffer |
float32 |
| Fixed-point | rounded int32 |
int16, clipped to [−32768, 32767] |
Writes use overwrite=True. Saving again with the same
name and folder replaces the previous file.
Generators fill the array in ten contiguous slices
Ik = [ ⌊N(k−1)/10⌋ , ⌊N k / 10⌋ ) , k = 1,…,10 (37)
and call progress_callback(10k) after each slice.
The UI processes events so the circular gauge can repaint.
h5py, pandas.read_parquet, scipy.io.loadmat.load('name.mat') and read the fs variable.