Save Results to a File

Contents

Save Results to a File#

Data can be saved in different file formats. To do this, specify the file path in the constructor. The file extension determines the file format. Supported file formats include: - .nc or .cdf for NetCDF files - .zarr for Zarr files - .pkl for Pickle files Pickle files offer the advantage of storing generic Python objects but are not platform-independent. The following example demonstrates saving a dataset to a NetCDF file:

from sweepexp import sweepexp

sweep = sweepexp(
    func = lambda a, b: { "rat": a / b },
    parameters = { "a": [1, 4], "b": [1, 2] },
    save_path = "results.cdf",
)
sweep.run()

# Save the results to a file
sweep.save()

When executing this code a second time, a file named results.cdf already exists. In this case, the data is automatically loaded when the sweepexp object is initialized. Thus, calling sweep.run() during the second execution does not perform any experiments, as their status is already marked as “C” (completed).

However, calling sweep.save() again will raise an error because the file already exists. To overwrite the file, use the parameter mode=”w”:

sweep.save(mode="w")

Autosave#

To prevent data loss, set the auto_save parameter to True. When enabled, the data is automatically saved after each completed experiment. This ensures that, in the event of an unexpected program crash, completed experiments do not need to be repeated.

from sweepexp import sweepexp

sweep = sweepexp(
    func = lambda a, b: { "rat": a / b },
    parameters = { "a": [1, 4], "b": [1, 2] },
    save_path = "results.cdf",
    auto_save = True,
)
sweep.run()
Contents