architecture.md

Architecture v1

This is the architecture definition of the first version

The QPU is a specialized interface used to access parameters and calibrations that are tracked between quantum experiments and enable detection of staleness

Concept

The data is stored in a two level hierarchy which can be thought of as a filesystem.

The first level is folder or directory, and the second level is file

QPU is an easy interface to save files in directories, while keeping the history of each file. QPU will also handle the versioning using a git-based single-branched model.

Create Script

Bundled with this package, there should be a script (installed in bin) that will help the users create new databases. The python API itself focuses on manipulating such database

entropy-qpu init

This creates a clear separation between the initial setup work, and the day-to-day work

API

from entropylab_qpudb import QpuDatabase

# Create the object connecting to the database on the python side
db = QpuDatabase("./db")

# (Optional) Set the entire content of the database
initial_dictionary = {
    "folder1": {
        "file1": 6,
        "file2": 6
    },
    "folder2": {
        "file5": [1, 2, 3],
        "file8": np.array([1 + 0.4 * 1j, 2, 3])
    }
}
db.set_all(initial_dictionary)

Working with individual files and folders can be done using explicit methods

db.get("folder2", "file5").value = 60
print(db.get("folder2", "file5").value) # prints 60

db.get("folder2", "file5") returns a wrapping object QpuParameter which enable access to the content and the metadata (file modification datetime)

It is possible to access the items using properties

db.folder2.file5.value = 60

This should be equivalent