THULab/qiskit_calibration_drift

Qiskit Calibration Drift (TsFile)

Apache TsFile version of phanerozoic/qiskit-calibration-drift.

Overview

Calibration parameters from IBM Quantum Heron processors joined to ambient and space-weather conditions at the time of each measurement, collected by a GitHub Action that polls backend.properties() every 30 minutes. The pinned revision covers 10,680,076 calibration events (2026-01-31 .. 2026-05-20) on one backend across 3 calibration properties (1- and 2-qubit), keyed on (backend, property, qubit_a, qubit_b); each event also records the calibration age, a chip-wide recalibration event id, and joined weather / geomagnetic / solar conditions. Designed for time-series forecasting of qubit drift and environmental coupling studies.

• Events: 10,680,076; properties: 3; qubit indices: 156 (qubit_a); value units: 2; 10 chip-wide recalibration events in range.
• observed_time is used as the series time axis at microsecond precision (the source timestamps carry us; ms would collide for 335 rows).

Schema (TsFile structure)

Data is sharded by the conversion tool into qiskit_calibration_drift_<n>.tsfile.

• Time (INT64, microseconds) — observed_time (UTC epoch us).
• backend, property, qubit_a, qubit_b, seq (TAG, STRING) — the measured device dimension; qubit_b is NA for single-qubit properties, and seq (0/1) keeps apart the 335 rows sharing the same key and exact microsecond.
• value, calibration_age_seconds (FIELD, DOUBLE)
• calibrated_time_us, snapshot_update_time_us (FIELD, DOUBLE) — the other source timestamps as epoch us (snapshot_update_time is missing for ~2.8% of rows and stays NaN).
• is_failure_ceiling, is_new_measurement (FIELD, BOOLEAN, nullable)
• property_family, unit, scope, chipwide_recal_event_id (FIELD, STRING)
• latitude, longitude, solar_zenith_deg, temperature_c, pressure_hpa, humidity_pct and the space-weather joins bz_gsm_nt, neutron_flux, kp_index, ap_index, Ap_daily, SN, f107_observed_sfu, f107_adjusted_sfu, solar_flux_sfu, dst_nt (FIELD, DOUBLE)

Rows are sorted by (TAG..., Time); all 31 source columns are represented (three source timestamps moved to Time / *_us fields).

Usage

Install the Apache TsFile Python SDK (pip install tsfile) and read a converted file:

from pathlib import Path
from tsfile import TsFileReader

path = Path("qiskit_calibration_drift_1.tsfile")
with TsFileReader(str(path)) as reader:
    schemas = reader.get_all_table_schemas()
    print("tables:", list(schemas))
    table_name = next(iter(schemas))
    table = schemas[table_name]
    columns = [column.get_column_name() for column in table.get_columns()]
    print("columns:", columns)
    field_names = [
        column.get_column_name()
        for column in table.get_columns()
        if column.get_column_name() not in {"Time", "time"}
    ]
    if field_names:
        with reader.query_table(table_name, field_names[:3], batch_size=1024) as result:
            batch = result.read_arrow_batch()
            if batch is not None:
                print(batch.to_pandas().head())

Source & license

• Original dataset: https://huggingface.co/datasets/phanerozoic/qiskit-calibration-drift
• License: CC-BY-4.0