CycleTimeEvents

Cycle times per part number from a part-number string signal and a cycle-complete trigger (boolean or integer). A cycle time is the time between consecutive completions.

ts_shape.events.production.cycle_time_tracking · Production guide · API reference

Example

import pandas as pd
import ts_shape
from ts_shape.events.production.cycle_time_tracking import CycleTimeEvents

df = pd.concat([
    ts_shape.make_id_signal("line1:part_number", ["PN-100", "PN-200"], hold=100),
    ts_shape.make_timeseries(["line1:cycle_done"], n_points=200, value_column="value_bool"),
], ignore_index=True)

ct = CycleTimeEvents(df)
ids = dict(part_id_uuid="line1:part_number", cycle_trigger_uuid="line1:cycle_done")
result = ct.cycle_time_statistics(**ids, max_cycle_gap="5min")
print(result[["part_number", "count", "min_seconds", "avg_seconds", "max_seconds"]])
slow = ct.detect_slow_cycles(**ids, threshold_factor=1.5)
trend = ct.cycle_time_trend(**ids, part_number="PN-100", window_size=20)

Methods

Method	Returns
cycle_time_by_part(...)	one row per cycle: systime (completion), part_number, cycle_time_seconds
cycle_time_statistics(...)	per part: count, min_seconds, avg_seconds, max_seconds, std_seconds, median_seconds
detect_slow_cycles(..., threshold_factor=1.5)	cycles slower than threshold_factor × the part's median: adds median_seconds, deviation_factor, is_slow
cycle_time_trend(..., part_number, window_size=20)	one part: cycle_time_seconds, moving_avg, trend (improving/stable/degrading)
hourly_cycle_time_summary(...)	per hour and part: cycles_completed, avg_/min_/max_cycle_time

All methods take part_id_uuid and cycle_trigger_uuid.

Notes

• Set max_cycle_gap (e.g. "5min"): longer intervals between completions are dropped as downtime. Without it, a break or jam between two completions becomes one very long cycle that inflates max/avg and is flagged as slow.
• The slow-cycle threshold is relative to each part's median, so mixed part numbers need no per-part limits.