Skip to content

PartProductionTracking¤

Track production quantities by part number from part-ID and counter signals.

Module: ts_shape.events.production.part_tracking Guide: Production Monitoring

When to Use¤

Use for part-level production reporting. Tracks how many of each part number were produced per hour, day, or custom period. Essential for production planning reconciliation and for computing per-part OEE when a line produces multiple SKUs.

Quick Example¤

The tracker reads long-format timeseries data: one row per signal reading, identified by a uuid column. The part number lives in a string signal (value_string) and the production count in a counter signal (value_integer).

from ts_shape.events.production.part_tracking import PartProductionTracking
import pandas as pd
import numpy as np

t = pd.date_range("2024-01-01 06:00", periods=480, freq="1min")

# Part-number signal (which part is running) and a counter signal
parts = pd.DataFrame({
    "uuid": "part_id",
    "systime": t,
    "value_string": (["PN-100"] * 160 + ["PN-200"] * 160 + ["PN-100"] * 160),
})
counter = pd.DataFrame({
    "uuid": "part_counter",
    "systime": t,
    "value_integer": np.arange(1, 481),
})
df = pd.concat([parts, counter], ignore_index=True)

pt = PartProductionTracking(df, time_column="systime")
hourly = pt.production_by_part(part_id_uuid="part_id", counter_uuid="part_counter", window="1h")
daily = pt.daily_production_summary(part_id_uuid="part_id", counter_uuid="part_counter")
totals = pt.production_totals(part_id_uuid="part_id", counter_uuid="part_counter")
print(hourly.head(10))

# If the counter resets (shift/part change, controller restart), enable
# reset handling so production is not silently dropped:
hourly = pt.production_by_part(
    part_id_uuid="part_id", counter_uuid="part_counter",
    window="1h", handle_resets=True,
)
resets = pt.detect_resets(part_id_uuid="part_id", counter_uuid="part_counter")

Key Methods¤

Method Purpose Returns
production_by_part(part_id_uuid, counter_uuid, window='1h', handle_resets=False) Count parts produced per part number per time window DataFrame with window, part_id, and quantity (plus resets when handle_resets=True)
daily_production_summary(part_id_uuid, counter_uuid, handle_resets=False) Aggregate daily production totals by part number DataFrame with date, part_id, and daily quantity
production_totals(part_id_uuid, counter_uuid, handle_resets=False) Compute total production by part number over the entire date range DataFrame with part_id and total quantity
detect_resets(part_id_uuid, counter_uuid) Find when the counter was reset DataFrame with timestamp, part_id, count_before, count_after, drop

Counter Resets¤

Production counters are assumed to increase as parts are produced. Many counters reset back to zero (or a lower value) — at a shift change, a part change, or a controller restart. By default (handle_resets=False), a window where the counter drops reports quantity = max(0, last_count - first_count), which clamps to 0 and loses all production in that window.

Pass handle_resets=True to count production correctly across resets. The quantity is then the sum of per-reading increments, where a drop is treated as a reset that contributes the new counter value. The result gains a resets column counting resets per window.

# Reset-aware hourly production
hourly = pt.production_by_part(
    part_id_uuid="part_id", counter_uuid="part_counter",
    window="1h", handle_resets=True,
)

# Inspect exactly when the counter reset
resets = pt.detect_resets(part_id_uuid="part_id", counter_uuid="part_counter")
print(resets)   # timestamp, part_number, count_before, count_after, drop

Tips & Hints¤

Enable reset handling when counters reset

If your counter resets at part changes, shift boundaries, or controller restarts, pass handle_resets=True so production is not silently dropped. Use detect_resets() first to confirm whether and when your counter resets.

Related modules

See Also¤