Creating dev branch to base calc channels, rules, runs, and channels off

[ian-sift]

Data:

Getting data for multiple channels:
0: voltage: 100 points. Avg: 5.99
1: mainmotor.velocity: 29 points. Avg: 5.586206896551724
Start time: 2025-07-02 21:17:28.878926+00:00, End time: 2025-07-02 21:18:24.878785+00:00
2: gpio.12v: 89 points. Avg: 1.0
3: gpio.charge: 178 points. Avg: 0.8932584269662921
4: gpio.heater: 267 points. Avg: 0.6666666666666666
5: gpio.led: 356 points. Avg: 1.3089887640449438

=== Testing cache with varying time ranges ===

Test 1: Exact same time range no run_id (should hit cache)

Test 2: Subset of time range (should hit cache if overlapping)

Test 3: Extended time range earlier (should hit cache for overlapping portion)

Test 4: Different time encompassed range (should hit cache)

Test 5: No time range specified (should miss cache)
0: gpio.12v: 100 points. Avg: 1.0
1: gpio.charge: 100 points. Avg: 0.88
2: gpio.heater: 100 points. Avg: 0.19
3: gpio.led: 100 points. Avg: 3.07
4: voltage: 100 points. Avg: 5.02
5: mainmotor.velocity: 100 points. Avg: 5.24

Test 6: No time range specified again (should hit cache)
0: gpio.12v: 189 points. Avg: 1.0
1: gpio.charge: 189 points. Avg: 0.8359788359788359
2: gpio.heater: 189 points. Avg: 0.20105820105820105
3: gpio.led: 189 points. Avg: 3.1481481481481484
4: voltage: 200 points. Avg: 5.505
5: mainmotor.velocity: 129 points. Avg: 5.317829457364341

=== Cache Performance Summary ===
Original call: 0.3581 seconds
Exact time range no run_id: 0.0039 seconds (92.2x faster)
Subset time range: 0.0033 seconds (107.2x faster)
Extended time range earlier: 0.0892 seconds (4.0x faster)
Different time range: 0.0049 seconds (73.5x faster)
No time range: 0.8140 seconds (2.3x slower)
No time range repeat: 0.0023 seconds (360.6x faster) 

Ingestion https://app.development.siftstack.com/share/6E47Xye
Currently a little slow and not setup for bitfields or enums yet
Ingestion time: 0:01:07.627055 seconds
Ingested 2550 datapoints
Ingestion rate: 37.71 datapoints/second

[niliayu]

Yes, we should be able to leave this unspecified -- the org ID can be resolved via the API key on the backend.

[niliayu]

For your consideration, there is a bytes type available now that we might want to start looking at including in places like this.

[niliayu]

For consideration: maybe urllib.parse?