Pytest on TofuPilot

pytest is a code-first Python test framework built around test_* functions and assert statements, with hardware setup handled via fixtures.

With TofuPilot, you can deploy pytest suites to your stations via Git push and stream live test data to both the kiosk and the dashboard, with zero configuration.

Getting started

To get started with pytest on TofuPilot:

• Clone our pytest starter template to your favorite Git provider and deploy it on TofuPilot.

Pytest Starter

Pure pytest suite with asserts promoted to measurements

tofupilot/template-pytest-starter

• If you have a pytest suite, push it to your Git provider and import it from the New Procedure wizard.

Integration

The TofuPilot CLI runs your pytest suite natively under an embedded plugin and streams tests, asserts, and logs to the dashboard.

def test_supply_voltage():
    voltage = read_voltage()
    assert 4.8 <= voltage <= 5.2, "Supply voltage [V]"

Tests

In pytest, every collected test_* function is one unit of execution. TofuPilot maps each test to a phase row.

Measurements

The plugin parses each test's AST at collection time and promotes recognized assert patterns to typed measurements with limits and a runtime-captured value.

def test_temperature():
    temp = read_temperature()
    assert 20 <= temp <= 80, "Case temperature [°C]"

The assert message becomes the measurement docstring. Append [unit] to set the units field on numeric measurements.

Logs

Standard Python logging records flow through TofuPilot. The connector attaches a handler to the root logger at session start, so any module-level call is captured.

import logging

def test_calibration():
    logging.info("Calibration started")
    assert calibrate() == "OK"

Logs are stored at the run level; per-test association is not preserved.

Run metadata

pytest has no built-in concept of a unit under test. TofuPilot reads defaults from pyproject.toml to identify the run:

[tool.tofupilot]
serial_number = "SN-0001"
part_number = "PCB01"
revision_number = "A"
batch_number = "2024-001"
auto_identify = false

When auto_identify is false (default), the CLI prompts the operator for the serial number on stdin before collection. When true, the defaults are used directly — the right choice for stations where a barcode scanner fills the serial in advance.

Fixtures

pytest fixtures run as-is in Python — setup, teardown, scoping, and dependency injection work unchanged. The dashboard does not surface fixture instances or lifecycle events; only the data emitted from inside a test (measurements, logs) is captured.

Offline upload

When the bench can lose connectivity, the CLI queues runs locally and uploads them when the network comes back. Each queued run keeps its original timestamp, so analytics stay accurate even when uploads land hours later. See tofupilot queue.

Unsupported features

The following pytest features run as-is in Python, but TofuPilot does not surface them on the dashboard.