feat(api, performance-metrics): track analysis execution

[DerekMaggio]

Overview

Closes https://opentrons.atlassian.net/browse/EXEC-399

Changelog

opentrons-shared-data Project

• Add shared types to performance/dev_types.py
• Define SupportsTracking mypy Protocol
• Move RobotContextState from performance-metrics to shared-data

performance-metrics Library

• Export RobotContextTracker to allow utilization of library
• Add constants file, which contains file names for different performance metrics. I figured this was just as easy as providing a temporary name and then changing it later
• RobotContextTracker
  • Conditionally import timing functions based on OS. This is to address CI actions that were failing on Windows
  • RobotContextTracker is a structural subtype of SupportsTracking

robot_server Project

• Add performance-metrics library as a dev dependency

api Project

• Add performance-metrics library as a dev dependency
• Add singleton initialization of RobotContextTracker inside the api project
• Decorate _analyze function with tracker
• Initialize the performance_metrics_data directory inside the data directory on the robot

Test Plan

• Add unit test cases for tracking analysis
• Verify on OT-2
• Verify on Flex

Review requests

• Should I move RobotContextStates to shared_data?
• Is the weirdness import weirdness I did in test_cli.py okay?

from opentrons import get_robot_context_tracker


# Enable tracking for the RobotContextTracker
# This must come before the import of the analyze CLI
context_tracker = get_robot_context_tracker()
context_tracker._should_track = True

from opentrons.cli.analyze import analyze  # noqa: E402

Risk assessment

Low, hidden behind feature flag which defaults to being False