mgi-pathology 0.0.1

Python SDK for compiling protocols for the MGI Pathology workstation.

mgi_pathology

Python SDK for compiling protocols for the MGI Pathology moving-liquid workstation.

A protocol script defines a metadata dict and a run(ctx) function; the SDK executes the function against a fresh Context, captures every command and form-layer step, and emits a validated ProtocolDefinition JSON that the workstation's runtime backend consumes.

Install

pip install mgi_pathology

Quick start

metadata = {
    "protocolName": "minimal demo",
    "author": "you",
    "description": "",
}


def run(ctx):
    ctx.init(ingredients={"0": ("Sample", "#FF6B6B")})
    pipette = ctx.load_pipette("p1000_flex8")
    gripper = ctx.load_gripper("gripper")
    tiprack = gripper.load_labware_to(
        "mdk_96_filtertiprack_1000ul",
        "1-1",
        stackId="s-1",
    )
    pipette.pick_up_tip(tiprack["A1"])
    pipette.drop_tip()

Compile to JSON:

mgi_pathology_py demo.py -o demo.json

Deck

The MGI Pathology deck has 7 lanes plus a fixed trash slot:

Lane	Positions
1	1-1, 1-2, 1-3, 1-4, 1-5
2	2-1, 2-2, 2-3, 2-4, 2-5
3	3-1, 3-2, 3-3, 3-4
4	4-1, 4-2, 4-3, 4-4
5	5-1, 5-2, 5-3, 5-4
6	6-1, 6-2, 6-3, 6-4
7	7-1, 7-2, 7-3, 7-4
trash	trash

The trash slot is fixed and does not need to be configured.

Hardware

• One pipette: p1000_flex8 (8-channel, 1000 µL).
• One gripper: gripper.
• Optional temperature module: pathologyTemperatureModuleV1.
• Three side stacks reachable by the gripper: s-1, s-2, s-3.

Documentation

See getting_started.md for a longer walkthrough.

---

Getting started

Protocol structure

Every protocol script is a regular Python file that defines two top-level objects:

• metadata — a dict with protocolName, author, description. Optional keys are passed through to the emitted JSON unchanged.
• run(ctx) — a function that receives a mgi_pathology.Context and uses it to describe the protocol.

The compiler executes run(ctx) against a fresh Context, captures every command emitted, and produces a ProtocolDefinition JSON.

Initialization

ctx.init(ingredients={"0": ("Sample", "#FF6B6B"), "1": ("Buffer", "#4ECDC4")})

Unlike the prepall workstation, the pathology deck shape is fixed (7 lanes, 29 slots, fixed trash bin), so init only takes optional ingredients — a registry of named liquids, keyed by id, with (displayName, displayColor) values.

init must be called once before any other Context method.

Loading hardware

pipette = ctx.load_pipette("p1000_flex8")
gripper = ctx.load_gripper("gripper")
temp    = ctx.load_module("pathologyTemperatureModuleV1", "5-1")
adapter = ctx.load_labware("mgi_pathology_plate_carrier_adapter", "3-2")

Slots are referenced by string identifiers like "3-2" (lane 3, position 2). See LocationName in mgi_pathology.protocol_api.deck for the full list.

Feeding labware off side stacks

The gripper can pull labware off three side stacks s-1, s-2, s-3 and drop it on the deck:

tiprack = gripper.load_labware_to(
    "mdk_96_filtertiprack_1000ul",
    "1-1",            # destination slot
    stackId="s-1",
    code="T1",        # optional operator-facing label
)

Each call materialises a fresh Labware and registers it on the context, so subsequent pipette operations can target it by reference. The destination can be a slot string ("1-1"), an existing Labware (stack on top), or a Module.

Moving labware between deck locations

gripper.move_labware(plate, "4-2")           # to a slot
gripper.move_labware(plate, temp_module)     # onto a module
gripper.move_labware(plate, base_adapter)    # stack on another labware

Liquid handling

Liquid handling lives on Pipette and follows the Opentrons-style API:

pipette.pick_up_tip(tiprack["A1"])
pipette.aspirate(100, source["A1"])
pipette.dispense(100, dest["A1"])
pipette.drop_tip()

For higher-level workflows, use pipette.transfer(...), pipette.distribute(...), pipette.consolidate(...), and pipette.mix(...).

Compiling

mgi_pathology_py path/to/protocol.py -o output.json

The CLI is also available as python -m mgi_pathology path/to/protocol.py.

Trash

pipette.drop_tip() (no argument) drops the tip into the fixed trash bin. The backing loadLabware for the trash slot is emitted lazily the first time drop_tip() is called.