CobotAR-Protocol
A collection of protobuf messages for the CobotAR-project
Current version: 1.0.8
Documentation
Find it here: documentation/README.md
New version
- Make changes
make publish bump={patch,minor,major}
NATS subject convention
CRUD
- add:
edit.<domain>.add - update:
edit.<domain>.update - delete:
edit.<domain>.delete - get:
get.<domain>.one - getByIds:
get.<domain>.some - getAll:
get.<domain>.all
Events
Subject shape: <kind>.<domain>.<type>.<scope..>
where:
- kind:
{command | event | plan | telemetry} - domain:
{robot | task | asset | ...} - type: event type as lower_snake_case string token (
start_stop,started,task_completed, etc.) - scope optional identifiers for routing (
robotId,taskId,assetId,lineId, etc.). Generally following this:<kind>.<domain>.<type>.<site>.<workcell>.<entityId>[.<planId>]
Examples:
command.robot.start_stop.workcellA.robot-1
plan.robot.path.workcellA.robot-1.plan-P123
telemetry.robot.state.workcellA.robot-1
event.robot.started.workcellA.robot-1
event.plan.cancelled.workcellA.plan-P123
Conceptual design of Assembly processes
A. Product definition (What the thing is)
- Parts
- Assemblies / sub-assemblies
- Product BOM / hierarchy
- Part occurrences / instances
- CAD / 3D model references
- Mating / connection hints if you have them
B. Process template (How the thing is normally assembled)
- Process recipe
- Sequences / sub-sequences
- Tasks / operations / steps
- Preconditions / ordering / alternatives / variants
- Required tool roles, skill requirements, actor constraints
- Operator text
C. Execution instance (What is happening right now on the floor)
- Running process
- Current state of sequences/tasks
- Assigned human(s)/robot(s)
- Start/end timestamps
- Current station
- Evidence / completion / faults / overrides
D. Resource library (What can be used to perform the work)
- Humans
- Robots
- Tools
- Fixtures
- Assets
- Stations / cells
- Models
E. Capability layer (Why a given resource can or cannot do a task)
- Skills
- Capability profiles
- Tool roles
- Constraints
- Validity / training / retraining / calibration / certification
Distinguish type from instance
Example: Part. In assembly we usually need both:
- Part definition: “M4x12 screw”
- Part occurrence: “the third M4x12 screw in this assembly”
thus distinguishing between the two is important and must be supported and done consistently.
Minimal TaskDefinition example
TaskDefinition {
id: "task-fasten-cover-screw-1"
name: "Fasten cover screw 1"
instruction_text: "Fasten screw 1 to 0.8 Nm"
task_type: TASK_TYPE_FASTEN
target: {
target_node_id: "node-screw-1"
target_part_definition_id: "part-m4x12"
local_target: {
anchor_id: "node-screw-1"
}
}
tool_requirements: [{
role: TOOL_ROLE_APPLY_TORQUE,
required_properties: [TOOL_PROPERTY_TORQUE_CONTROLLED],
minimum_capability: {
min_torque_nm: 0.72,
max_torque_nm: 0.88
}
}]
skill_requirements: [{
skill_id: "SK-APPLY-CONTROLLED-TORQUE",
minimum_level: SKILL_LEVEL_QUALIFIED
}]
validation: {
require_tool_feedback: true
allow_manual_confirmation: true
manual_confirmation_min_level: SKILL_LEVEL_QUALIFIED
}
execution_policy: {
assignment_preference: TASK_ASSIGNMENT_PREFERENCE_EITHER
actor_constraint: {
allowed_actor_kinds: [ACTOR_KIND_HUMAN, ACTOR_KIND_ROBOT]
collaboration_mode: COLLABORATION_MODE_SEQUENTIAL_HRC
}
}
}
Variant handling
Types of variants
Bucket A — Structural variants
Affect product tree. Modeled in product.
Examples:
- hinge side
- optional handle
- different fastener family
- extra sensor module
Bucket B — Workflow variants
Affect tasks/sequences/skills/tools/validation. Modeled in recipe applicability or recipe branching.
Examples:
- door must be flipped
- adhesive cure step required
- fragile part requires two-person lift
- robot can only assemble one orientation
Bucket C — Parametric variants
Do not change workflow structure. Passed at run-time.
Examples:
- color
- language
- customer text
- batch number
- cosmetic finish code
Handle variants without recipe explosion
For a product family, do this:
-
One ProductDefinition
- Contains all structural possibilities.
-
Few ProcessRecipes
Only split recipes where the workflow truly differs. For example:
- recipe-door-standard
- recipe-door-flipped-fixture
- recipe-door-adhesive
Not:
- red-left-short-handle
- red-left-long-handle
- blue-right-short-handle
- …
-
One ProcessRun per actual order/build
Contains:
- selected product configuration
- runtime parameters
Example
Product dimensions
- hinge_side = left | right
- color = red | green | blue
- handle = short | long
Recipe
recipe-door-standardis applicable when:- handle = short or long
- color = any
- hinge_side = left
recipe-door-flippedis applicable when:- hinge_side = right
They need to be split into two recipes because right-hinge requires a flip operation and different fixture.
Process run
When a real order starts the selection might look like this:
product_selection:
hinge_side = right
color = green
handle = long
thus the system chooses the recipe-door-flipped recipe and the recipe stays the same regardless of color.
Metadata
Release files for cobotar-protocol 1.0.8
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | |
|---|---|---|---|
| cobotar_protocol-1.0.8.tar.gz | 1.1 MB | Details |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Reset |
|---|---|---|---|---|
| cobotar_protocol-1.0.8-py3-none-any.whl | Python 3 | none | any | Details |
Total release size: 1.4 MB
Release files / cobotar_protocol-1.0.8.tar.gz
| Download URL | cobotar_protocol-1.0.8.tar.gz |
|---|---|
| Size | 1.1 MB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
812953fcebf10e6c182b86d62f4f62001287d14cec26260a6ca1de7cc78d75f8
|
|
BLAKE2b-256 checksum How to use checksums |
9a89671f0973ea043479b2692de064e7988b79b11149c58048101c27c7c98cd7
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/7.0.0 CPython/3.13.14
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Sep 13, 2026.
Transparency logRelease files / cobotar_protocol-1.0.8-py3-none-any.whl
| Download URL | cobotar_protocol-1.0.8-py3-none-any.whl |
|---|---|
| Size | 276.0 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
279c5c4af0ab40742f73977f6ff8256c79b7c4d8c17c1d7d3f89aa044a00ab1a
|
|
BLAKE2b-256 checksum How to use checksums |
9f3427e7f63f954da438a59dcb0c859afde004ffb57f561b3382f2c2440a01bd
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/7.0.0 CPython/3.13.14
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Sep 13, 2026.
Transparency log