Initial subject tracking implementation

[brosenberg42]

Issues:

• Add initial subject tracking implementation openmpf#1832

Related PRs:

• Feat/subject tracking openmpf#1833
• Add OrToolsSubjectComponent openmpf-components#378
• Initial subject tracking implementation #84
• Feat/subject tracking openmpf-docker#219
• Add protobuf build commands openmpf.github.io#191

---

This change is 

[jrobble]

Reviewed 8 of 8 files at r1, 2 of 2 files at r2, all commit messages.
Reviewable status: all files reviewed, 2 unresolved discussions (waiting on @brosenberg42)

---

subject/api/mpf_subject_api/__init__.py line 48 at r1 (raw file):

MediaId = NewType('MediaId', str)

DetectionComponentType = NewType('DetectionComponentType', str)

Please provide examples for the DetectionComponentType in a comment above this line.

---

subject/examples/PythonSubjectComponent/subject_component/__init__.py line 48 at r1 (raw file):

                entity = get_single_track_entity(track_id)
                entities.append(entity)
                relationships.append(get_relationship(entity, detection_job.media_id))

Please update this to create a relationship between this entity and the one before it. Having a relationship with two entities is more realistic.

[jrobble]

Reviewed 1 of 1 files at r3, all commit messages.
Reviewable status: all files reviewed, 1 unresolved discussion (waiting on @brosenberg42)

[brosenberg42]

Reviewable status: 7 of 9 files reviewed, 1 unresolved discussion (waiting on @jrobble)

---

subject/api/mpf_subject_api/__init__.py line 48 at r1 (raw file):

Previously, jrobble (Jeff Robble) wrote…

Please provide examples for the DetectionComponentType in a comment above this line.

Done.

---

subject/examples/PythonSubjectComponent/subject_component/__init__.py line 48 at r1 (raw file):

Previously, jrobble (Jeff Robble) wrote…

Please update this to create a relationship between this entity and the one before it. Having a relationship with two entities is more realistic.

That will result in every entity being transitively related to every other. It will also allow entities that are never in the same media to have a relationship. Half the code will exist just to address those issues. That means the component developer needs to figure out which half of the code to ignore. I think having a simple example is better than a realistic example. If a realistic example is required, then the or-tools component is a better example.

[jrobble]

Reviewed 2 of 2 files at r4, all commit messages.
Reviewable status: all files reviewed, 1 unresolved discussion (waiting on @brosenberg42)

---

subject/examples/PythonSubjectComponent/subject_component/__init__.py line 48 at r4 (raw file):

                entity = get_single_track_entity(track_id)
                entities.append(entity)
                relationships.append(get_relationship(detection_job.media_id, entity))

As discussed, please relate each entity back to one other entity. They can all be related to the same one, or hand-picked after the loop. Only important thing is that a relationship has two entities.

[brosenberg42]

Reviewable status: 8 of 12 files reviewed, 1 unresolved discussion (waiting on @brosenberg42 and @jrobble)

---

subject/examples/PythonSubjectComponent/subject_component/__init__.py line 48 at r4 (raw file):

Previously, jrobble (Jeff Robble) wrote…

As discussed, please relate each entity back to one other entity. They can all be related to the same one, or hand-picked after the loop. Only important thing is that a relationship has two entities.

Done.

[jrobble]

Reviewed 4 of 4 files at r5, all commit messages.
Reviewable status: all files reviewed, 1 unresolved discussion (waiting on @brosenberg42)

---

detection/api/mpf_component_api/timing.py line 53 at r5 (raw file):

class Timing:

Why is this part if the API and not the component util?

[brosenberg42]

Reviewable status: all files reviewed, 1 unresolved discussion (waiting on @jrobble)

---

detection/api/mpf_component_api/timing.py line 53 at r5 (raw file):

Previously, jrobble (Jeff Robble) wrote…

Why is this part if the API and not the component util?

Because it does not rely on any outside libraries. I think we should move everything in component util that does not rely on an external library to the api package. If you agree, I'll create an issue for that.

[jrobble]

Reviewable status: all files reviewed, 1 unresolved discussion (waiting on @brosenberg42)

---

detection/api/mpf_component_api/timing.py line 53 at r5 (raw file):

Previously, brosenberg42 wrote…

Because it does not rely on any outside libraries. I think we should move everything in component util that does not rely on an external library to the api package. If you agree, I'll create an issue for that.

That makes sense to me. Yes, please create an issue for it.

[brosenberg42]

Reviewable status: all files reviewed, 1 unresolved discussion (waiting on @jrobble)

---

detection/api/mpf_component_api/timing.py line 53 at r5 (raw file):

Previously, jrobble (Jeff Robble) wrote…

That makes sense to me. Yes, please create an issue for it.

This is the issue: openmpf/openmpf#1876