EPIC: Review/update connectivity queries

[dosumis]

We now have decent, working connectivity queries but should review and tweak.

Example:

https://v2.virtualflybrain.org/org.geppetto.frontend/geppetto?id=VFB_jrchjte4&i=VFB_00101567,VFB_jrchjte4

TermInfo:

Suggestions:

1. remove display of connectivity as relationships (better to just have queries)
2. Improve query text

->

• Show connected regions
• Show connected neurons

neuron-region query results

Suggested header changes

• _Downstream_Synapses -> outputs (all)

• _Tbars -> outputs (T-bars)

• _Upstream_Synapses -> Inputs

(Note - at time of ticket writing, this upstream column is empty - but this should be fixed soon see #297)

Neuron-neuron queries

Looks pretty good now - but here's some poss changes

1. Suggested header changes:

• Type -> Partner_neuron_type

• _Downstream_Synapses  -> outputs

• _Upstream_Synapses -> inputs

????

2. Suggested additional column:
   • Partner neuron symbol
   • Partner_neuron_gross_type - this could be especially useful if we add more gross cell types - e.g. local interneuron; Kenyon cell? MBON?

[dosumis]

@admclachlan @mmc46 assigning to you to review/comment on suggestions before @Robbie1977 implements any changes.

[admclachlan]

Suggestions:

TermInfo:

• Connectivity displayed as relationships: Remove
• Query text: Suggested text possibly ambiguous?
• How is: 'Show per-region synaptic connectivity weights', 'Show per-neuron synaptic connectivity weights' (maybe too long).

Neuron-Region:

• Headers: Does 'outputs' and 'inputs' makes it sound like this is a count of neurons instead of individual synapses?
• Images: always offset to the right of the header? makes the table look a bit messy. Might be better to centre all columns?

Neuron-Neuron:

• Type -> Partner_neuron_type: less ambiguous, could also do the same in neuron-neupil table?
• outputs/inputs: also think this is possibly ambiguous?
• partner_neuron_gross_type: potentially very useful

[admclachlan]

Also for neuron-region:

• can we show images even if they are not from the same template? Would be better than having some blank results in the images column?
• using non-primary rois leads to overlap and potential confusion. For the example neuron above, the connectivity table gives three rows all with 3215 down and 128 up because the three overlapping regions shown are reporting the same connectivity. This makes it look like there is much more connectivity than there actually is.
• blocked regions: some regions are blocked from being displayed as they don't have images, but I still think it's useful to see the connectivity even if users can't see the regions.

Part of solving this may be to use the primary, non-overlapping rois but I'm not sure if those have full coverage.

[mmc46]

For Term info

Show connected regions

Think "Show connectivity per region" would be better.

Show connected neurons

Is fine.

Neuron Region

Headers: Does 'outputs' and 'inputs' makes it sound like this is a count of neurons instead of individual synapses?

Inputs and outputs would be fine, these words are usually used to mean the synapses.

Images: always offset to the right of the header? makes the table look a bit messy. Might be better to centre all columns?

I think all columns are left aligned by default on the site, but making the images ones centred would definitely be better.

can we show images even if they are not from the same template? Would be better than having some blank results in the images column?

Not sure on this one. Same template would be better, if people want to display them in conjunction with the neurons. Not showing one makes it clear that there isn't one in that template, whereas showing another template would make it confusing. I guess the regions could be 'merged' together to generate others, 'MB(R)' for example, though I'm not sure how difficult that is.

using non-primary rois leads to overlap and potential confusion. For the example neuron above, the connectivity table gives three rows all with 3215 down and 128 up because the three overlapping regions shown are reporting the same connectivity. This makes it look like there is much more connectivity than there actually is.

Related to the above missing images. I think it's fine for now, better to be explicit than to try and work with the region hierarchy.

blocked regions: some regions are blocked from being displayed as they don't have images, but I still think it's useful to see the connectivity even if users can't see the regions.

Which ones are these? What's the difference between the ones without images that are shown and these?

Neuron-Neuron:

Type -> Partner_neuron_type:

Yes.

outputs/inputs: also think this is possibly ambiguous?

As mentioned in Neuron-Region, this is fine.

Partner neuron symbol

I'm not convinced this would add much to the table.

Partner_neuron_gross_type

This one would be useful, depending on space. It would be good to see how it might squash the remaining columns, and how that looks.

[dosumis]

Decisions:

• [x] Delete relationships on term info - see https://github.com/VirtualFlyBrain/Management/issues/279
• [x] query text - we realised that the text needs to include the term name as it appears elsewhere (e.g. in results table header, where context is not obvious)=>
  • Show neurons connected to X;
  • Show connectivity per region for X.
• [x] Headers: As in first comment except 'outputs (all)' -> outputs
• [x] Alignment: centre all columns
  5. Additional columns:
     • [ ] for now just add gross type on neuron-neuron. Might want consider symbol later as could be useful for non Hemibrain neurons.

[dosumis]

@Robbie1977 - Wondering about progress in changing query text?