Issue #160 #163
Issue #160 #163
Conversation
|
Hmm, this results in: and lots of duplicate ids for each neuron path. The path features shouldn't have |
|
@napakalas this is now all looking good in terms making a rat map without duplicate ids turning up -- over to you to see if the inconsistent geojson ids are resolved. |
|
I’m generating all the maps now and will let you know once they’re ready. |
|
Confirm that the results are now consistent between ./annotations and ./pathways. All paths in all maps now have connectivity, and they all load correctly in CQ. However, there is missing node information that needs to be discussed (new issue?). For example, in the case of This is consistent and correct. These nodes are available in |
|
The missing node issue only occurs in the human maps (male and female); other maps are fine. |
|
@napakalas with and randomly checking a few of these ids in Why have different geometric shapes been identified as being the same?? Fixing the ids in the SVG should resolve the missing node issue. |
|
Yeah, those should be fixed by differentiating left and right in the flatmap source. I’ll fix that and create a PR. The case I mentioned earlier is different. It’s related to elements with the .group markup, such as: |
|
Here is the difference between the renderings of aacar-15 in the human and rat maps: the related spines are not highlighted in the human map, but they are highlighted in the rat map. male human: rat: |
|
Some objects, such as the white communicating ramus, do not have ontology terms that specify laterality. As a result, both left and right positions share the same ontology term (e.g., ILX:0793209), but they are given different IDs (S43_T3-Right and S43_T3-Left). 
|
Maybe this is a question for @nickerso? Or do we have a convention of say keeping to the left? Or simply continue being random? I suspect that we've now dealt with the original #160 issue and this PR can be merged. There though are three new issues that need to be raised and resolved:
|
@napakalas please test this and see if it resolves the inconsistent geojson ids.
An error is now logged when attempting to add a feature with a duplicate id, and we make sure mapmaker doesn't internally assign duplicate ids.