search

jefferis / vfbr

R Package to Access the Virtual Fly Brain project API
http://jefferis.github.io/vfbr/
2 stars 1 forks source link

vfb_nblast('VFB_00000001', target="GMR-Gal4", n = 10) not returning valid VFB ID's #10

Closed Robbie1977 closed 6 years ago

Robbie1977 commented 6 years ago 
      id  score
1   <NA> 0.5375
2   <NA> 0.5124
3   <NA> 0.4930
4   <NA> 0.4894
5   <NA> 0.4648
6   <NA> 0.4431
7   <NA> 0.4423
8   <NA> 0.4393
9   <NA> 0.4230
10  <NA> 0.4226
jefferis commented 6 years ago

It doesn't work for GMR nblast. It does for flycircuit single neuron.

jefferis commented 6 years ago

@Robbie1977 I assume it should also return VFB ids for GMR lines?

jefferis commented 6 years ago

Let's assume I have 22A05. What solr query should I use to get GMR ids @Robbie1977?

Robbie1977 commented 6 years ago

should be the same as with flycircuit but with asterisks surrounding *22A05*

Robbie1977 commented 6 years ago

wondering if it might be better to use neo4j for this as can specify exact types you want returning. This should be achieved in solr by filtering on id.

The issue is avoiding genetic components being returned as well as the individual image

Robbie1977 commented 6 years ago

@jefferis isn't this already returning the list: https://github.com/jefferis/vfbr/blob/master/R/vfbids.R#L71

Robbie1977 commented 6 years ago

@jefferis you're just not using them (gmr_vfbid) in https://github.com/jefferis/vfbr/blob/master/R/vfb_nblast.R#L60

jefferis commented 6 years ago

Hmm, actually I currently use queries against label_s for flycircuit ids since no asterisks required.

jefferis commented 6 years ago

Take a look at 8a3b39bc5912b2de1158c71cc82bf2ae7aceeefb and see if you are happy.

jefferis commented 6 years ago

More generally is this VFB id always the one that I want to return for a GMR id? It seems tied to the particular image for that e.g. VFB_00021621 = GMR_93D09_AE_01_03-fA01b.

Robbie1977 commented 6 years ago

Yep will work fine for now - a bit worried we might change labels to R10A09 format in longterm as Janelia have requested this. Also as we add more lines having GMR_ baked in might be an issue.

Robbie1977 commented 6 years ago

It is the Individual example Image but that was the example the NBLAST analysis was based on so I think it's more valid. We could return the Class but then would require extra steps to find an example image to display with target

Robbie1977 commented 6 years ago

@jefferis how's a fix coming? Hoping to get this live for neurofly.

jefferis commented 6 years ago

closed by 8a3b39b

Robbie1977 commented 6 years ago

getting the occasional error:

replacement has 39 rows, data has 40

In call:
`[[<-.data.frame`(`*tmp*`, "id", value = c("VFB_00021833", "VFB_00021915", 
"VFB_00021465", "VFB_00022468", "VFB_00021992", "VFB_00021621", 
"VFB_00021516", "VFB_00021575", "VFB_00022546", "VFB_00022215", 
"VFB_00022228", "VFB_00022249", "VFB_00021175", "VFB_00021216", 
"VFB_00021295", "VFB_00021283", "VFB_00021133", "VFB_00020527", 
"VFB_00020466", "VFB_00020602", "VFB_00020924", "VFB_00020861", 
"VFB_00020841", "VFB_00020880", "VFB_00020719", "VFB_00020734", 
"VFB_00020670", "VFB_00022975", "VFB_00020372", "VFB_00023253", 
"VFB_00023243", "VFB_00023052", "VFB_00023044", "VFB_00023070", 
"VFB_00022614", "VFB_00022663", "VFB_00022807", "VFB_00023142", 
"VFB_00023199"))