Closed nathandunn closed 7 years ago
Should just be a matter of adding .csv
to the other file-types for pass-through and caching.
Shouldn't be too hard to fix or cache.
So, the problem here is that we want to use an identical csv include for both JBrowse and Apollo.
The include URL looks like this:
{ "type": "csv", "url": "data/Nvec_v10/includes/metadata.csv" }
I can usually get one or the other to work, but we really want a single trackList.json for all of them.
It seems to work, but getting the JBrowse and the Apollo URL to sync is a bit difficult.
Getting:
INFO: GroovyPagesServlet: "/WEB-INF/grails-app/views/jbrowse/data.gsp" not found
Organism directory:
/var/other_data/jbrowse/data/Nvec_v10/
JBrowse link is:
https://XXX/jbrowse/?data=data/Nvec_v10
The 404 in Apollo:
GET http://XXX:8080/apollo/17243046418848101441634387168/jbrowse/data/Pmar_v11/includes/pmar_metadata.csv 404 (Not Found)
and the trackList.json is (which works for JBrowse and should work for Apollo):
{ "type": "csv", "url": "data/Nvec_v10/includes/metadata.csv" }
FYI @srobb1
So I defined a similar case to reproduce. Created the root directory:
/opt/apollo/
Yeast is:
/opt/apollo/yeast
and the include is:
/opt/apollo/yeast/include/myTrackData.csv
In JBrowse I do a symlink:
ln -s /opt/apollo/ data
This works for JBrowse (http://localhost/jbrowse/?data=data%2Fyeast) localhost and fails for Apollo:
{ "type": "csv", "url": "data/yeast/include/myTrackData.csv" }
This works for Apollo and fails for JBrowse:
{ "type": "csv", "url": "data/include/myTrackData.csv" }
The problem is that this include should be exploded to be data/yeast instead of data if the initial is not found as data
is a keyword in JBrowse / Apollo.
It fails because it replaces "data" with the proper absolute path. In this case we're provided with the proper path, which includes data
.
params.path: yeast/include/myTrackMetaData.csv directory /opt/apollo/yeast/
File not found: /opt/apollo/yeast//yeast/include/myTrackMetaData.csv
Should be:
/opt/apollo/yeast/include/myTrackMetadata.csv
Fix should be if not found and start and end are the same, just remove the prefix of the second one.
This should be working (on branch) using an include:
{ "type": "csv", "url": "data/yeast/include/myTrackMetaData.csv" }
If the file is not found and it detects an overlap, it automatically merges it.
====
Getting a 404 on this:
http://XXX:8080/apollo/17243046418848101441634387168/jbrowse/data/Pmar_v11/includes/pmar_metadata.csv 404 (Not Found)
From metadata loaded this way:
http://gmod.org/wiki/JBrowse_Configuration_Guide#Loading_Track_Metadata_from_Files