biodalliance: alternative start page to avoid hardcoding a reference

[mgcam]

given sample accession number or similar, get the reference from mongo

[mgcam]

A good analysis and suggested way forward by @jmtcsngr:

With the current situation we are using a static page for each reference when starting a genome browser. So it is possible to add tracks to that browser using a specially crafted url which includes the accession. This is not friendly because if a user wants to visualise a data set the user needs to know in advance the reference needed. It will be better if the user just needs the accession and we provide the corresponding reference.

So we are thinking about something to solve the problem:

Having the accession we can query the mongo database to find a reference. We can modify npg_ranger server so it supports an extra url which allows for queries regarding the reference. The response of the query (likely a JSON) can then include the path/name of the reference needed.

We then need to create a small user interface (in a web page) which allows for the user to input the accession and use the service to resolve the reference. Once the reference is found, access to the corresponding fai file can give general information about regions available in that reference. With the reference and regions it will be possible to provide default start-end if the user did not provide them, do basic validation of chromosomes and maybe regions. With reference, chromosome and region we can initialise a genome browser object to display the data the user requested. The user can then add extra tracks for that reference or start over again with a new accession.