Support for additional parameters to loadTrack

[mbookman]

A coming feature for Google Cloud Storage will be the ability to specify a billing project id for requests to GCS. We need a way to pass this project id in igv.browser.loadTrack().

The specified field is the userProject query parameter.

For reference, support for this in the google-cloud-python library is described here:

• Add support for "requester pays" operations googleapis/google-cloud-python#3474

and implemented here:

• Add '{Bucket,Blob}.user_project' properties. googleapis/google-cloud-python#3490
• Pass 'user_project' if set for Bucket API requests googleapis/google-cloud-python#3492

[jrobinso]

So if I add a configuration option that allows you to pass through any query parameter would that address the issue? Something like

query: "userProject=sflkelk&somethingElse=foobar"

This string would get appended to all requests for this track, with leading ? or & added as required.

[mbookman]

I would expect that would work.

[vanetten]

Any updates on this Jim?

[jrobinso]

@mbookman Have you tried just including the parameter in the url you set in the track configuration object? The url can include query parameters, I'm not sure we need anything else to support this.

[jrobinso]

You can now set the userProject parameter, or any other parameters you need, in the URL.

[javadnoorb]

Hi. How do you set the query parameter in the url? Does it work with uri as well?

[jrobinso]

@javadnoorb I don't understand your question, are you asking what the Google parameter name? That's probably a Google question, igv.js doesn't care it just passes through any parameters it finds.

[javadnoorb]

@jrobinso I'm trying to set payer project in jupyter-igv. The only Google option I can find is this. But this is based on URL (not URI) and requires extra JSON (or XML) file. So I don't see how this can be set in the URI. I'm asking this because IGV desktop has an implementation where you can set the payer project (here). I was hoping asking my question here would help find a similar solution in jupyter-igv. Also apologies since this thread is not directly related to jupyter version of IGV.

[jrobinso]

@javadnoorb I don't understand the distinction you are making between URI and URL, igv.js only supports http/https protocol so what is the difference?

In a URL you can set the user project as a query parameter: &userProject=[PROJECT_ID]

[javadnoorb]

@jrobinso I've tried &userProject=[PROJECT_ID] in the URLs without success. By URI I just meant the google cloud storage paths which start which 'gs://'. Maybe my question is more suited for jupyter-igv section (but I just assumed the backend is the same igv.js). There you can use the URIs directly and skip http/https, as in:

import igv
b = igv.Browser({"genome":'hg19})
b.load_track(
{
  'type': "alignment",
  'format': "bam",
  'name': "NA12889",
  'url': "gs://genomics-public-data/platinum-genomes/bam/NA12889_S1.bam",
  'indexURL': "gs://genomics-public-data/platinum-genomes/bam/NA12889_S1.bam.bai",
})

[jrobinso]

OK, there might be a bug then. Could you try the https equivalent url and see if it works? That's not a long term solution but is information. Yes igv-jupyter uses igv.js.

[jrobinso]

igv.js translates the gs url to its https equivalent, it might be stripping the parameter off in the process. You could watch this from the network tab.

[javadnoorb]

No, it doesn't work either way. To confirm, I tried a URL with the following structure:

https://storage.cloud.google.com/[BUCKET_NAME]/[PATH_TO_BAM_FILE]/my.bam?authuser=2&userProject=[PROJECT_ID]

I copied the first part of the path (https://storage.cloud.google.com/[BUCKET_NAME]/[PATH_TO_BAM_FILE]/my.bam?authuser=2) directly from google console.

FYI, I don't encounter this issue in the IGV Desktop. There I log in to my account, set the project and then set the URI and things work.