BD2KGenomics / ga4gh-integration-deprecated

Tracking for ga4gh-integration projects
1 stars 2 forks source link

Code demonstrating parallel API requests #26

Open david4096 opened 8 years ago

david4096 commented 8 years ago

Parallelize some requests in order to demonstrate how an API allows an application developer to balance connection.

For example, quickly generate a list of variants in a region by splitting that region into a series of requests for some window. For example, a parallelized request for variants between 0-5kb might be asked as 5 requests, each for 0-1kb, 1-2kb and so on.

diekhans commented 8 years ago

Note that the current API doesn't directly support partitioning range queries. Since the API queries for overlapping regions and variants may be longer than one base long, naively partitioning by window can result in the same variant being return multiple times.

The client would have to provide addition filtering to prevent multiple counting. An approach to this is adding an additional, non-overlapping criteria for using a variant. An easy way is to only use variants for a connection that start withing the given window.

David Steinberg notifications@github.com writes:

Parallelize some requests in order to demonstrate how an API allows an application developer to balance connection.

For example, quickly generate a list of variants in a region by splitting that region into a series of requests for some window. For example, a parallelized request for variants between 0-5kb might be asked as 5 requests, each for 0-1kb, 1-2kb and so on.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub, or mute the thread.*