Force a time out on long running blast.pdb() jobs
Running blast.pdb() on large databases when the NCBI server load is high can take an unreasonable amount of time. I suggest we have an optional time.out=90 seconds with a warning and a link to the server results.
Comments (7)
-
-
reporter because I run it in a script and want it to exit so I am can continue without babysitting...
-
How about setting time.out=90 and try=3? We try three times to re-connect to the server and then stop if all of them fail.
-
reporter Sounds good. In the error msg we can also print-out the (broken) result URL.
-
the timeout is a bit annoying since I anyway need to fetch the blast results in order to continue the analysis. can we have no timeout as default?
-
reporter OK, let's have NULL or -1 as default but keep the option to timeout. I have been using the HMMer 3 server quite a bit recently and think it could make a nice alternative for single sequence and 'fasta' 'pdbs' multiple sequence search.
-
reporter - changed status to resolved
This is now implemented along with other improvements on the 'feature_pdb_dataframe' branch
- Log in to comment
hmm.. why would timing out be helpful? I anyway need the blast results, and if I don't I can just Crtl-C to stop it. Instead, could we write some meaningful output to the user after some time with the never ending dots (e.g. some server results..?)