Casper da Costa-Luis
Casper da Costa-Luis
Yup that explains why the first entry `new[0,0,0] == old[0,0,0]` but not the rest.
I pushed it here; it's a no-conflict rebase.
> Latest performance comparison (with #1317 + [UCL/STIR#1592](https://github.com/UCL/STIR/pull/1592) + #1314) - the last column shows timings we can get with `asarray`. > > ``` > test sirf as_array + numpy...
> what is the minimum numpy version for this? There is no minimum; the API is very old.
Is this ready for merge @KrisThielemans @evgueni-ovtchinnikov? Are we ok with the tests? Also I'm unsure which line(s) need `STIR_VERSION` checks or if this PR really depends on https://github.com/UCL/STIR/pull/1592
> Careful both. conflicting updates and rebases again I believe We're on a call no worries. Think it's fixed now.
Argh missed `--non-interactive` in my debug commit :see_no_evil: > `docopt` should really print the failing command line We should not be using `os.system` to run things. At a minimum we...
@microsoft-github-policy-service agree
Ah @afourney your question(s) imply a fundamental misunderstanding. Clearly some docs are missing here. Unsure about how much detail you'd prefer to see in the GitHub flavoured markdown README, though....
ok rebased & removed `shtab` now