-
Notifications
You must be signed in to change notification settings - Fork 647
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
363 - Check Writers all still work #1015
Comments
XTC and TRR are still OK |
I'm looking at GRO -- should I'm getting a traceback indicating that |
Yeah so |
Looks like the 363 branch doesn't currently support setting resid values of arbitrarily large size?
|
Hmm, that looks like you're overflowing a |
My vote is for the core code to handle the data type promotion automatically. Probably first by trying to promote to Although in practice the fallbacks won't likely ever occur, so performance won't be affected in any substantial way, it does enable testing flexibility and general robustness. Looks like it is this block of code in 193 class GroupBase(_MutableBase):
194 """Base class from which a Universe's Group class is built.
195
196 """
197 def __init__(self, ix, u):
198 # indices for the objects I hold
199 self._ix = np.asarray(ix, dtype=np.int64) Conversely, if enough people object I can just remove some digits from the original unit test (which I wrote a few months ago). |
I would say that int64 is more then big enough for the not to distant future. >>> np.iinfo(np.int64)
iinfo(min=-9223372036854775808, max=9223372036854775807, dtype=int64) I don't think we would be close to this number simulating cells with about a Quintilian different atoms. Also there is no int128 in numpy (Which would go up to ~1.7e38 that is roughly enough to number every atom in every human being, assuming each human has ~ 100 e26 atoms [if they weight around 100 Kg]) |
So I just checked your test number in the gro-file test. It is really to large for an int64. I think you can safely change the number to |
Ok, fair enough. It is mostly designed to test truncation behaviour, not any real number of particles. |
@richardjgowers What behaviour do you want for |
@tylerjereddy sorry, what I meant by blank was without any topology attributes, so for example with GRO I didn't assume there'd be resnames and caught and warned about the output. |
This is an umbrella issue for all the Writer checks for the 363 merge
All writers should still work, regardless of the AtomGroup/Universe supplied (as before). This might (now) necessitate some guessing of attributes, (PDBWriter requires
resids
, these might not exist).For this, each Writer need checking! Each Writer has a card in the project board. When you start checking a writer, move the card to in progress to avoid duplication of efforts. If a Writer proves stubborn, you can convert the card into an issue and raise the problem.
Requirements:
core.groupbase.make_Universe
to generate Universes with arbitrary attributes)!The text was updated successfully, but these errors were encountered: