Read single precision fluid field in RAMSES#5207
Read single precision fluid field in RAMSES#5207AnatoleStorck wants to merge 16 commits intoyt-project:mainfrom
Conversation
|
Hi! Welcome, and thanks for opening this pull request. We have some guidelines for new pull requests, and soon you'll hear back about the results of our tests and continuous integration checks. Thank you for your contribution! |
chrishavlin
added
enhancement
|
For reviewers: testing is complicated as all the RAMSES datasets at hand are stored in double precision! |
chrishavlin
left a comment
chrishavlin
left a comment
There was a problem hiding this comment.
LGTM! left some minor coding style comments...
| fields = [ "Potential,f", "Potential-old,f", "x-acceleration,f", "y-acceleration,f", "z-acceleration,f" ] | ||
|
|
||
| Importantly, the order of the fields should match the order in which they are written in the RAMSES output files. | ||
| yt will also assume that each entry is formed of a field name followed by its type (``f`` for single precision, ``d`` for double precision), separated by a comma. Field names containing commas are not supported, other precisions (e.g. integers) are not supported either. All fields in a given file type (gravity, hydro, ...) need to have the same precision. |
There was a problem hiding this comment.
is it worth a mention that though yt can read in the single precision case, they will be converted to double precision for use within yt?
| DETECTED_FIELDS = {} # type: ignore | ||
|
|
||
|
|
||
| def check_field_precision(fields: list[tuple[str, str]]) -> bool: |
There was a problem hiding this comment.
Since this function will only ever return True or throw an error (i.e., it will never return False), I'd suggest changing it a bit. Rename it to validate_field_precision and remove the bool returns.
| bool | ||
| True if all fields have the same precision, False otherwise. | ||
| """ | ||
| if not fields: |
There was a problem hiding this comment.
if you're using this to check for a list of length 0, can you change this to check for len(fields) == 0 explicitly? if fields can actually be None, then the typing should be updated to list[tuple[str, str]] | None.
| if all fields are single precision, False if double precision. | ||
| """ | ||
| field_names = [field for field, _ in fields] | ||
| check_field_precision(fields) |
There was a problem hiding this comment.
oh hey, you're already ignoring the output bool.
| # Case 2: fields are provided by users in the config | ||
| fields = cls.load_fields_from_yt_config() | ||
| ok = len(fields) > 0 | ||
| # Are fields are provided by users in the config? |
There was a problem hiding this comment.
| # Are fields are provided by users in the config? | |
| # Are fields provided by users in the config? |
3 participants
PR Summary
The current back-end for reading RAMSES fluid fields assumes the data is double precision. This PR (in collaboration with @cphyc ) is an initial expansion of the RAMSES fluid reader to read single precision fields. For my simulations, the RT fields are written in single precision so this change is needed to read them in.
PR Checklist