[v14 - Ready] Fix bug where expand_registered_functions mutated original reaction system
#929
Conversation
src/registered_functions.jl
Outdated
| end | ||
| return expr | ||
| end | ||
| # If applied to a Reaction, return a reaction with its rate modified. | ||
| function expand_registered_functions(rx::Reaction) | ||
| Reaction(expand_registered_functions(rx.rate), rx.substrates, rx.products, | ||
| Reaction(expand_registered_functions!(deepcopy(rx.rate)), rx.substrates, rx.products, |
There was a problem hiding this comment.
Is this appropriate use of deepcopy? I could not come up with a better way to do it. If we really want to avoid it it would be possible to first check the expression whether there are any registered functions, and only apply the function in this case
There was a problem hiding this comment.
Also, I think technically this by could be rewritten using @set (since expanding functions really shouldn't change anything more of the reaction), however, I was reluctant to introduce any more changes than necessary.
There was a problem hiding this comment.
It would be more memory efficient to first check if the given expression would be expanded, and only then deepcopy and expand, but otherwise this is fine as a less memory efficient version.
There was a problem hiding this comment.
(Note, I'm not saying this needs to be done in this PR, just that it would be good to switch to that eventually.)
TorkelE
changed the title
expand_registered_functions mutated original reaction systemexpand_registered_functions mutated original reaction system
src/registered_functions.jl
Outdated
| end | ||
| return expr | ||
| end | ||
| # If applied to a Reaction, return a reaction with its rate modified. | ||
| function expand_registered_functions(rx::Reaction) | ||
| Reaction(expand_registered_functions(rx.rate), rx.substrates, rx.products, | ||
| Reaction(expand_registered_functions!(deepcopy(rx.rate)), rx.substrates, rx.products, |
There was a problem hiding this comment.
It would be more memory efficient to first check if the given expression would be expanded, and only then deepcopy and expand, but otherwise this is fine as a less memory efficient version.
|
Sorry, totally forgot about this one. I think I will add the checks, it should be fairly straightforward with the new symbolic functions. |
|
Ok, I have changed around the internals (no longer using |
Co-authored-by: Sam Isaacson <isaacsas@users.noreply.github.com>
Previously
expand_registered_functionswould unintentionally apply the operation on the input system. E.g.would result in both
rsandrs_expanded_funcshaving their functions expanded. This fixes this, and adds a test to ensure this remains the case.