Updates for upstream breaking changes #290
Conversation
|
Should do the compats too? |
Co-authored-by: "Samuel Isaacson" <isaacsas@users.noreply.github.com>
|
latexify failed again?! |
|
Looks like you forgot to evaluate the second reaction network in the file before generating the third and fourth test? |
| @@ -13,7 +13,7 @@ Reexport = "189a3867-3050-52da-a836-e630ba90ab69" | |||
| Catlab = "0.7.3, 0.8, 0.9, 0.10" | |||
| Latexify = "0.13.5, 0.14" | |||
| MacroTools = "0.5.5" | |||
| ModelingToolkit = "4.4.1" | |||
| ModelingToolkit = "5" | |||
There was a problem hiding this comment.
Bump Catalyst too for a new release?
test/latexify.jl
Outdated
| \ce{ \varnothing &<=>[\frac{p_{a} \left( B\left( t \right) \right)^{n}}{k^{n} + \left( B\left( t \right) \right)^{n}}][d_{a}] A}\\ | ||
| \ce{ \varnothing &<=>[p_{b}][d_{b}] B}\\ | ||
| \ce{ 3 B &<=>[r_{a}][r_{b}] A} | ||
| \ce{ \varnothing &->[K1^{n1} v1^{2} \left( \mathrm{X4}\left( t \right) \right)^{n1} \mathrm{inv}\left( K1^{n1} + \left( \mathrm{X2}\left( t \right) \right)^{n1} \right) \mathrm{inv}\left( K1^{n1} + \left( \mathrm{X4}\left( t \right) \right)^{n1} \right)] X1}\\ |
There was a problem hiding this comment.
This is for the wrong reaction network, first evaluate the network on line 78,
r = @reaction_network begin
(hill(B, p_a, k, n), d_a), 0 ↔ A
(p_b, d_b), 0 ↔ B
(r_a, r_b), 3B ↔ A
end p_a k n d_a p_b d_b r_a r_bthen generate this test and the last test.
isaacsas
changed the title
|
If tests now pass I think this should be gtg. |
|
@shashi seems like I'm now seeing crashes in a non-Latexify test. (The SDE test.) Does that pass for you? |
|
@TorkelE We seem to have a reproducible test failure here with the latest ModelingToolkit and SymbolicUtils in the SDE tests. Could you take a look? |
|
FWIW, the test passes locally on my computer when I just include the SDE test file, but fails when I run |
|
Weird, bussy today, but will hopefully be able to take a look tomorrow. |
…ng generated functions...
|
@TorkelE I dropped checking the solution in those SDE tests, and instead switched to just testing if the generated |
|
I was a bit ambitious and figured that for a solution over a long-enough timeframe the means and stds would be similar enough, but if the f and g functions are identical there should be no problem (and if there is, it's unlikely to be in this package). Doing it like this is good to me, especially when it saves us some headache. And as you noted, the test was bugged anyway. |
|
@shashi finally done! Thanks for all your help migrating us over! |
needs SciML/ModelingToolkit.jl#720 and the latest release of SymbolicUtils.jl