Keep MC running even when solver fails.

[lptolik]

Code that prevents MC calculation from stop when solver return non-success status.

[coveralls]

Pull Request Test Coverage Report for Build 17582725066

Details

• 7 of 14 (50.0%) changed or added relevant lines in 4 files are covered.
• 1 unchanged line in 1 file lost coverage.
• Overall coverage decreased (-0.1%) to 54.503%

---

Changes Missing Coverage	Covered Lines	Changed/Added Lines	%
src/types.jl	0	1	0.0%
src/monte_carlo.jl	4	6	66.67%
src/simulate.jl	3	5	60.0%
src/solution_interface.jl	0	2	0.0%

Files with Coverage Reduction	New Missed Lines	%
src/solution_interface.jl	1	18.84%

Totals
Change from base Build 17380559039:	-0.1%
Covered Lines:	696
Relevant Lines:	1277

---

💛 - Coveralls

[coveralls]

Pull Request Test Coverage Report for Build 17582725066

Details

• 7 of 14 (50.0%) changed or added relevant lines in 4 files are covered.
• 1 unchanged line in 1 file lost coverage.
• Overall coverage decreased (-0.1%) to 54.503%

---

Changes Missing Coverage	Covered Lines	Changed/Added Lines	%
src/types.jl	0	1	0.0%
src/monte_carlo.jl	4	6	66.67%
src/simulate.jl	3	5	60.0%
src/solution_interface.jl	0	2	0.0%

Files with Coverage Reduction	New Missed Lines	%
src/solution_interface.jl	1	18.84%

Totals
Change from base Build 17380559039:	-0.1%
Covered Lines:	696
Relevant Lines:	1277

---

💛 - Coveralls

[coveralls]

Pull Request Test Coverage Report for Build 17582725066

Details

• 7 of 14 (50.0%) changed or added relevant lines in 4 files are covered.
• 1 unchanged line in 1 file lost coverage.
• Overall coverage decreased (-0.1%) to 54.503%

---

Changes Missing Coverage	Covered Lines	Changed/Added Lines	%
src/types.jl	0	1	0.0%
src/monte_carlo.jl	4	6	66.67%
src/simulate.jl	3	5	60.0%
src/solution_interface.jl	0	2	0.0%

Files with Coverage Reduction	New Missed Lines	%
src/solution_interface.jl	1	18.84%

Totals
Change from base Build 17380559039:	-0.1%
Covered Lines:	696
Relevant Lines:	1277

---

💛 - Coveralls

[coveralls]

Pull Request Test Coverage Report for Build 17582725066

Details

• 7 of 14 (50.0%) changed or added relevant lines in 4 files are covered.
• 1 unchanged line in 1 file lost coverage.
• Overall coverage decreased (-0.1%) to 54.503%

---

Changes Missing Coverage	Covered Lines	Changed/Added Lines	%
src/types.jl	0	1	0.0%
src/monte_carlo.jl	4	6	66.67%
src/simulate.jl	3	5	60.0%
src/solution_interface.jl	0	2	0.0%

Files with Coverage Reduction	New Missed Lines	%
src/solution_interface.jl	1	18.84%

Totals
Change from base Build 17380559039:	-0.1%
Covered Lines:	696
Relevant Lines:	1277

---

💛 - Coveralls

[coveralls]

Pull Request Test Coverage Report for Build 17582725066

Details

• 7 of 14 (50.0%) changed or added relevant lines in 4 files are covered.
• 1 unchanged line in 1 file lost coverage.
• Overall coverage decreased (-0.1%) to 54.503%

---

Changes Missing Coverage	Covered Lines	Changed/Added Lines	%
src/types.jl	0	1	0.0%
src/monte_carlo.jl	4	6	66.67%
src/simulate.jl	3	5	60.0%
src/solution_interface.jl	0	2	0.0%

Files with Coverage Reduction	New Missed Lines	%
src/solution_interface.jl	1	18.84%

Totals
Change from base Build 17380559039:	-0.1%
Covered Lines:	696
Relevant Lines:	1277

---

💛 - Coveralls

[ivborissov]

@lptolik Thanks a lot for discovering this issue and for proposing a fix. We discussed it with @metelkin and reviewed how our current workflow is intended to work. At the moment, we save simulation results even when they don’t finish with a successful retcode (e.g. interrupted by maxiter or other termination conditions).

As you pointed out, this approach fails if termination happens at the very first step and no simulation points are saved (empty vectors). We see two possible solutions:

1. Your proposal: save results only if the simulation finishes with a successful retcode.
2. Our original design: save interrupted simulations as well, but handle the case where no timepoints are saved. This case is rare for ODEs (where we at least save at t=0), but it can occur in DAEs when the solution at t=0 is unknown.

We’re leaning towards option 2, since we have modeling scenarios where interrupted simulations still provide valuable information (e.g. tumor growth reaching a stopping condition). We’ve implemented this alternative fix here: #72 . Let us know what you think about this approach.

.