feat: Json

[Siel]

This PR aims to provide a JSON definition for models. This definition is compatible with EXA, this means JSON models can be compiled into Equation.

[mhovd]

Should we hold off on this to reduce the size of the PR?
And rather wait for normalized-rows to be merged into main, then target main for this PR?

[Siel]

Should we hold off on this to reduce the size of the PR? And rather wait for normalized-rows to be merged into main, then target main for this PR?

However you prefer, we can wait until normalized-api is merged on main. Right now it is like this because I need them all to run the latest version of papir.

[github-actions]

Bencher Report

Branch	feat/json
Testbed	mhovd-pgx

⚠️ WARNING: No Threshold found!

Without a Threshold, no Alerts will ever be generated.

• Latency (nanoseconds (ns))

Click here to create a new Threshold
For more information, see the Threshold documentation.
To only post results if a Threshold exists, set the --ci-only-thresholds flag.

Click to view all benchmark results

Benchmark	Latency	nanoseconds (ns)
Analytical vs ODE/One-compartment IV/Analytical	📈 view plot ⚠️ NO THRESHOLD	834.67 ns
Analytical vs ODE/One-compartment IV/ODE	📈 view plot ⚠️ NO THRESHOLD	17,874.00 ns
Analytical vs ODE/One-compartment oral/Analytical	📈 view plot ⚠️ NO THRESHOLD	829.86 ns
Analytical vs ODE/One-compartment oral/ODE	📈 view plot ⚠️ NO THRESHOLD	25,521.00 ns
Analytical vs ODE/Two-compartment IV/Analytical	📈 view plot ⚠️ NO THRESHOLD	919.78 ns
Analytical vs ODE/Two-compartment IV/ODE	📈 view plot ⚠️ NO THRESHOLD	26,154.00 ns
Analytical vs ODE/Two-compartment oral/Analytical	📈 view plot ⚠️ NO THRESHOLD	938.01 ns
Analytical vs ODE/Two-compartment oral/ODE	📈 view plot ⚠️ NO THRESHOLD	28,999.00 ns
Conditional dose modification	📈 view plot ⚠️ NO THRESHOLD	1,216.20 ns
Create large dataset (100 subjects)	📈 view plot ⚠️ NO THRESHOLD	60,034.00 ns
Data expand complex (1h intervals)	📈 view plot ⚠️ NO THRESHOLD	28,712.00 ns
Data expand simple (1h intervals)	📈 view plot ⚠️ NO THRESHOLD	506.81 ns
Data expand with additional time	📈 view plot ⚠️ NO THRESHOLD	39,700.00 ns
Filter exclude subjects	📈 view plot ⚠️ NO THRESHOLD	31,831.00 ns
Filter include subjects	📈 view plot ⚠️ NO THRESHOLD	7,985.90 ns
Modify all bolus doses	📈 view plot ⚠️ NO THRESHOLD	1,187.00 ns
Modify all infusion doses	📈 view plot ⚠️ NO THRESHOLD	1,218.00 ns
SubjectBuilder multi-occasion	📈 view plot ⚠️ NO THRESHOLD	268.57 ns
SubjectBuilder simple	📈 view plot ⚠️ NO THRESHOLD	103.42 ns
SubjectBuilder with covariates	📈 view plot ⚠️ NO THRESHOLD	288.20 ns
one_compartment	📈 view plot ⚠️ NO THRESHOLD	19,412.00 ns
one_compartment_covariates	📈 view plot ⚠️ NO THRESHOLD	25,816.00 ns
readme 20	📈 view plot ⚠️ NO THRESHOLD	300,990.00 ns
two_compartment	📈 view plot ⚠️ NO THRESHOLD	21,270.00 ns

🐰 View full continuous benchmarking report in Bencher