minor updates

Author: matthiaskoenig

src/sbmlsim/fit/TODO.md

@@ -0,0 +1,26 @@
+# TODO
+Important features:
+- support simple cluster deployment (full serialization of problem & settings)
+- (support complex calculated data AUC, ...)
+
+*IO*
+- [ ] serialization of problem(s) to PEtab (v2) + additional metadata (i.e. all simulations, data, metadata); i.e. the complete problem is stored in PEtab; store problem and metadata (for filtering & validation) => full problem => create subset problems
+- [ ] inject model parameters/changes and serialize SBML
+- [ ] generation of subsets of problems easily (filter) => new PEtab problems; all serialized; 
+
+*Optimization*
+- simulate only timepoints which are necessary
+- clear access to optimization function;
+- robust handling of failing optimizations;
+
+* Evaluation & reports*
+- simple calculation of metrics based on a set of paramters; i.e PETab problem + parameter set allow to calculate all metrics
+  - [ ] Calculate metrics: PRED, IPRED, IRES, AIC, RMSE (based on given set of parameters) and report; Access to cost function;
+- store results of parameter optimization (parameter sets with metrics) -> merge results
+- confidence intervals using profile Likelihoods of cost function; => subsequent analysis
+- improve plots; Bland Altmann; Waterfall
+- separate reports from parameter optimization (just needs problem definition & model parameters); Creates report for given parameter set, trainings data & evaluation data
+- interactive reports (quarto); => use best of both worlds R, nlmixr2 , Pypesto; petab; much better interactive/static reports as overview
+
+Profile-Likelihood analysis
+- Parameter identifiability analysis;