More flexible handling of Gaugings, stage time series or more generally input/output datasets

[benRenard]

For instance: edit imported data within BaRatinAGE, sort/search tools, add new columns, etc. Keeping in mind, for the longer term, that in the context of BaM these data could be anything.

As for figures we need to be realistic, we don't want to reinvent excel!

[benRenard]

linked to this: Gestion des en-têtes et format des fichiers d'import

J'oublie souvent de sauter une ligne d'en-tête quand je prépare mon fichier CSV jaugeages, et je perds le premier jaugeage sans toujours m'en apercevoir... C'est ma faute, mais je signale au cas où d'autres le font aussi. Ca se gèrera sans doute à un niveau plus général, avec la généralisation BaM.

[benRenard]

Also: how to handle missing values?

---

original request: nodata-gaps in the stage series should not be plotted with connecting lines in the graph

---

For the moment MV are allowed (in BaRatin/BaM) only for observed outputs (i.e. in the calibration dataset), using a MV value (-9999 I think by default). However missing data are not allowed in inputs, and I think this makes sense: a model cannot run without inputs, and given the generality of BaM, it cannot be its responsibility to decide how to fill it in.

However allowing missing data could maybe be allowed in BaRatinAGE: BaRatinAGE would then just remove them before calling BaM. But I'm not convinced it's worth the effort: how about dynamical models for instance?

[benRenard]

And another one: offer some pre-processing tools for input datasets? For instance subsampling tools would be particularly valuable - but is it BaRatinAGE's job, or should rather be done by some external app?

[benRenard]

Another one:

---

It may be cool that HYDRO HTVAR format and/or SANDRE XML format be accepted to import stage series. And maybe also that another flow series (and another rating curve?) can be imported and plotted for comparison of the BaRatin results with other results.

---

More generally: what formats are allowed for input/output datasets? If we decide to offer some translation utilities to convert between BaRatinAGE and external formats, then my feeling is that they should probably not be coded in java: I would favor a fortran executable that could also be called from the command line, or from R or anything else,

[IvanHeriver]

There are many ideas. Here is a summary:

• [ ] edit imported data within BaRatinAGE, sort/search tools, add new columns (in the context of BaM, data can be anything)
• [x] dealing with file headers when importing a file: fixed in BaRatinAGE v3.0
• [x] dealing with missing values: supported in BaRatinAGE v3.0
  • import with missing value is supported
  • calibration data with missing value is supported however, BaM won't run until the corresponding data point are deactivated
  • prediction data with missing value is supported. data points are discarded for BaM to make the prediction and reinserted afterwards
• [ ] add subsampling tools
• [ ] support for importing specific formats for input dataset:
  • [ ] HYDRO HTVAR format
  • [ ] SANDRE XML format
• [ ] support for exporting to specific formats:
  • [x] rating curve export to Barème format