Structure for applications

[sashahafner]

For @AnneMortensen 's files.

[sashahafner]

From Anne: https://github.com/AU-BCE-EE/tric-fil-mod/commit/6e61ea247bcebdd2821fe53ad41cafe924f8e53a

[sashahafner]

Looks like some good progress Anne, but remember that you need subdirectories for plots, output, and input data too. Could be something like this, with a separate subdir for each experiment or application:

applications
|-- exper01
     |--meas-data
     |--output
     |--scripts
     |--plots
|--exper02

[sashahafner]

Or maybe group by type of application, like this

applications
|--lab-experiments
|    |--exper01
|    |  |--meas-data
|    |   |--output
|    |   |...
|    |...
|--model-sensitivity
|    |--sim01
|    |    |--scripts

[sashahafner]

Ideally we want is a structure that:

• is easy to understand
• makes it easy to find particular results
• is reproducible, so we can re-run any earlier work
• is ordered in a way, so we can see/remember what work came first, and how the pieces might relate to each other

But don't worry too much about getting it perfect. Most important thing now is that you can find your way around.

[AnneMortensen]

Okay, thank you :-) Would you still suggest that all data treatment (raw data to processed data) refer back to one single file, like the way mod_funcs works? And how about plotting, would you copy and reuse that script, or make them all refer back to one file?

[sashahafner]

Okay, thank you :-) Would you still suggest that all data treatment (raw data to processed data) refer back to one single file, like the way mod_funcs works? And how about plotting, would you copy and reuse that script, or make them all refer back to one file?

Do you mean one data file? Or one script? Here are my suggestions. If the data are naturally in a single file, refer to it. If they are separate because they were from separate experiments, keep them separate (don't add unnecessary manual steps). If you mean scripts, keep them separate unless you want plan to analyze/plot all data together. The copy of mod_func is just to avoid multiple copies of the model itself--I would not apply the same concept to data analysis/plotting scripts.

[AnneMortensen]

I think this issue can be used for discussing the structure of the "comparison" graph structure as well. The primary issue is currently in "/Applications/Lab experiments/experiment 5"

Description of how I do it now:

1) "Data treatment functions/DT_func" takes files from "Raw_data" and makes files in "Processed_data" Manually run this file once per replicate, changing only the file name (28 times)

2) "/Scripts for plotting/Making_Compare_pH_data" takes files from "Processed_data" and makes files in "Compare_pH_data". Manually run this file once per sub-experiment (7times) both changing the the name of the "lab_parameters" file and the experiment number. (4 numbers changed pr. run)

3) "/Scripts for plotting/Plotting_pH_differences" takes files from "Compare_pH_data" and makes a plot and saves that in "Plots". Only runs once

I think we can do better than that, but I may need a little guidiance on how to approach it best...

[AnneMortensen]

New structure complete - this uses a master excel spreadsheet for laboratory parameters (time of start, pH ect.), and a loop so that all plots (experiment 2-6, profiles and concentration as a function of time) can be updated at once. Also did a folder called "Tasks" for taking a smaller amount of data and working on that, eg. to adjust k, which can then easily be updated in all plots at once.