paul-florentin-charles / bcn-rainfall-models Goto Github PK

Check this link.
From the above link:

• SPIs for short accumulation periods (e.g., SPI-1 to SPI-3) are indicators for immediate impacts
  such as reduced soil moisture, snowpack, and flow in smaller creeks;
• SPIs for medium accumulation periods (e.g., SPI-3 to SPI-12) are indicators for reduced stream
  flow and reservoir storage; and
• SPIs for long accumulation periods (SPI-12 to SPI-48) are indicators for reduced reservoir and
  groundwater recharge, for example

• Organize python files differently
• Rename folders consequently
• ...

Find an already existing (variance) or build a function to retrieve the "dispersion" of rainfall data for the whole dataframe or within an optional specific timeframe.

As per usual, the function would most likely be a class function of YearlyRainfall.

It could be useful for monthly or seasonal data, to find which are the most unpredictable and which are the stablest ones.

We could do that by storing data as csv and loading it.
We could also do that by avoiding loading elements many times for nothing (loading once the dataset is enough)

So they will be easier to manipulate; notably to add Pydantic or marshmallow models as body params or responses.

Title says it all...
Not a user-friendly feature, but could be nice for retrieving some data fastly.
For instance: python cli.py rainfall:average 1970 1999

This will notably allow to use:

• "|" instead of "Union"
• Match/case statement

Following errors ought to be managed:

• Time mode is monthly but no month is set
• Time mode is seasonal but no season is set

Y-axis label of rainfall normals plot should refer to a percentage and not a mm value.
--> Check plotting decorator to fix the mistake

As a reminder... (Since models are not done yet)

CSV export only allows export of all data; would be nice to:

1. be able to restrict it to specific dates.
2. be possible to restrict it to given columns.

Flasgger, unfortunately, does not seem to be compatible with most recent versions of Python.
We could simply use FastAPI to set up our API, and keep Flask later as a back supporting our front version.

Operate the following transformations for functions plot_rainfall(...) and plot_normal(...) in class YearlyRainfall

show=False -> show: Optional[bool] = False

Experiment and integrate Plotly for rainfall data visualisation.

Nothing to declare.

Optional parameters allow flexibility but overusing them might complexify futur usages of the core classes & the API that is being built. It also makes the behaviour less predictable sometime.

Code a generic plot function that would take a label for the X values and an array of labels for all the Y values to plot.
Optional parameters to decide the kind of plot (curve, scatter, histo, etc.) , colors and more are to be added as well.

Instead of using "/".join([...]) or f-strings.

Long term issue.
Might be interesting to store data in a lightweight db instead of directly using external data.

Make them as strings for easier treatment via Fast API
-> Will imply to change how MonthlyRainfall/SeasonalRainfall models are instantiated.

More a project than an actual issue

• Use Flask combined with Flasgger (support for Swagger with Flask)
• API should be as light as possible
• API should allow retrieving rainfall data according to several criteria
  • after a defined year
  • by year/season/month
  • ...
• API should allow computing on rainfall data
  • K-Means
  • Linear Regression
  • ...
• API should retrieve data as plots & CSVs

The function should pass 4 arguments:

1. rainfall value (float)
2. operator('>' or '<')
3. begin year (int) [optional]
4. end year (int) [optional]