show_knmi.py

import pandas as pd
import streamlit as st
from datetime import datetime
import platform

    
from show_knmi_functions.utils import show_weerstations, help,  list_to_text,check_from_until, find_date_for_title, download_button,get_weerstations, get_data
from show_knmi_functions.does_rain_predict_rain import does_rain_predict_rain
from show_knmi_functions.polar_plot import polar_plot, polar_debug
from show_knmi_functions.show_warmingstripes import show_warmingstripes
from show_knmi_functions.show_plot import show_plot
from show_knmi_functions.plot_percentiles import plot_percentiles
from show_knmi_functions.show_aantal_keren import show_aantal_keren
from show_knmi_functions.show_per_periode import show_per_periode
from show_knmi_functions.spaghetti_plot import spaghetti_plot
from show_knmi_functions.show_year_histogram_animation import show_year_histogram_animation
from show_knmi_functions.last_day import last_day
from show_knmi_functions.anomaly import anomaly
from show_knmi_functions.neerslagtekort import neerslagtekort, neerslagtekort_meerdere_stations
# except:
#     from utils import show_weerstations, help,  list_to_text,check_from_until, find_date_for_title, download_button,get_weerstations, get_data
#     from does_rain_predict_rain import does_rain_predict_rain
#     from polar_plot import polar_plot, polar_debug
#     from show_warmingstripes import show_warmingstripes
#     from show_plot import show_plot
#     from plot_percentiles import plot_percentiles
#     from show_aantal_keren import show_aantal_keren
#     from show_per_periode import show_per_periode
#     from spaghetti_plot import spaghetti_plot
#     from show_year_histogram_animation import show_year_histogram_animation
#     from last_day import last_day
#     from neerslagtekort import neerslagtekort, neerslagtekort_meerdere_stations
# INSPRIATION : https://weatherspark.com/m/52666/10/Average-Weather-in-October-in-Utrecht-Netherlands
# https://radumas.info/blog/tutorial/2017/04/17/percentile-test.html

pd.set_option('future.no_silent_downcasting', True) 
pd.options.mode.chained_assignment = None  
try:
    #delete the accessor to avoid warning 
    del pd.DataFrame.loess
except AttributeError:
    pass


@st.cache_data (ttl=60 * 60 * 24)
def getdata_wrapper(stn, fromx, until):
    url = f"https://www.daggegevens.knmi.nl/klimatologie/daggegevens?stns={stn}&vars=TEMP:SQ:SP:Q:DR:RH:UN:UX:EV24:FHX&start={fromx}&end={until}"
    
    df = get_data(url)
     
    if platform.processor():
        df = df[(df["YYYYMMDD"] >= fromx) & (df["YYYYMMDD"] <= until)]
    return df, url


    
def interface():
    """Kies het weerstation, de begindatum en de einddatum

    Returns:
        df, het weerstation, begindatum en einddatum (laatste drie als string)
    """
    mode = st.sidebar.selectbox(
        "Modus (kies HELP voor hulp)", ["doorlopend per dag", "aantal keren", "specifieke dag","last day",
                                        "jaargemiddelde", "maandgemiddelde", "per dag in div jaren", "per week in div jaren", 
                                        "per maand in div jaren",  "spaghetti plot","anomaly", "percentiles", 
                                        "polar plot/radar chart", "show year histogram animation",
                                        "does rain predict rain","neerslagtekort","neerslagtekort_meerdere", "warmingstripes",
                                        "show weerstations", "help", "polar_debug"], index=19
    )
    if mode !=  "neerslagtekort_meerdere":
        weer_stations = get_weerstations()
        weerstation_namen = []
        for w in weer_stations:
            weerstation_namen.append(w[1])
        weerstation_namen.sort()

        gekozen_weerstation = st.sidebar.selectbox(
            "Weerstation", weerstation_namen, index=4
        )
        for w in weer_stations:
            if gekozen_weerstation == w[1]:
                stn = w[0]
    else:
        stn = None

    DATE_FORMAT = "%m/%d/%Y"
    start_ = "2019-01-01"
    today = datetime.today().strftime("%Y-%m-%d")
    from_ = st.sidebar.text_input("startdatum (yyyy-mm-dd) from 1-1-1900", start_)
    until_ = st.sidebar.text_input("enddatum (yyyy-mm-dd)", today)

    if mode not in ["does rain predict rain","neerslagtekort","neerslagtekort_meerdere"] :
        show_options = [
            "temp_min",
            "temp_avg",
            "temp_max",
            "gevoelstemperatuur_avg",
            "gevoelstemperatuur_max",
            "T10N",
            "zonneschijnduur",
            "perc_max_zonneschijnduur",
            "glob_straling",
            "neerslag_duur",
            "neerslag_etmaalsom",
            "neerslag_etmaalsom_div_duur",
            "RH_min","RH_max",
            "spec_humidity_knmi_derived",
            "abs_humidity_knmi_derived",
            "globale_straling_log10",
        ]


        what_to_show = st.sidebar.multiselect("Wat weer te geven", show_options, "temp_max")
       
        if (mode != "anomaly") & (mode!= "warmingstripes") & (mode!= "last day"):
            graph_type = st.sidebar.selectbox("Graph type (plotly=interactive)", ["pyplot", "plotly"], index=1)


            wdw = st.sidebar.number_input("Window smoothing curves", 1, 999, 7)
            wdw2 = st.sidebar.number_input("Window smoothing curves 2 (999 for none)", 1, 999, 999)
            if wdw2 != 999:
                sma2_how = st.sidebar.selectbox("SMA2 How", ["mean", "median"], 0)
            else:
                sma2_how = None
            centersmooth =  st.sidebar.selectbox(
                "Smooth in center", [True, False], index=0
                )
            st.sidebar.write("Smoothing niet altijd aanwezig")
            show_ci =  st.sidebar.selectbox(
                "Show CI", [True, False], index=1
                )
            if show_ci:
                wdw_ci = st.sidebar.slider("Window confidence intervals", 1, 100, 20)
            else:
                wdw_ci = 1
            show_loess =  st.sidebar.selectbox(
                "Show loess", [True, False], index=1
                )
            show_parts =  st.sidebar.selectbox(
                "Show parts", [True, False], index=1
                )
            if show_parts:
                no_of_parts = st.sidebar.slider("Number of parts", 1, 10, 5)
            else:
                no_of_parts = None
            groupby_ = st.sidebar.selectbox("Groupby", [True, False], index=1)
        else:
            # (mode == "anomaly") & (mode== "warmingstripes")
            wdw, wdw2,sma2_how, what_to_show, gekozen_weerstation, centersmooth, graph_type,show_ci,show_loess, wdw_ci,show_parts, no_of_parts, groupby_ = None,None,None,what_to_show,gekozen_weerstation,None,None,None,None,None,None, None, None

    else:
        # ["does rain predict rain","neerslagtekort","neerslagtekort_meerdere"] 
        wdw, wdw2,sma2_how, what_to_show, gekozen_weerstation, centersmooth, graph_type,show_ci,show_loess, wdw_ci,show_parts, no_of_parts, groupby_ = None,None,None,"neerslag_etmaalsom",None,None,None,None,None,None,None, None, None
    
   
    return stn, from_, until_, mode, groupby_, wdw, wdw2,sma2_how, what_to_show, gekozen_weerstation, centersmooth, graph_type,show_ci, show_loess, wdw_ci,show_parts, no_of_parts
    


    
def action(stn, from_, until_, mode,groupby_, wdw, wdw2, sma2_how, what_to_show, gekozen_weerstation, centersmooth, graph_type, show_ci, show_loess, wdw_ci,show_parts, no_of_parts):
    what_to_show_as_txt = list_to_text(what_to_show)
    FROM, UNTIL = check_from_until(from_, until_)
    if (stn !=None) and (mode != "help"):
        df_getdata, url = getdata_wrapper(stn, FROM.strftime("%Y%m%d"), UNTIL.strftime("%Y%m%d"))
        df = df_getdata.copy(deep=False)
    
    if groupby_:
        groupby_how = st.sidebar.selectbox("Groupby", ["year", "year_month"], index=1)
        groupby_what = st.sidebar.selectbox("Groupby",["sum", "mean"], index=1)
        if groupby_what == "sum":
            df = df.groupby([df[groupby_how]], sort = True).sum(numeric_only = True).reset_index()
        elif groupby_what == "mean":
            df = df.groupby([df[groupby_how]], sort = True).mean(numeric_only = True).reset_index()
        datefield = groupby_how
    else:
        datefield = "YYYYMMDD"
    # with st.expander("Dataframe"):
    #     st.write(df)
    #     download_button(df)
    
    if mode == "help":
        help()
    elif mode == "does rain predict rain":
        does_rain_predict_rain(df)
    elif mode == "last day":
       
        last_day(df, gekozen_weerstation, what_to_show)
    elif mode == "show weerstations":
        show_weerstations()
    elif mode == "neerslagtekort":
        neerslagtekort(df)
    elif mode == "neerslagtekort_meerdere":
        st.subheader("Neerslagtekort")
        #try:
        neerslagtekort_meerdere_stations(FROM, UNTIL)
        # except:
        #     st.error("Under construction")
        #     st.stop()
    elif mode == "per dag in div jaren":
        show_per_periode(df, gekozen_weerstation, what_to_show, "per_dag", graph_type)
        datefield = None
        title = f"{what_to_show_as_txt} van {from_} - {until_} in {gekozen_weerstation}"
    elif mode == "spaghetti plot":
        wdw_interval = st.sidebar.number_input("window for smoothing the 95% interval",1,len(df),9)
        last_year =  st.sidebar.selectbox(
            "Show / highlight last year", [True, False], index=0
            )
        mean_ =  st.sidebar.selectbox(
            "Show mean value", [True, False], index=0
            )
        show_quantiles =  st.sidebar.selectbox(
            "Show quantiles [2.5-97.5]", [True, False], index=1
            )
        sd_all =  st.sidebar.selectbox(
            "Show CI calculated with a stdev overall", [True, False], index=1
            )
        sd_day =  st.sidebar.selectbox(
            "Show CI calculated with a stdev per day", [True, False], index=1
            )
        
        spaghetti =  st.sidebar.selectbox(
            "Show spaghetti", [True, False], index=0
            )
        gradient =  st.sidebar.selectbox(
            "Show Gradient", ["None", "Pubu", "Purd", "Greys", "Plasma"], index=0
            )
        cumulative = st.sidebar.selectbox(
            "Show cumulative values", [True, False], index=0
            )
        spaghetti_plot(df, what_to_show, wdw, wdw_interval,sd_all, sd_day, spaghetti, mean_, last_year, show_quantiles, gradient, cumulative, True)
    elif mode == "per maand in div jaren":
        show_per_periode(df, gekozen_weerstation, what_to_show, "maandgem", graph_type)
        datefield = None
        title = f"{what_to_show_as_txt} van {from_} - {until_} in {gekozen_weerstation}"
    elif mode == "per week in div jaren":
        show_per_periode(df, gekozen_weerstation, what_to_show, "weekgem", graph_type)
        datefield = None
        title = f"{what_to_show_as_txt} van {from_} - {until_} in {gekozen_weerstation}"
    elif mode == "aantal keren":
        show_aantal_keren(df, gekozen_weerstation, what_to_show)
    elif mode == "anomaly":
        anomaly(df, what_to_show)
    elif mode == "percentiles":
        plot_percentiles(df,  gekozen_weerstation, what_to_show, wdw, centersmooth)
    elif mode == "warmingstripes":
        st.write(gekozen_weerstation)
        title = f"{what_to_show_as_txt} van {from_} - {until_} in {gekozen_weerstation}"
        mode  = st.sidebar.selectbox(
            "Mode", ["Classic", "New", "Matplotlib"], index=0
            )
        show_warmingstripes(df, what_to_show, title,mode )
        
    elif mode == "polar plot/radar chart":
        how = st.sidebar.selectbox(
            "Scatter / line", ["scatter", "line"], index=0
            )
        polar_plot(df,  what_to_show, how)
    elif mode == "polar_debug":
       
        polar_debug(df)
    elif mode =="show year histogram animation":
        show_year_histogram_animation(df, what_to_show)
        
    else:
        # all commands who use show_plot() and show_warmingstripes()
        if mode == "doorlopend per dag":
            title = f"{what_to_show_as_txt} van {from_} - {until_} in {gekozen_weerstation}"
        else:
            datefield = "YYYY"
            if mode == "jaargemiddelde":
                df = df.groupby(["YYYY"], sort=True).mean(numeric_only = True).reset_index()
                #st.write(df)
                title = f"Jaargemiddelde {what_to_show_as_txt}  van {from_[:4]} - {until_[:4]} in {gekozen_weerstation}"
                st.sidebar.write(
                    "Zorg ervoor dat de einddatum op 31 december valt voor het beste resultaat "
                )

            elif mode == "maandgemiddelde":
                
                month_min = st.sidebar.number_input("Maand",1,12,6)
                month_max = st.sidebar.number_input("Maand",1,12,8)
                if month_min > month_max:
                    st.error("Max must be higher or equal than min")
                    st.stop()

                df = df[(df["MM"] >= month_min) & (df["MM"] <= month_max)].reset_index().copy(deep=True)
                df = df.groupby(by=["year"]).mean(numeric_only=True).reset_index() # werkt maar geeft geen 0 waardes weer 
                title = f"Maandgemiddelde (maand = {month_min}-{month_max}) - {what_to_show} in {gekozen_weerstation}"
               
            elif mode == "specifieke dag":
                day = st.sidebar.number_input("Dag", 1, 31, 1, None, format="%i")
                months = [
                    "januari",
                    "februari",
                    "maart",
                    "april",
                    "mei",
                    "juni",
                    "juli",
                    "augustus",
                    "september",
                    "oktober",
                    "november",
                    "december",
                ]
                month = months.index(st.sidebar.selectbox("Maand", months, index=0)) + 1
                df = df[
                    (df["YYYYMMDD"].dt.month == month) & (df["YYYYMMDD"].dt.day == day)
                ]
                
                title = f"{what_to_show_as_txt}  op {find_date_for_title(day,month)} van {from_[:4]} - {until_[:4]} in {gekozen_weerstation}"
                st.sidebar.write(
                    "Zorg ervoor dat de datum in de gekozen tijdrange valt voor het beste resultaat "
                )
        
        show_plot(df, datefield, title, wdw, wdw2, sma2_how, what_to_show, graph_type, centersmooth, show_ci, show_loess, wdw_ci, show_parts, no_of_parts)
        # show_warmingstripes(df, what_to_show, title, False)
    try:
        st.sidebar.write(f"URL to get data: {url}")
    except:
        # in case of neerslag tekort meerdere
        pass

def main():
    
    st.write("Weather NL.")
    stn, from_, until_, mode,groupby_, wdw, wdw2, sma2_how, what_to_show, gekozen_weerstation, centersmooth, graph_type, show_ci, show_loess, wdw_ci,show_parts, no_of_parts = interface()
    action(stn, from_, until_, mode, groupby_,  wdw, wdw2,sma2_how,  what_to_show, gekozen_weerstation, centersmooth, graph_type, show_ci, show_loess,wdw_ci,show_parts, no_of_parts)

if __name__ == "__main__":
    main()