2 com

wehs_helpers.py

@@ -0,0 +1,215 @@
+import pandas as pd
+import numpy as np
+
+DATA_FOLDER = '../Daten_pcm_export/'
+
+def load():
+    file_list = ['klienten_pcm.csv', 'faelle_pcm.csv' , 'massnahmen_pcm.csv','produkte_pcm.csv', #  'faelle2_pcm_20210919.csv'
+             'platzierungen_pcm.csv','aufsuchende_familienarbiet_pcm.csv','tagesstruktur_pcm.csv',
+                'mutter_export_20210819.csv','vater_export_20210819.csv']
+    df = {file[:3]:pd.read_csv(DATA_FOLDER+file,delimiter=';',skiprows=[1]) for file in file_list}
+
+    preprocess_kli(df['kli'])
+    preprocess_fae(df['fae'])
+    preprocess_mas(df['mas'])
+    preprocess_eltern(df['mut'],'MUTTER')
+    preprocess_eltern(df['vat'],'VATER')
+
+
+    df['kli']['Nmas'] = key_count(df['kli'],df['mas'],'CDW_KLIENT_ID')
+    df['kli']['Npla'] = key_count(df['kli'],df['pla'],'CDW_KLIENT_ID')
+    df['kli']['Nauf'] = key_count(df['kli'],df['auf'],'CDW_KLIENT_ID')
+    df['kli']['Ntag'] = key_count(df['kli'],df['tag'],'CDW_KLIENT_ID')
+    df['kli']['Nfae'] = key_count(df['kli'],df['fae'],'CDW_KLIENT_ID')
+    df['kli']['Npro'] = key_count(df['kli'],df['pro'],'CDW_KLIENT_ID')
+    df['kli']['Nbetr'] = df['kli']['Npla'] + df['kli']['Nauf'] + df['kli']['Ntag']
+    df['kli']['Nmut'] = key_count(df['kli'],df['mut'],'CDW_SOURCE_KEY')
+    df['kli']['Nvat'] = key_count(df['kli'],df['vat'],'CDW_SOURCE_KEY')
+    df['kli']['Neltern'] = df['kli']['Nvat'] + df['kli']['Nmut']
+
+    return df
+
+# ------
+def preprocess_fae(df):
+    df['FALL_AUFNAHME_DATUM'] = df['FALL_AUFNAHME_DATUM'].apply(parse_date)
+    df['FALL_AUFNAHME_Jahr'] = df['FALL_AUFNAHME_DATUM'].apply(get_year)
+
+# ------
+def preprocess_kli(df):
+    df['KLIENT_GEBURTSDATUM'] = df['KLIENT_GEBURTSDATUM'].apply(parse_date)
+    df['KLIENT_Geburtsjahr'] = df['KLIENT_GEBURTSDATUM'].apply(get_year)
+    df['PLZ_KLIENT_WOHNORT'] = df['PLZ_KLIENT_WOHNORT'].apply(parse_plz)
+
+# ------
+def preprocess_eltern(df, tag):
+    df['CDW_SOURCE_KEY']= df['ID_K_ELTERN_FK'].apply(lambda s: s.strip(' ')) 
+    df['GEBURTSDATUM_YYMD_'+tag] = df['GEBURTSDATUM_YYMD_'+tag].apply(parse_date).astype(np.datetime64)
+    df['Geburtsjahr_'+tag] = df['GEBURTSDATUM_YYMD_'+tag].apply(get_year)
+    df['PLZ_GESETZ_WOHNSITZ_'+tag] = df['PLZ_GESETZ_WOHNSITZ_'+tag].apply(parse_plz)
+
+# ------
+def preprocess_mas(df,prefix = 'zgb'):
+    ZGBdata = [['308 Abs. 1 + 2 (+3)',
+          '308',
+          '308a321',
+          'Beistandschaft Rat und Tat, besondere Befugnisse, elterliche Sorge entsprechend beschränkt'],
+         ['308 Abs. 1 + 2',
+          '308',
+          '308a21',
+          'Erziehungsbeistandschaft und besondere Befugnisse'],
+         ['308 Abs. 2', '308', '308a2', 'Beistandschaft, besondere Befugnisse'],
+         ['308 Abs. 1', '308', '308a1', 'Erziehungsbeistandschaft'],
+         ['306 Abs. 2', '306', '306', 'Interessenkollision/Verhinderung der Eltern'],
+         ['* 309 / 308', '308', '308u09', 'Beistandschaft Vaterschaft/Unterhalt'],
+         ['310 / 308',
+          '310',
+          '310u08',
+          'Aufhebung des Aufenthaltsbestimmungsrechts/Beistandschaft'],
+         ['307', '307', '307', 'Kindesschutz (Weisung, Erziehungsaufsicht)'],
+         ['327a (311/312/298 Abs. 2)',
+          '327',
+          '327',
+          'Minderjährige unter Vormundschaft'],
+         ['308 Abs. 2',
+          '308',
+          '308a2v',
+          'Beistandschaft Regelung Vaterschaft/Unterhalt'],
+         ['310 / 308 / 314b',
+          '310',
+          '310u08u14',
+          'Aufhebung des Aufenthaltsbestimmungsrechts/Beistandschaft/FU'],
+         ['299 / 300 ZPO',
+          'ZPO',
+          'ZPO',
+          'Verfahrensvertretung in familienrechtlichen Verfahren'],
+         ['Art. 18 BG-HAÜ', 'BG', 'BG18', 'Adoptionsvormundschaft'],
+         ['308 Abs. 2', '308', '308a2u', 'Beistandschaft Regelung Unterhalt'],
+         ['324 / 325',
+          '324',
+          '324',
+          'Schutz des Kindesvermögens (Aufsicht/Beistandschaft)'],
+         ['314abis Abs.2 Ziff.2', '314', '314', 'Verfahrensbeistandschaft vor KESB'],
+         ['Art. 17 BG-HAÜ', 'BG', 'BG17', 'Adoptionsbeistandschaft'],
+         ['308 Abs. 2 + 3',
+          '308',
+          '308a32',
+          'Beistandschaft, besondere Befugnisse, elterliche Sorge entsprechend beschränkt']]
+
+    ZGBdf = (pd.DataFrame(ZGBdata, columns=['ZGBARTIKEL','ZGBkey1','ZGBkey2','ZGBTEXTformatiert'])
+             .set_index('ZGBTEXTformatiert')
+             .sort_values(by='ZGBkey2') )
+
+    df['ZGBTEXTformatiert'] = df.ZGBTEXT.apply(lambda s: s.strip(' ').replace(' / ','/'))
+    df['ZGBkey1'] = ZGBdf['ZGBkey1'].reindex(df['ZGBTEXTformatiert']).values
+    df['ZGBkey2'] = ZGBdf['ZGBkey2'].reindex(df['ZGBTEXTformatiert']).values
+
+
+# ============== helpers ================================================================   
+import datetime
+def parse_date(x, cutyear = 2022):
+    if np.isreal(x):
+        if np.isnan(x):
+            x = ''
+        else:
+            x = str(int(x))
+
+    if (len(x) == 8):
+        date = np.datetime64('-'.join([x[:4],x[4:6],x[-2:]]))
+        if date.astype(object).year <= cutyear:
+            return date
+        else:
+            return np.datetime64('nat')
+    else:
+        return np.datetime64('nat')
+
+def get_year(d):
+    try:
+        if type(d) == np.datetime64:
+            return d.astype(object).year  
+        else:
+            return int(d.year)
+    except:
+        return -1  
+
+def parse_plz(x):
+    # PLZ are strings - especially for foreign countries: DL-12345
+    if type(x) == str:
+        return x
+    else:
+        if np.isnan(x):
+            return ''
+        else:
+            return str(x)
+
+
+
+# ============== aggregation functions ==================================================
+
+def one_count(df,column, count_col_name='Counts', 
+              percentage = True, percentage_col_name = 'Percentage',minpct=0):
+    '''
+    one_count(df,column, count_col_name='Counts', 
+              percentage = True, percentage_col_name = 'Percentage')
+    '''
+
+    out = (df.assign(**{count_col_name:1})
+            .groupby([column])[[count_col_name]].count()
+            .fillna(0).astype(int)
+            .sort_values(by=count_col_name,ascending = False) )
+    if percentage:
+        out[percentage_col_name] = (100*out[count_col_name] / out[count_col_name].sum() ).round(1)
+        out = out[out[percentage_col_name] >= minpct]
+
+    return out
+
+def two_count(df,index_col, column_col, dense=False):
+    out = (df.assign(counts=1)
+            .groupby([index_col,column_col])[['counts']].count()
+            .reset_index()
+            .pivot(index=index_col, columns=column_col,values='counts')
+            .fillna(0).astype(int) )
+    if dense:
+        s = out.stack()
+        out = s[s>0].sort_values(ascending=False).reset_index().rename(columns={0:'Counts'})
+    return out
+
+def key_count(target_df,count_df,key):
+    return (count_df.assign(counts=1)
+            .groupby(key)[['counts']].count()
+            .reindex(target_df[key]).fillna(0).astype(int)['counts'].values) 
+
+def group_col_to_list(df,groupby,values,sortby=None,joinfun=None):
+    '''
+    group_col_to_list(df,groupby,values,sortby=None,joinfun=None)
+    
+    groups a column into lists and returns a series objct
+    
+    Arguments
+    ---------
+    
+    df : DataFrame
+    
+    groupby : str
+        column name of df for groupby
+        
+    values : str
+        column name of df for values
+        
+    sortby : str, default = None
+        column name of df to sort the values
+        
+    joinfun : function handle, default = None
+        function which takes the list as an input and outputs something else (e.g. concatenated string)
+    '''
+
+    def applyfun(df):
+        if sortby is not None:
+            df = df.sort_values(sortby)
+        vallist = df[values].values.tolist()
+        if joinfun is None:
+            return vallist
+        else:
+            return joinfun(vallist)
+
+    return df.groupby(groupby).apply(applyfun)
+