add universal functions, to import to notebook, ex. ensure_rain/snow

src/my_package/util.py

@@ -1,3 +1,7 @@
+import numpy as np
+import pandas as pd
+
+# Function to replace the norcid 'æøå'
 def replace_nordic(city_name):
     for letter in city_name:
         if letter in 'æøå':
@@ -6,7 +10,73 @@ def replace_nordic(city_name):
             city_name = city_name.replace('å', 'aa')
     return city_name
 
-
+# Function to convert from kelvin to celsius temp
 def kelvin_to_celsius(temp_in_kelvin):
     temp_in_celsius = temp_in_kelvin - 273.15
-    return temp_in_celsius
+    return temp_in_celsius
+
+# Fucntion to check if there are a ['rain.1h'] column in the dataset, if not make one
+def ensure_rain_column(df):
+    try:
+        _ = df['rain.1h']
+
+    # If no rain, make the rain column and fill it with NaN
+    except KeyError:
+        print("'Rain' is not present in the JSON file.")
+        df['rain.1h'] = np.nan
+
+    return df
+
+# Fucntion to check if there are a ['snow.1h'] column in the dataset, if not make one
+def ensure_snow_column(df):
+    try:
+        _ = df['snow.1h']
+
+    # If no snow, make the snow column and fill it with NaN
+    except KeyError:
+        print("'Snow' is not present in the JSON file.")
+        df['snow.1h'] = np.nan
+
+    return df
+
+# Function to fill NaN values in ['rain.1h'] columns with 0
+def fill_rain_column(df):
+    try: 
+        # Replace the NaN with 0, using pandas function
+        df['rain.1h'] = df['rain.1h'].fillna(0)
+
+    except KeyError:
+        print(["'rain.1h', not in df"])
+
+    return df
+
+# Function to fill NaN values in ['snow.1h'] columns with 0
+def fill_snow_column(df):
+    try: 
+        # Replace the NaN with 0, using pandas function
+        df['snow.1h'] = df['snow.1h'].fillna(0)
+
+    except KeyError:
+        print(["'snow.1h', not in df"])
+
+    return df
+
+# Function to 'normalize' the dataset, with index-changing and dropping meta-data 
+def extract_city_df(weather_data):
+    if 'list' in weather_data:
+        # Normalize the json for better readability
+        df = pd.json_normalize(weather_data['list'])
+
+        # Delete duplicates based on the dt row, all the other values can appear more than once, but the date should only appear once
+        df = df.drop_duplicates(subset=['dt'])
+
+        # The weather column dosnt have any releated information, therefor we delete it
+        df = df.drop(columns="weather")
+
+        # Convert 'dt' column from Unix timestamp to datetime and set it as the index
+        df['dt'] = pd.to_datetime(df['dt'], unit='s')
+        df.set_index('dt', inplace=True)
+        return df
+
+    else:
+        return None