feat: updated algo for offers

app.py

@@ -14,10 +14,10 @@ PASS = os.getenv("PASSWORD")
 # create an interface and limit output's width for the dataframe bu
 list_iface = gr.Interface(fn=compute_offer,
             inputs=[gr.File(label="Upload CSV", type="file"),
-                    gr.Slider(1, 365, value=30, step=1, label="Days", info="Number of days to look back"),
-                    gr.Slider(5000, 100000, value=15000, step=1, label="Minimum Sent", info="Minimum number of emails sent"),
-                    gr.Dropdown(["Comcast", "Yahoo", "Hotmail", "Aol"], value="Comcast", label="Domain"),
-                    gr.Radio(["Newsletters", "Offers"], label="Type", value="Newsletters"),
+                    gr.Slider(1, 365, value=3, step=1, label="Days", info="Number of days to look back"),
+                    gr.Slider(5000, 10000000, value=1000000, step=1, label="Minimum Sent", info="Minimum number of emails sent"),
+                    gr.Dropdown(["Comcast", "Yahoo", "Hotmail", "Aol"], value="Yahoo", label="Domain"),
+                    gr.Radio(["Newsletters", "Offers"], label="Type", value="Offers"),
                     gr.Textbox(label="Exclude list", info="Example: INH,MNP", value="INH,DHI,HHP,RTA,JVR,HTH,FNC,SCD,ENH,WIP")],
             outputs="dataframe")
 

utils/gradio_utils.py

@@ -256,7 +256,7 @@ def compute_offer(csv_file, days_lookback, min_sent, domain, offer_type, x_list)
 
     cols = ['Campanie', 'Oferta', 'Nume', 'Server', 'User',
         'Lista Custom', 'Data', 'HClicks', 'Clicks', 'Unscribers', 'Openers',
-        'Click Open', 'Leads', 'CLike', 'Complains', 'Traps', 'Send']
+        'Click Open', 'Leads', 'CLike', 'Complains', 'Traps', 'Send', 'ECPM']
     # df_all = raw_df[raw_df['Nume'].str.contains('|'.join(cmp_list))]  #1
     # df_all = df_all[df_all['Domeniu'] == 'Comcast'] #2
     exclude_list = df_all[(df_all['Data'] > (pd.Timestamp('now') - pd.Timedelta(days=days_lookback))) \
@@ -265,9 +265,12 @@ def compute_offer(csv_file, days_lookback, min_sent, domain, offer_type, x_list)
     
     df_all = df_all[df_all['Send'] > int(min_sent)]
     df_all = df_all[cols]
-    df_all['Click Open'] = df_all['Click Open'].str.replace('%', '').astype(float)
     # fixed a blank line in the csv
     df_all = df_all[df_all["Oferta"] != " "]
+    
+    df_all['Click Open'] = df_all['Click Open'].str.replace('%', '').astype(float)
+    df_all['ECPM'] = df_all['ECPM'].astype(float)
+
 
     # Limit the characters in the "Nume" column
     # df_all["Nume"] = df_all["Nume"].apply(_limit_chars)
@@ -282,6 +285,11 @@ def compute_offer(csv_file, days_lookback, min_sent, domain, offer_type, x_list)
                         ]
     elif offer_type == "Offers":
         df_all = df_all[~df_all['Nume'].str.startswith("Aeon News")]
+        df_all = df_all[~df_all['Nume'].str.contains("NU SE TRIMITE")]
+        df_all = df_all[~df_all['Nume'].str.contains("de testat")]
+        df_all = df_all[~df_all['Nume'].str.contains("_TEST")]
+        df_all = df_all[~df_all['Nume'].str.contains("CPM")]
+        df_all = df_all[~df_all['Nume'].str.contains("RESTRICTED")]
 
     if x_list != "":
         x_list = x_list.split(',')
@@ -294,11 +302,18 @@ def compute_offer(csv_file, days_lookback, min_sent, domain, offer_type, x_list)
 
     df_all.reset_index(drop=True, inplace=True)
 
-    final_df = df_all.groupby(["Oferta", "Nume"])\
-        .agg( N=('Oferta', 'count'), send_avg=('Send', 'mean'), CO=('Click Open', 'mean'))\
-        .sort_values(['CO', 'N'], ascending=False)
-    final_df['send_avg'] = final_df['send_avg'].astype(int)
-    final_df['CO'] = final_df['CO'].round(2).astype(float)
-    final_df.reset_index(inplace=True)
-
+    if offer_type == "Newsletters":
+        final_df = df_all.groupby(["Oferta", "Nume"])\
+            .agg( N=('Oferta', 'count'), send_avg=('Send', 'mean'), CO=('Click Open', 'mean'))\
+            .sort_values(['CO', 'N'], ascending=False)
+        final_df['send_avg'] = final_df['send_avg'].astype(int)
+        final_df['CO'] = final_df['CO'].round(2).astype(float)
+        final_df.reset_index(inplace=True)
+    elif offer_type == "Offers":
+        final_df = df_all.groupby(["Oferta", "Nume"])\
+            .agg( N=('Oferta', 'count'), send_avg=('Send', 'mean'), ECPM=('ECPM', 'mean'))\
+            .sort_values(['ECPM', 'N'], ascending=False)
+        final_df['send_avg'] = final_df['send_avg'].astype(int)
+        final_df['ECPM'] = final_df['ECPM'].round(2).astype(float)
+        final_df.reset_index(inplace=True)
     return final_df