Add application file

app.py

@@ -1,5 +1,6 @@
 import os
 import uuid
+import numpy as np
 import pandas as pd
 import streamlit as st
 import huggingface_hub as hh
@@ -24,7 +25,6 @@ GROUPS_INFO_PATH = os.path.join(CACHE_PATH, "groups-info")
 
 REQUIRED_COLUMNS = ['accommodation_id', 'user_id'] + [f'review_{i}' for i in range(1, 11)]
 
-
 API = hh.HfApi(token=TOKEN)
 
 
@@ -54,12 +54,26 @@ def refresh_data():
 refresh_data()
 
 
+def get_match_index(row):
+    for i in range(1, 11):
+        if row['review_id'] == row[f'review_{i}']:
+            return i
+    return np.inf
+
+
 def calculate_metrics(df_pred):
     df_gt = pd.read_csv(os.path.join(GT_PATH, 'val_matches.csv'))
+    if len(df_pred) != len(df_gt):
+        raise Exception("Your predictions file should contain {} rows, only {} rows were found in the file".format(
+            len(df_gt), len(df_pred)
+        ))
+
+    df_merged = pd.merge(df_gt, df_pred, how='left', on=['accommodation_id', 'user_id']).fillna('')
+    df_merged['match_index'] = df_merged.apply(get_match_index, axis=1)
+    df_merged['mrr10'] = df_merged['match_index'].apply(lambda x: 1/x)
+    df_merged['precision10'] = df_merged.apply(lambda x: 1 if x != np.inf else 0)
 
-    mrr10 = 0.3
-    precision10 = 0.2
-    return mrr10, precision10
+    return df_merged['mrr10'].mean(), df_merged['precision10'].mean()
 
 
 def get_group_name_by_email(email):
@@ -148,7 +162,7 @@ def render():
                 submission_tab.markdown("**Group name:** " + group_name)
                 submission_tab.markdown("**Model name:** " + model_name)
                 submission_tab.markdown("**Revision:** " + str(revision) +
-                                        f"(out of {MAX_SUBMISSIONS} allowed submissions)")
+                                        f" (out of {MAX_SUBMISSIONS} allowed submissions)")
 
                 submission_tab.write("### Submission results")
                 submission_tab.markdown("**MRR@10:** {:.4f}".format(mrr10))