From 96e82c57997891a4eabcee8ead5acebb5811a428 Mon Sep 17 00:00:00 2001
From: Theo Trosman <49092919@est.ort.edu.ar>
Date: Thu, 9 Apr 2026 11:04:58 -0300
Subject: [PATCH] Refactor CSV processing and SQL generation

---
 loader/load.py | 123 ++++++++++++++++++++++++++++---------------------
 1 file changed, 71 insertions(+), 52 deletions(-)

diff --git a/loader/load.py b/loader/load.py
index 2650601..4763896 100644
--- a/loader/load.py
+++ b/loader/load.py
@@ -1,60 +1,79 @@
 import csv
-import sys, os
+import sys
+from pathlib import Path
 
-# Reading from a CSV file
-with open("mctl.csv", "r") as file:
-    with open("users.sql", "w") as user_query:
-        user_query.write(
-            """
+INPUT_FILE = "mctl.csv"
+USER_OUTPUT = "users.sql"
+GAME_OUTPUT = "games.sql"
+
+
+def parse_row(row):
+    user_id = row[1]
+    last_game_date = row[2]
+
+    attempts = 11 if row[3] == "0" else int(row[4])
+
+    return {
+        "user_id": user_id,
+        "last_game_date": last_game_date,
+        "attempts": attempts,
+    }
+
+
+def read_csv(limit=None):
+    data = []
+
+    with open(INPUT_FILE, newline="") as file:
+        reader = csv.reader(file)
+        next(reader)  # skip header
+
+        for i, row in enumerate(reader, start=1):
+            data.append(parse_row(row))
+
+            if limit and i >= limit:
+                break
+
+    return data
+
+
+def generate_users_sql(data):
+    values = ",\n".join(
+        f"('{d['user_id']}', '{d['last_game_date']}')" for d in data
+    )
+
+    return f"""
 INSERT INTO public.user (user_id, last_game_date)
 VALUES
+{values}
+ON CONFLICT (user_id)
+DO UPDATE SET
+    last_game_date = excluded.last_game_date;
 """
-        )
-        with open("games.sql", "w") as game_query:
-            game_query.write(
-                f"""
-    INSERT INTO public.game_count (user_id, attempts, game_count)
-    VALUES
-    """
-            )
-            reader = csv.reader(file)
-            for i, row in enumerate(reader):
-                if i == 0:
-                    continue
-
-                user_query.write(
-                    f"""
- ('{row[1]}', '{row[2]}'),"""
-                )
-                if row[3] == "0":
-                    attempts = "11"
-                else:
-                    attempts = row[4]
-
-                game_query.write(
-                    f"""
-    ('{row[1]}', {attempts}, 1),"""
-                )
-
-                if len(sys.argv) > 1 and i > int(sys.argv[1]):
-                    break
-            game_query.seek(game_query.tell() - 1, os.SEEK_SET)
-            game_query.write("")
-            game_query.write(
-                """
+
+
+def generate_games_sql(data):
+    values = ",\n".join(
+        f"('{d['user_id']}', {d['attempts']}, 1)" for d in data
+    )
+
+    return f"""
+INSERT INTO public.game_count (user_id, attempts, game_count)
+VALUES
+{values}
 ON CONFLICT (user_id, attempts)
-    DO UPDATE SET
+DO UPDATE SET
     game_count = public.game_count.game_count + 1;
-            """
-            )
+"""
 
-        user_query.seek(user_query.tell() - 1, os.SEEK_SET)
-        user_query.write("")
-        user_query.write(
-            """
-ON CONFLICT (user_id)
-    DO UPDATE SET
-    last_game_date = excluded.last_game_date
-;
-        """
-        )
+
+def main():
+    limit = int(sys.argv[1]) if len(sys.argv) > 1 else None
+
+    data = read_csv(limit)
+
+    Path(USER_OUTPUT).write_text(generate_users_sql(data))
+    Path(GAME_OUTPUT).write_text(generate_games_sql(data))
+
+
+if __name__ == "__main__":
+    main()