Make dumpstuff more efficient

example.py

@@ -0,0 +1,31 @@
+"""Exemple d'utilisation des dumps."""
+
+import json
+from collections import Counter
+
+
+def parse(filename):
+    with open(filename, "r") as file:
+        return json.load(file)
+
+
+if __name__ == "__main__":
+    Articles = {a["id"]: a for a in parse("article.dump.json")}
+    Operations = parse("operation.dump.json")
+
+    # Dépenses faites en K-Fêt depuis K-Psul
+    total_expenses = -sum((float(op["amount"]) for op in Operations))
+    print(f"total expenses: {total_expenses:.2f}€")
+
+    # Montant moyen d'une transaction
+    nb_ops = len(Operations)
+    print(f"Average transaction amount: {total_expenses / nb_ops:.2f}€")
+
+    # Nombre de ventes par article
+    articles_sales = Counter()
+    for operation in Operations:
+        articles_sales[operation["article"]] += operation["number"]
+    print("Most sold articles:")
+    for id, nb in articles_sales.most_common(3):
+        article_name = Articles[id]["name"]
+        print(f"  - {article_name} : {nb}")

kfet/management/commands/dumpstuff.py

@@ -0,0 +1,76 @@
+import json
+import random
+from typing import Dict
+
+from django.core.management.base import BaseCommand
+
+from kfet.models import Account, Article, Operation
+
+
+def gen_anonymisation_table() -> Dict[int, str]:
+    random.seed()
+    hashes = {}
+    for account_id in Account.objects.values_list("id", flat=True):
+        h = random.getrandbits(128)
+        hashes[account_id] = "{:032x}".format(h)
+    return hashes
+
+
+def dump_articles(filename: str) -> None:
+    articles = [
+        {
+            "id": article.id,
+            "name": article.name,
+            "price": str(article.price),
+            "category": article.category.name,
+            "box_type": article.box_type,
+            "box_capacity": article.box_capacity,
+        }
+        for article in Article.objects.all()
+    ]
+    with open(filename, "w") as file:
+        json.dump(articles, file, indent=4)
+
+
+def dump_operations(filename: str, accounts_hashes: Dict[int, str]) -> None:
+    not_canceled_purchases = (
+        Operation.objects.filter(type=Operation.PURCHASE)
+        .filter(canceled_at__isnull=True)
+        .values_list(
+            "amount",
+            "article__id",
+            "article_nb",
+            "group__at",
+            "group__is_cof",
+            "group__on_acc__id",
+        )
+    )
+    operations = [
+        {
+            "amount": str(amount),
+            "article": article_id,
+            "number": nb,
+            "date": str(at),
+            "is_cof": is_cof,
+            "on_account": accounts_hashes[on_acc],
+        }
+        for amount, article_id, nb, at, is_cof, on_acc in not_canceled_purchases
+    ]
+    with open(filename, "w") as file:
+        json.dump(operations, file, indent=4)
+
+
+class Command(BaseCommand):
+    help = 'Dump un historique "anonymisé".'
+
+    def handle(self, *args, **options):
+        # XXX. This is not great for privacy.
+        accounts_hashes = gen_anonymisation_table()
+
+        article_file = "article.dump.json"
+        self.stdout.write('Dumping articles to "{}"'.format(article_file))
+        dump_articles(article_file)
+
+        operation_file = "operation.dump.json"
+        self.stdout.write('Dumping operations to "{}"'.format(operation_file))
+        dump_operations(operation_file, accounts_hashes)