From 8baaa1d339b37d767b79b2074b92909eb617ca1f Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Martin=20P=C3=A9pin?= <kerl@wkerl.me>
Date: Thu, 6 Feb 2020 19:46:08 +0100
Subject: [PATCH] Add randomly generated account ids in kfet dump

---
 kfet/management/commands/dumpstuff.py | 62 +++++++++++++++------------
 1 file changed, 35 insertions(+), 27 deletions(-)

diff --git a/kfet/management/commands/dumpstuff.py b/kfet/management/commands/dumpstuff.py
index fe8a3eea..d13810e5 100644
--- a/kfet/management/commands/dumpstuff.py
+++ b/kfet/management/commands/dumpstuff.py
@@ -1,59 +1,67 @@
 import json
+import random
+from typing import Dict
 
 from django.core.management.base import BaseCommand
 
-from kfet.models import Article, Operation, OperationGroup
+from kfet.models import Account, Article, Operation
 
 
-def dump_articles(filename: str):
-    articles = Article.objects.values_list(
-        "id", "name", "price", "category__name", "box_type", "box_capacity"
-    )
+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": id,
-            "name": name,
-            "price": str(price),
-            "category": cat,
-            "box_type": box_type,
-            "box_capacity": box_capacity,
+            "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 id, name, price, cat, box_type, box_capacity in articles
+        for article in Article.objects.all()
     ]
     with open(filename, "w") as file:
         json.dump(articles, file, indent=4)
 
 
-def dump_operations(filename: str):
-    operations = (
-        Operation.objects.filter(type=Operation.PURCHASE)
-        .filter(canceled_at__isnull=True)
-        .values_list(
-            "amount", "article__id", "article_nb", "group__at", "group__is_cof"
-        )
+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
     )
     operations = [
         {
-            "amount": str(amount),
-            "article": id,
-            "number": nb,
-            "date": str(at),
-            "is_cof": is_cof,
+            "amount": str(operation.amount),
+            "article": operation.article.id,
+            "number": operation.article_nb,
+            "date": str(operation.group.at),
+            "is_cof": operation.group.is_cof,
+            "on_account": accounts_hashes[operation.group.on_acc.id],
         }
-        for amount, id, nb, at, is_cof in operations
+        for operation in not_canceled_purchases
     ]
     with open(filename, "w") as file:
         json.dump(operations, file, indent=4)
 
 
 class Command(BaseCommand):
-    help = "Dump un historique anonymisé."
+    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)
+        dump_operations(operation_file, accounts_hashes)