From 3f4a41ba4207663df3578cfd085c3ae87c5c75d5 Mon Sep 17 00:00:00 2001
From: imperosol <thgirod@hotmail.com>
Date: Wed, 19 Nov 2025 13:51:38 +0100
Subject: [PATCH] refactor detection of the need to merge `PageRev`

---
 core/models.py      | 32 +++++++++++++++++++++++++++++++-
 core/views/forms.py | 23 +++++------------------
 2 files changed, 36 insertions(+), 19 deletions(-)

diff --git a/core/models.py b/core/models.py
index b84139ae..be3b5cc4 100644
--- a/core/models.py
+++ b/core/models.py
@@ -23,12 +23,13 @@
 #
 from __future__ import annotations
 
+import difflib
 import string
 import unicodedata
 from datetime import timedelta
 from io import BytesIO
 from pathlib import Path
-from typing import TYPE_CHECKING, Self
+from typing import TYPE_CHECKING, Final, Self
 from uuid import uuid4
 
 from django.conf import settings
@@ -1344,6 +1345,9 @@ class PageRev(models.Model):
     The content is in PageRev.title and PageRev.content .
     """
 
+    MERGE_TIME_THRESHOLD: Final[timedelta] = timedelta(minutes=20)
+    MERGE_DIFF_THRESHOLD: Final[float] = 0.2
+
     revision = models.IntegerField(_("revision"))
     title = models.CharField(_("page title"), max_length=255, blank=True)
     content = models.TextField(_("page content"), blank=True)
@@ -1385,6 +1389,32 @@ class PageRev(models.Model):
     def is_owned_by(self, user: User) -> bool:
         return any(g.id == self.page.owner_group_id for g in user.cached_groups)
 
+    def similarity_ratio(self, text: str) -> float:
+        """Similarity ratio between this revision's content and the given text.
+
+        The result is a float in [0; 1], 0 meaning the contents are entirely different,
+        and 1 they are strictly the same.
+        """
+        # cf. https://docs.python.org/3/library/difflib.html#difflib.SequenceMatcher.ratio
+        return difflib.SequenceMatcher(None, self.content, text).quick_ratio()
+
+    def should_merge(self, other: Self) -> bool:
+        """Return True if `other` should be merged into `self`, else False.
+
+        It's considered the other revision should be merged into this one if :
+
+        - it was made less than 20 minutes after
+        - by the same author
+        - with a similarity ratio higher than 80%
+        """
+        return (
+            not self._state.adding  # cannot merge if the original rev doesn't exist
+            and self.author == other.author
+            and (other.date - self.date) < self.MERGE_TIME_THRESHOLD
+            and (not other._state.adding or other.revision == self.revision + 1)
+            and self.similarity_ratio(other.content) >= (1 - other.MERGE_DIFF_THRESHOLD)
+        )
+
 
 def get_notification_types():
     return settings.SITH_NOTIFICATIONS
diff --git a/core/views/forms.py b/core/views/forms.py
index c39d36b2..58b5c598 100644
--- a/core/views/forms.py
+++ b/core/views/forms.py
@@ -20,9 +20,9 @@
 # Place - Suite 330, Boston, MA 02111-1307, USA.
 #
 #
-import difflib
 import re
-from datetime import date, datetime, timedelta
+from copy import copy
+from datetime import date, datetime
 from io import BytesIO
 
 from captcha.fields import CaptchaField
@@ -390,14 +390,11 @@ class PageRevisionForm(forms.ModelForm):
 
         - less than 20 minutes ago
         - by the same author
-        - with a diff ratio higher than 20%
+        - with a similarity ratio higher than 80%
 
         then the latter will be edited and the new revision won't be created.
     """
 
-    TIME_THRESHOLD = timedelta(minutes=20)
-    DIFF_THRESHOLD = 0.2
-
     class Meta:
         model = PageRev
         fields = ["title", "content"]
@@ -409,21 +406,11 @@ class PageRevisionForm(forms.ModelForm):
         super().__init__(*args, instance=instance, **kwargs)
         self.author = author
         self.page = page
-        self.initial_content = instance.content if instance else ""
-
-    def diff_ratio(self, new_str: str) -> float:
-        return difflib.SequenceMatcher(
-            None, self.initial_content, new_str
-        ).quick_ratio()
+        self.initial_obj: PageRev = copy(self.instance)
 
     def save(self, commit=True):  # noqa FBT002
         revision: PageRev = self.instance
-        if (
-            revision._state.adding
-            or revision.author != self.author
-            or revision.date + self.TIME_THRESHOLD < now()
-            or self.diff_ratio(revision.content) < (1 - self.DIFF_THRESHOLD)
-        ):
+        if not self.initial_obj.should_merge(self.instance):
             revision.author = self.author
             revision.page = self.page
             revision.id = None  # if id is None, Django will create a new record