Add option to only act on duplicate messages

kdeldycke · May 20, 2021 · 73629a4 · 73629a4
1 parent 7ba6bc3
commit 73629a4
Show file tree

Hide file tree

Showing 3 changed files with 22 additions and 7 deletions.
diff --git a/mail_deduplicate/__init__.py b/mail_deduplicate/__init__.py
@@ -118,6 +118,7 @@ class Config:
     default_conf = {
         "dry_run": False,
         "input_format": False,
+        "only_act_on_duplicates": False,
         "force_unlock": False,
         "hash_only": False,
         "hash_headers": HASH_HEADERS,

diff --git a/mail_deduplicate/cli.py b/mail_deduplicate/cli.py
@@ -93,6 +93,13 @@ def validate_regexp(ctx, param, value):
     "supports maildir and mbox format. Use this option to open up other box "
     "format, or bypass unreliable detection.",
 )
+@click.option(
+    "-o",
+    "--only-act-on-duplicates",
+    is_flag=True,
+    default=False,
+    help="Perform any actions only on duplicate messages, not on unique messages.",
+)
 @click.option(
     "-u",
     "--force-unlock",
@@ -239,6 +246,7 @@ def mdedup(
     ctx,
     dry_run,
     input_format,
+    only_act_on_duplicates,
     force_unlock,
     hash_only,
     hash_header,
@@ -346,6 +354,7 @@ def mdedup(
     conf = Config(
         dry_run=dry_run,
         input_format=input_format,
+        only_act_on_duplicates=only_act_on_duplicates,
         force_unlock=force_unlock,
         hash_only=hash_only,
         hash_headers=hash_header,

diff --git a/mail_deduplicate/deduplicate.py b/mail_deduplicate/deduplicate.py
@@ -48,7 +48,8 @@
         ("mail_hashes", "Number of unique hashes."),
         (
             "mail_unique",
-            "Number of unique mails (which where automatically added to selection).",
+            "Number of unique mails (which where automatically added to selection "
+            "unless --only-act-on-duplicates was passed on the command line).",
         ),
         (
             "mail_duplicates",
@@ -404,9 +405,10 @@ def select_all(self):
             if mail_count == 1:
                 logger.debug("Add unique message to selection.")
                 self.stats["mail_unique"] += 1
-                self.stats["mail_selected"] += 1
-                self.stats["set_single"] += 1
-                candidates = mail_set
+                if self.conf.only_act_on_duplicates is False:
+                    self.stats["mail_selected"] += 1
+                    self.stats["set_single"] += 1
+                    candidates = mail_set
 
             # We need to resort to a selection strategy to discriminate mails
             # within the set.
@@ -457,9 +459,12 @@ def check_stats(self):
         # Mail grouping by hash.
         assert self.stats["mail_retained"] >= self.stats["mail_unique"]
         assert self.stats["mail_retained"] >= self.stats["mail_duplicates"]
-        assert self.stats["mail_retained"] == (
-            self.stats["mail_unique"] + self.stats["mail_duplicates"]
-        )
+        if self.conf.only_act_on_duplicates:
+            assert self.stats["mail_retained"] == self.stats["mail_duplicates"]
+        else:
+            assert self.stats["mail_retained"] == (
+                self.stats["mail_unique"] + self.stats["mail_duplicates"]
+            )
         # Mail selection stats.
         assert self.stats["mail_retained"] >= self.stats["mail_skipped"]
         assert self.stats["mail_retained"] >= self.stats["mail_discarded"]