From 9ae1021074da3049a608c5f01d2cc3318d117c4c Mon Sep 17 00:00:00 2001
From: benoit74 <benoit74@users.noreply.github.com>
Date: Wed, 31 Jan 2024 11:47:01 +0100
Subject: [PATCH] Allow to add a ZIM scraper suffix via CLI argument

---
 src/warc2zim/__about__.py | 2 +-
 src/warc2zim/converter.py | 4 +++-
 src/warc2zim/main.py      | 6 ++++++
 3 files changed, 10 insertions(+), 2 deletions(-)

diff --git a/src/warc2zim/__about__.py b/src/warc2zim/__about__.py
index 494af57f..11fb1b2f 100644
--- a/src/warc2zim/__about__.py
+++ b/src/warc2zim/__about__.py
@@ -1 +1 @@
-__version__ = "2.0.0-dev0"
+__version__ = "2.0.0-dev1"
diff --git a/src/warc2zim/converter.py b/src/warc2zim/converter.py
index 96f33100..c0b399bd 100644
--- a/src/warc2zim/converter.py
+++ b/src/warc2zim/converter.py
@@ -142,6 +142,8 @@ def __init__(self, args):
 
         self.written_records = self.total_records = 0
 
+        self.scraper_suffix = args.scraper_suffix
+
     def init_env(self):
         # autoescape=False to allow injecting html entities from translated text
         env = Environment(
@@ -240,7 +242,7 @@ def run(self):
             Illustration_48x48_at_1=self.illustration,
             Tags=";".join(self.tags),
             Source=self.source,
-            Scraper=f"warc2zim {get_version()}",
+            Scraper=f"warc2zim {get_version()}{self.scraper_suffix}",
         ).start()
 
         for filename in importlib.resources.files("warc2zim.statics").iterdir():
diff --git a/src/warc2zim/main.py b/src/warc2zim/main.py
index b459f0f0..d2582c1c 100644
--- a/src/warc2zim/main.py
+++ b/src/warc2zim/main.py
@@ -79,6 +79,12 @@ def main(raw_args=None):
         default="",
     )
 
+    parser.add_argument(
+        "--scraper-suffix",
+        help="Additional string to append as a suffix to ZIM Scraper metadata, in "
+        "addition to regular warc2zim value",
+    )
+
     args = parser.parse_args(args=raw_args)
     converter = Converter(args)
     return converter.run()