From 3f46c2b54618e92e05bd2c7635fb6e9343cdd960 Mon Sep 17 00:00:00 2001
From: Fledge Shiu <xzk0701@gmail.com>
Date: Wed, 19 Jul 2023 11:57:22 -0400
Subject: [PATCH] Refactor test_collections in scraper.

---
 nautiluszim/scraper.py | 17 ++++++++++++-----
 1 file changed, 12 insertions(+), 5 deletions(-)

diff --git a/nautiluszim/scraper.py b/nautiluszim/scraper.py
index ba42240..2c71816 100644
--- a/nautiluszim/scraper.py
+++ b/nautiluszim/scraper.py
@@ -339,6 +339,9 @@ def test_collection(self):
         nb_files = sum([len(i.get("files", [])) for i in self.json_collection])
         logger.info(f"Collection loaded. {nb_items} items, {nb_files} files")
 
+        self.test_files()
+
+    def test_files(self):
         with zipfile.ZipFile(self.archive_path, "r") as zh:
             all_names = zh.namelist()
 
@@ -356,16 +359,20 @@ def test_collection(self):
                 except ValueError:
                     missing_files.append(entry["title"])
 
+        duplicate_file_names = set(
+            [
+                filename
+                for filename in all_file_names
+                if all_file_names.count(filename) > 1
+            ]
+        )
+
         if missing_files:
             raise ValueError(
                 "File(s) referenced in collection but missing:\n - "
                 + "\n - ".join(missing_files)
             )
-        duplicate_file_names = set([
-            filename
-            for filename in all_file_names
-            if all_file_names.count(filename) > 1
-        ])
+
         if duplicate_file_names:
             raise ValueError(
                 "Files in collection are duplicate:\n - "