From b1d30cadcf827ec9ac069da3910d6ec7d37d7d0c Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Rapha=C3=ABl=20Vinot?= <raphael@vinot.info>
Date: Fri, 29 Jul 2022 11:02:51 +0200
Subject: [PATCH] fix: Attempt to fix the stats page

---
 lookyloo/capturecache.py | 15 ++++-----------
 lookyloo/lookyloo.py     | 15 ++++++++++++++-
 2 files changed, 18 insertions(+), 12 deletions(-)

diff --git a/lookyloo/capturecache.py b/lookyloo/capturecache.py
index 290329e..2b464d1 100644
--- a/lookyloo/capturecache.py
+++ b/lookyloo/capturecache.py
@@ -117,6 +117,7 @@ class CapturesIndex(Mapping):
                     and not cc.incomplete_redirects):
                 self.__cache[uuid] = cc
                 return self.__cache[uuid]
+        # The tree isn't cached yet
         try:
             tree = load_pickle_tree(capture_dir, capture_dir.stat().st_mtime)
         except TreeNeedsRebuild:
@@ -230,7 +231,7 @@ class CapturesIndex(Mapping):
             lock_file.unlink(missing_ok=True)
         return tree
 
-    def _set_capture_cache(self, capture_dir: Path, tree: Optional[CrawledTree]=None) -> CaptureCache:
+    def _set_capture_cache(self, capture_dir: Path, tree: CrawledTree) -> CaptureCache:
         '''Populate the redis cache for a capture. Mostly used on the index page.
         NOTE: Doesn't require the pickle.'''
         with (capture_dir / 'uuid').open() as f:
@@ -257,17 +258,9 @@ class CapturesIndex(Mapping):
                 cache['timestamp'] = har.initial_start_time
                 cache['url'] = har.root_url
                 if har.initial_redirects and har.need_tree_redirects:
-                    if not tree:
-                        # try to load tree from disk
-                        tree = load_pickle_tree(capture_dir, capture_dir.stat().st_mtime)
                     # get redirects
-                    if tree:
-                        cache['redirects'] = json.dumps(tree.redirects)
-                        cache['incomplete_redirects'] = 0
-                    else:
-                        # Pickle not available
-                        cache['redirects'] = json.dumps(har.initial_redirects)
-                        cache['incomplete_redirects'] = 1
+                    cache['redirects'] = json.dumps(tree.redirects)
+                    cache['incomplete_redirects'] = 0
                 else:
                     cache['redirects'] = json.dumps(har.initial_redirects)
                     cache['incomplete_redirects'] = 0
diff --git a/lookyloo/lookyloo.py b/lookyloo/lookyloo.py
index 6b4c134..7defd80 100644
--- a/lookyloo/lookyloo.py
+++ b/lookyloo/lookyloo.py
@@ -968,7 +968,20 @@ class Lookyloo():
         stats: Dict[int, Dict[int, Dict[str, Any]]] = {}
         weeks_stats: Dict[int, Dict] = {}
 
-        for cache in self.sorted_capture_cache():
+        # Load the archived captures from redis
+        archived: List[CaptureCache] = []
+        p = self.redis.pipeline()
+        for directory in self.redis.hvals('lookup_dirs_archived'):
+            p.hgetall(directory)
+        for cache in p.execute():
+            if not cache:
+                continue
+            try:
+                archived.append(CaptureCache(cache))
+            except Exception:
+                continue
+
+        for cache in self.sorted_capture_cache() + archived:
             date_submission: datetime = cache.timestamp
 
             if date_submission.year not in stats: