From f994060149f68a1205e59dae9e99868ac12acd10 Mon Sep 17 00:00:00 2001
From: xenofem <xenofem@xeno.science>
Date: Sun, 3 Mar 2024 02:56:46 -0500
Subject: [PATCH] suggest works with similar titles that aren't already in the
 same series

---
 dlibrary/dlibrary.py | 55 ++++++++++++++++++++++++++++++++++++++++++--
 1 file changed, 53 insertions(+), 2 deletions(-)

diff --git a/dlibrary/dlibrary.py b/dlibrary/dlibrary.py
index 3498072..03f7f4e 100755
--- a/dlibrary/dlibrary.py
+++ b/dlibrary/dlibrary.py
@@ -821,6 +821,47 @@ def copy_recursive(src, dest):
         else:
             shutil.copyfile(item, dest / item.name)
 
+
+memoized_similarities = {}
+
+def similarity(a, b):
+    if len(a) < len(b):
+        shorter = a
+        longer = b
+    else:
+        shorter = b
+        longer = a
+    if len(shorter) == 0:
+        return 0
+
+    if (shorter, longer) in memoized_similarities:
+        return memoized_similarities[(shorter, longer)]
+
+    skip_similarity = similarity(shorter[1:], longer)
+    match_idx = longer.find(shorter[0])
+    if match_idx == -1:
+        result = skip_similarity
+    else:
+        take_similarity = 1 + similarity(shorter[1:], longer[match_idx+1:])
+        result = max(skip_similarity, take_similarity)
+
+    memoized_similarities[(shorter, longer)] = result
+    return result
+
+def top(items, n, key):
+    if len(items) == 0:
+        return []
+    winners = items[:1]
+    for item in items[1:]:
+        if len(winners) < n or key(item) > key(winners[-1]):
+            for i in range(len(winners)):
+                if key(item) > key(winners[i]):
+                    winners.insert(i, item)
+                    break
+            if len(winners) > n:
+                winners.pop()
+    return winners
+
 def generate(args):
     jenv = Environment(
         loader=PackageLoader("dlibrary"),
@@ -869,15 +910,25 @@ def generate(args):
         }
         works.append(work)
 
-    for work in works:
+    for (idx, work) in enumerate(works):
+        def suggestion_priority(other_work):
+            if other_work is work:
+                return -2
+            if work['series'] and work['series'] == other_work['series']:
+                return -1
+            return similarity(work['title'], other_work['title'])
+        suggested = top(works, 6, suggestion_priority)
+
         work_dir = site_dir / 'works' / work['id']
         viewer_dir = work_dir / 'view'
         viewer_dir.mkdir(parents=True, exist_ok=True)
         with open(work_dir / 'index.html', 'w') as f:
-            f.write(work_template.render(depth=2, work=work, title=work['title'], suggested=works[:5]))
+            f.write(work_template.render(depth=2, work=work, title=work['title'], suggested=suggested))
         with open(viewer_dir / 'index.html', 'w') as f:
             f.write(viewer_template.render(depth=3, work=work, title=work['title']))
 
+        print(f'\x1b[2K\r{idx+1}/{len(works)} works processed...', end='')
+
     def make_categorization(categorization, query, work_filter, work_style_cards=False):
         categorization_dir = site_dir / categorization