Use indexes, heap-based top k for python as the other languages do

python/related.py

@@ -1,62 +1,28 @@
+import heapq
 import json
-from typing import Dict
-
-
-class Post:
-    def __init__(self, _id, title, tags):
-        self._id = _id
-        self.title = title
-        self.tags = tags
-
-    def __hash__(self):
-        return hash(self._id)
-
-    def __eq__(self, other):
-        if isinstance(other, Post):
-            return self._id == other._id
-        return False
-
-    def to_dict(self):
-        return {
-            "_id": self._id,
-            "title": self.title,
-            "tags": self.tags,
-        }
-
-
-class PostWithSharedTags:
-    def __init__(self, post, shared_tags):
-        self.post = post
-        self.shared_tags = shared_tags
+from collections import Counter
 
 
 def main():
     with open("../posts.json") as f:
         posts = json.load(f)
 
     tag_map = {}
-    for post in posts:
+    for idx, post in enumerate(posts):
         for tag in post["tags"]:
             if tag not in tag_map:
                 tag_map[tag] = []
-            tag_map[tag].append(Post(post["_id"], post["title"], post["tags"]))
+            tag_map[tag].append(idx)
 
     all_related_posts = []
-    for post in posts:
-        related_posts_dict: Dict[Post, int] = {}
-        for tag in post["tags"]:
-            for related_post in tag_map[tag]:
-                if related_post._id != post["_id"]:
-                    if related_post not in related_posts_dict:
-                        related_posts_dict[related_post] = 0
-                    related_posts_dict[related_post] += 1
-
-        sorted_posts = sorted(
-            related_posts_dict.items(), key=lambda x: x[1], reverse=True
-        )
-
-        num = min(5, len(sorted_posts))
-        top_posts = [p[0].to_dict() for p in sorted_posts[:num]]
+    for this_post_idx, post in enumerate(posts):
+        related_posts_list = Counter((p for tag in post["tags"] for p in tag_map[tag]))
+        related_posts_list[this_post_idx] = 0
+
+        top_posts = [
+            {k: posts[p][k] for k in ("_id", "title", "tags")}
+            for p in heapq.nlargest(5, related_posts_list, key=related_posts_list.get)
+        ]
 
         all_related_posts.append(
             {

run.sh

@@ -12,7 +12,6 @@ run_go() {
     echo "Running Go" &&
         cd ./go &&
         go build &&
-        #command time -f '%es %Mk' ./related
         if [ $HYPER == 1 ]; then
             command hyperfine -r 10 --show-output "./related"
         else
@@ -25,7 +24,7 @@ run_rust() {
         cd ./rust &&
         cargo build --release &&
         if [ $HYPER == 1 ]; then
-            command hyperfine -r 10 --show-output "./target/release/rust"
+            command hyperfine -r 20 --show-output "./target/release/rust"
         else
             command time -f '%es %Mk' ./target/release/rust
         fi
@@ -46,7 +45,7 @@ run_python() {
     echo "Running Python" &&
         cd ./python &&
         if [ $HYPER == 1 ]; then
-            command hyperfine -r 1 "python3 ./related.py"
+            command hyperfine -r 2 --show-output "python3 ./related.py"
         else
             command time -f '%es %Mk' python3 ./related.py
         fi