From 5be02807e354d4a2f58e8194330ab033fce2ed22 Mon Sep 17 00:00:00 2001
From: Mark Veidemanis <m@zm.is>
Date: Mon, 5 Sep 2022 07:20:30 +0100
Subject: [PATCH] Add cache indicator, time the fetch from Redis and round it

---
 core/lib/manticore.py | 50 ++++++++++++++++++++++++++++++++++++++++++-
 1 file changed, 49 insertions(+), 1 deletion(-)

diff --git a/core/lib/manticore.py b/core/lib/manticore.py
index 8d3dd76..447012a 100644
--- a/core/lib/manticore.py
+++ b/core/lib/manticore.py
@@ -1,12 +1,22 @@
+import json
+import logging
+import random
+import string
+import time
 from datetime import datetime
+from math import floor, log10
 from pprint import pprint
 
 import manticoresearch
 from django.conf import settings
+from siphashc import siphash
 
+from core import r
 from core.lib.processing import annotate_results, filter_blacklisted, parse_results
 from core.views import helpers
 
+logger = logging.getLogger(__name__)
+
 
 def initialise_manticore():
     """
@@ -22,6 +32,22 @@ def initialise_manticore():
 api_client, client = initialise_manticore()
 
 
+def initialise_caching():
+    hash_key = r.get("cache_hash_key")
+    if not hash_key:
+        letters = string.ascii_lowercase
+        hash_key = "".join(random.choice(letters) for i in range(16))
+        logger.debug(f"Created new hash key: {hash_key}")
+        r.set("cache_hash_key", hash_key)
+    else:
+        hash_key = hash_key.decode("ascii")
+        logger.debug(f"Decoded hash key: {hash_key}")
+    return hash_key
+
+
+hash_key = initialise_caching()
+
+
 def construct_query(query, size, index, blank=False):
     """
     Accept some query parameters and construct an OpenSearch query.
@@ -42,9 +68,27 @@ def construct_query(query, size, index, blank=False):
 
 
 def run_query(client, user, search_query):
+    start = time.process_time()
+    query_normalised = json.dumps(search_query, sort_keys=True)
+    hash = siphash(hash_key, query_normalised)
+    cache_hit = r.get(f"query_cache.{user.id}.{hash}")
+    if cache_hit:
+        print("Cache hit")
+        response = json.loads(cache_hit)
+        time_took = (time.process_time() - start) * 1000
+        # Round to 3 significant figures
+        time_took_rounded = round(time_took, 3 - int(floor(log10(abs(time_took)))) - 1)
+        response["took"] = time_took_rounded
+        response["cache"] = True
+        return response
     response = client.search(search_query)
     response = response.to_dict()
     filter_blacklisted(user, response)
+    print("Writing to cache")
+    to_write_cache = json.dumps(response)
+    r.set(f"query_cache.{user.id}.{hash}", to_write_cache)
+    r.expire(f"query_cache.{user.id}.{hash}", 30)
+    print("Written to cache")
     return response
 
 
@@ -166,7 +210,9 @@ def query_results(
     for source_iter in sources:
         add_top_tmp["bool"]["should"].append({"equals": {"src": source_iter}})
         total_count += 1
-    total_sources = len(settings.MANTICORE_MAIN_SOURCES) + len(settings.MANTICORE_SOURCES_RESTRICTED)
+    total_sources = len(settings.MANTICORE_MAIN_SOURCES) + len(
+        settings.MANTICORE_SOURCES_RESTRICTED
+    )
     if not total_count == total_sources:
         add_top.append(add_top_tmp)
 
@@ -293,4 +339,6 @@ def query_results(
         "card": results["hits"]["total"],
         "took": results["took"],
     }
+    if "cache" in results:
+        context["cache"] = results["cache"]
     return context