From 80c14628664a3a425fab1279fe989fcea25b9c3e Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Tao=20Bror=20Bojl=C3=A9n?= <tao@btao.org>
Date: Sat, 3 Aug 2019 00:25:39 +0300
Subject: [PATCH] improve search result relevance

---
 CHANGELOG.md                              |  1 +
 backend/lib/backend/api.ex                | 11 ++++-------
 backend/priv/elasticsearch/instances.json | 24 +++++------------------
 3 files changed, 10 insertions(+), 26 deletions(-)

diff --git a/CHANGELOG.md b/CHANGELOG.md
index 45dcbfe..d2e8c0b 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -15,6 +15,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 ### Changed
 
 - Performance improvements when opening the app on something that isn't the graph.
+- There are now fewer irrelevant search results.
 
 ### Deprecated
 
diff --git a/backend/lib/backend/api.ex b/backend/lib/backend/api.ex
index 2521054..8b3d017 100644
--- a/backend/lib/backend/api.ex
+++ b/backend/lib/backend/api.ex
@@ -115,6 +115,7 @@ defmodule Backend.Api do
         "sort" => "_score",
         "from" => from,
         "size" => page_size,
+        "min_score" => 1,
         "query" => %{
           "bool" => %{
             "filter" => %{
@@ -127,7 +128,8 @@ defmodule Backend.Api do
                 "multi_match" => %{
                   "query" => query,
                   "fields" => [
-                    "description.english"
+                    "description.english",
+                    "domain.english"
                   ]
                 }
               },
@@ -143,14 +145,9 @@ defmodule Backend.Api do
                 "wildcard" => %{
                   "domain.keyword" => %{
                     "value" => "*#{query}*",
-                    "boost" => 1
+                    "boost" => 50
                   }
                 }
-              },
-              %{
-                "match" => %{
-                  "domain.ngram^0.5" => query
-                }
               }
             ]
           }
diff --git a/backend/priv/elasticsearch/instances.json b/backend/priv/elasticsearch/instances.json
index 2151b60..714889b 100644
--- a/backend/priv/elasticsearch/instances.json
+++ b/backend/priv/elasticsearch/instances.json
@@ -1,21 +1,7 @@
 {
   "settings": {
     "number_of_shards": 1,
-    "number_of_replicas": 0,
-    "analysis": {
-      "analyzer": {
-        "ngramAnalyzer": {
-          "tokenizer": "ngramTokenizer"
-        }
-      },
-      "tokenizer": {
-        "ngramTokenizer": {
-          "type": "ngram",
-          "min_gram": 5,
-          "max_gram": 5
-        }
-      }
-    }
+    "number_of_replicas": 0
   },
   "mappings": {
     "_doc": {
@@ -23,12 +9,12 @@
         "domain": {
           "type": "text",
           "fields": {
-            "ngram": {
-              "type": "text",
-              "analyzer": "ngramAnalyzer"
-            },
             "keyword": {
               "type": "keyword"
+            },
+            "english": {
+              "type": "text",
+              "analyzer": "english"
             }
           }
         },