From 1ce5f0a920584e8681caa2c51011e72ac9c6299f Mon Sep 17 00:00:00 2001
From: itepechi <72330683+itepechi@users.noreply.github.com>
Date: Sun, 6 Aug 2023 08:08:56 +0900
Subject: [PATCH] Add a guide to customize the behavior of PGroonga

---
 README.md | 35 +++++++++++++++++++++++++++++++++++
 1 file changed, 35 insertions(+)

diff --git a/README.md b/README.md
index a77aca475..57fcb3ffc 100644
--- a/README.md
+++ b/README.md
@@ -194,6 +194,41 @@ podman restart akkoma-web
 
 You can revert the applied changes at any time by replacing `migrate` with `rollback`.
 
+#### Change PGroonga indexing options
+
+By default, PGroonga uses a bigram tokenizer with prefix matching enabled.
+
+If you're not happy with this, you can change the behavior using the `database set_pgroonga_options` command.
+
+If you want to support Japanese in your instance and do not want to match "東京都" (Tokyo-to) with "京都" (Kyoto), pass `TokenMecab` as the tokenizer.
+
+```sh
+podman exec -it akkoma-web \
+  pleroma_ctl \
+  database \
+  set_pgroonga_options "tokenizer='TokenMecab'"
+```
+
+If you want to disable prefix matching, e.g. you want "akko" to match "Akko is cute!" but not "Akkoma is cool!", pass `hash_table` as the lexicon type.
+
+```sh
+podman exec -it akkoma-web \
+  pleroma_ctl \
+  database \
+  set_pgroonga_options "lexicon_type='hash_table'"
+```
+
+If you want to set multiple options, simply concatenate them with a comma.
+
+_Do not add spaces after commas. **Elixir will complain!**_
+
+```sh
+podman exec -it akkoma-web \
+  pleroma_ctl \
+  database \
+  set_pgroonga_options "tokenizer='TokenMecab',lexicon_type='hash_table'"
+```
+
 ---
 
 ## Akkoma