From Fuzzy to Fluent: Semantic Code Search in Vim/Neovim

1.3K Views

November 04, 25

#vimconf

スライド概要

VimConf 2025 LT Slides by @Sirasagi62
Please check here for GIFs and other files that were not converted to PDF:
https://github.com/sirasagi62/talks/tree/main/2025/VimConf2025-LT

GIFなどのPDFで適切に変換されていない要素は上記リンクをご参照ください。

Gota Okabe

@sirasagi62

スライド一覧

Slides: https://github.com/sirasagi62/talks

またはPlayer版

埋め込む »CMSなどでJSが使えない場合

ダウンロード

関連スライド

猫でも分かるUnreal Engineの学び方 - 超初心者向け編 - 2023 v1.0

ue4 ue5 ue-beginner

エピックゲームズジャパン 1.6M

Unreal Engine5 Lumenの仕組みと肝心なところ

ue5 ue-rendering ue-lumen

エピックゲームズジャパン 1.4M

Meta XR SDK(V66-74)でQuestアプリを開発

spatial anchor unity quest pro shaperecognizeractivatestate oculus integration transformfeaturestateprovider building blocks transformrecognizeractivestate ovrsemanticclassification jointdeltaprovider ovrscenemanager jointvelocityactivestate オクルージョン sequenceactivestate scene manager ambisonic depth api metaxraudiosource playerlocomotor meta xr sdk quest3 ovrplayercontroller マルチモーダル meta haptics studio direct touch ui meta xr haptics sdk ovrspatialanchor ovrtrackedkeyboard hapticclipplayer fingerfeaturestateprovider hapticclip ワイドモーションモード wmm mruk mr utility kit voice sdk jointrotationactivestate meta horizon os ui set asw application spacewarp ovr metrics tool unityscene manager colocation discovery コロケーション mx ink passthrough camera api hand tracking microgestures webcamtexturemanager passthroughcamerautils cameraviewermanager hand pose selector recorder

あうぜん 1.1M

UE5レンダリングフロー総おさらい(2024) 基礎編！[CEDEC+KYUSHU 2024]

ue5 unreal engine ue-rendering

エピックゲームズジャパン 1.1M

猫でも分かる UE5.0, 5.1 におけるアニメーションの新機能について【CEDEC+KYUSHU 2022】

ue5 cedec+kyushu ue-animation ue-optimize ue-bp ue-physics ue-sequencer

エピックゲームズジャパン 1M

各ページのテキスト

From to Semantic Code Search in Vim/Neovim By OKABE Gota (@sirasagi62 1

I’m Sirasagi62 Sendai, Japan Tohoku Univ. TS/Go/Python/C++ … About 10 years 2

Finder 3

No, It’s 5

Motivation Finding code is difficult “Searching with different verb with same meaning” “Can’t think of what words to search” 6

Motivation Finding code is difficult “Searching with different verb with same meaning” “Can’t think of what words to search” → Cannot use synonym or sentences 7

What is Fluent Finder and flf-vim flf-vim flf-vim finds code semantically → AI understands the meaning of the code 8

Example 1: FlfDir flf-vim 10

10.

Problem flf-vim 11

11.

Problem ”FlfDir” takes a very long time flf-vim Example: Analyze vim source code → It takes 12

12.

Problem ”FlfDir” takes a very long time flf-vim Example: Analyze vim source code → It takes over 1 hour 13

13.

Problem ”FlfDir” takes a very long time flf-vim Example: Analyze vim source code → It takes over 1 hour Not 14

14.

Example 2: FlfBuf flf-vim 16

15.

Embedding Model Use SLMSmall Language Model) • Safe • Cost-Effective • Fast flf-vim

16.

Embedding Model “ruri-v370m-code-v-0.1” • Finetuned “ruri-v3” flf-vim

17.