Hello! 👋

👤 Hi, I am Yuichi Tateno, a Japanese software engineer — online as @hotchpotch or id:secondlife.

🔎 Currently: focused on IR (information retrieval) research and development.

🌐 My sites: hotchpotch.dev(en) | secon.dev(ja)

_{My profile icon is a robot head drooling into a puddle.}

NLP, IR, ML, DS & AI

• ⚡ fast-bunkai (75 stars) - 🇯🇵 Fast Japanese sentence splitting powered by Rust.
• 🪚 open_provence (65 stars) - Efficient context pruning for retrieval-augmented generation.
• ❓ JQaRA (45 stars) - 🇯🇵 Japanese question-answering benchmark for RAG evaluation.
• 📖 beko-translate (36 stars) - 🇯🇵 Native Mac translation for text and PDFs.
• ⚖️ HAKARI-Bench (30 stars) - Lightweight nano-scale benchmark for retrieval model decisions.
• 🎯 yast (21 stars) - Train sparse retrieval models with SPLADE.
• 🗄️ sqlite-vaporetto (20 stars) - 🇯🇵 Japanese full-text search for SQLite FTS5.
• 🌱 yasem (13 stars) - Simple efficient library for sparse SPLADE embeddings.
• 🧠 wikipedia-passages-jawiki-embeddings-utils (12 stars) - 🇯🇵 Build Japanese Wikipedia embeddings and FAISS indexes.
• 🦆 duckdb-vaporetto (11 stars) - 🇯🇵 Japanese full-text search extension for DuckDB.
• 🧭 JaCWIR - 🇯🇵 Casual Japanese web retrieval evaluation dataset.
• 🪞 similar-documents-cli - Find semantically similar documents from the command line.
• 🌍 trn - Fast translation CLI using Apple's native framework.
• 🗜️ ncd_classifier - Classify text using compression distance without parameters.
• 🧑‍🏫 youri-7b-sft-qa-context-jaqket - 🇯🇵 Fine-tune Japanese QA models with retrieval context.
• 🎓 fineweb-2-edu-japanese-classifier - 🇯🇵 Score educational quality in Japanese web content.

Tools & Other Projects

• 🔌 openai-api-server-via-codex (33 stars) - Serve an OpenAI-compatible API through Codex.
• 🏗️ hatena-group-static-generator - 🇯🇵 Export Hatena Group content as static pages.
• 👆 herdr-tiny-fingers - Copy visible terminal text using keyboard hints.
• 🦕 next-with-typesaurus-example - Example Next.js application using typed Firestore access.
• 🖼️ pixoo-spotify - Display Spotify artwork on a Pixoo screen.
• 🚀 rapids-svr-svc-marc_ja - 🇯🇵 Train accelerated classifiers on Japanese MARC records.

• 🚰 http-dump (69 stars) - Inspect HTTP requests captured through Ruby WebMock.
• 🗺️ react-navigation-mobx-example (52 stars) - Demonstrate MobX integration with React Navigation.
• 🐪 perldoc-vim (44 stars) - Browse Perl documentation without leaving Vim.
• 📋 pry-clipboard (42 stars) - Copy Ruby Pry session results to clipboard.
• 🌈 GlitchKit (36 stars) - Create visual glitch effects for Objective-C apps.
• 🔤 android-IconicFontEngine (29 stars) - Render icon fonts conveniently inside Android applications.
• 🧩 chrome-extension-starter (26 stars) - Minimal starter template for Google Chrome extensions.
• ✂️ rascut (26 stars) - Ruby utilities for building ActionScript projects quickly.
• 🧱 activerecord-raw-data (22 stars) - Access uncast database values from ActiveRecord models.
• 📝 dotfiles-vim (19 stars) - Personal Vim configuration and editor customizations.
• 🎄 advent-calendar2readlists (16 stars) - 🇯🇵 Convert Advent Calendar entries into readable lists.
• 🚂 as3rails2u (16 stars) - Reusable ActionScript helpers for Rails-oriented applications.
• 🗃️ jsdeferred-webdatabase (15 stars) - WebDatabase mapper built around JSDeferred control flow.
• 🧮 sql-abstract-javascript (11 stars) - Generate SQL expressions from structured JavaScript objects.
• 📈 fluent-plugin-graphite - Forward Fluentd metrics into Graphite monitoring systems.
• 📬 fluent-plugin-resque - Send Fluentd events through Resque background jobs.
• 🌉 uffy - Bridge ActionScript and JavaScript application code.
• 📡 Arduino-i2cdetect - Scan connected devices on Arduino I2C buses.
• 📦 faraday_middleware-msgpack - MessagePack middleware integration for Ruby Faraday.
• 👥 hatenausermulti-xul - 🇯🇵 Switch between multiple Hatena accounts in Firefox.
• 🎨 jscolor - Compact color manipulation helpers written in JavaScript.
• 🧑‍🔧 moco - Command-line client for the online mbed compiler.
• 🍀 perl-acme-momoiroclover - 🇯🇵 Playful Perl module inspired by Momoiro Clover.
• 🌡️ Arduino-HDC1000 - Arduino library for HDC1000 environmental sensors.
• 🔢 Arduino-SN74141 - Arduino driver for SN74141 Nixie tube controllers.
• 🔎 atok-direct-google-moshikashite - 🇯🇵 Google suggestion plugin for Japanese ATOK input.
• 🗣️ fluent_interface - Build fluent method chains in Ruby modules.
• 📰 hatena-blog-entry-mlt-example - 🇯🇵 Example related-entry search for Hatena Blog posts.
• 🪄 meta_framework-vim - Experimental metaprogramming framework operated through Vim.
• 📷 olympus-camera - Ruby wrapper for the Olympus Camera API.
• ⌚ rubyistokei-ios-motion - 🇯🇵 iOS Motion implementation of the Rubyist clock.
• 📁 tempdir - Create temporary directories with a Tempfile-like interface.
• 🛤️ vim-rails - Legacy Rails development helpers for the Vim editor.
• 🔄 digit_convertor - Convert numeric digits between common display formats.
• 💬 ruboty-aa - 🇯🇵 Generate ASCII art through a Ruboty plugin.
• 🎞️ anigif2png-js - Convert animated GIF data into PNG frames.
• ☁️ elastic-mapreduce - Unofficial Ruby client for Amazon Elastic MapReduce.
• 📸 google-photo-downloader-w-hatena - 🇯🇵 Migrate Google Photos images into Hatena services.
• ⭐ hatena-star-autopageload-googlechrome - 🇯🇵 Automatically load Hatena Stars inside Google Chrome.
• 🏠 hotchpotch.github.com - Legacy source for the personal GitHub Pages site.
• 📐 iphone-view-iframe - Preview webpages inside an iPhone-sized iframe.
• 🌸 kawaii-generator - 🇯🇵 Generate playful Japanese kawaii graphics with ActionScript.
• 🖌️ rack-convert-webp - Convert served images into WebP through Rack.
• 🧙 rmagick-1.15.17-no-rdoc - Package legacy RMagick without generated documentation.
• ⏰ YA-BiseiTokei-android - 🇯🇵 Android companion app for a Japanese voice clock.
• 😀 emoji-ime-dictionary-ascii - 🇯🇵 Type emoji names through Japanese input methods.
• 📚 readlists-anonymous - Access Readlists APIs without authenticated user sessions.
• ✏️ sketch - Personal experimental sketches and small code fragments.
• 📱 WebdbReactNativeExample - 🇯🇵 React Native examples for WEB+DB PRESS readers.
• 🀄 baidu-translate-api-ruby - Ruby client for the Baidu Translate API.
• 💡 hue_ir_receiver - Control Philips Hue lights from infrared signals.
• 🤖 hyatena-bot - 🇯🇵 Experimental automation bot for Japanese Hatena services.
• ✍️ mbed-autowriter - Automate source updates for online mbed projects.
• 📥 olympus-camera-dl - Download Olympus camera photos directly over Wi-Fi.
• 🛕 shakyo-kozos - 🇯🇵 Code transcription exercises using classic software structures.