Tool to quickly explain the usage of words in Japanese sentences by using Large Language models (Currently with GPT-4o-mini).
This Readme is a basic description of the program's functions.
Before running the program, you'll need to set up the following environment variables for API keys:
- OPENAI_API_KEY: Your API key for OpenAI.
- AZURE_SPEECH_API_KEY: Your API key for Azure Speech service (optional, for text-to-speech functionality).
To run the program, simply execute the jexplain.py file with Python.
- Ctrl+Win+Z: Explain the meaning of each word in the context of the sentence that is highlighted (autofocuses terminal window)
- Ctrl+Win+A: Ask a question to the AI
- Ctrl+Win+Tab: Translate the contents of the clipboard using ChatGPT
- Ctrl+Win+-: Aggressive translation. Uses chain of thought to generate a literary translation for difficult phrases.
- Ctrl+Win+K: Explain the meaning of each Kanji in the highlighted word.
- Ctrl+Win+]: Generate a mnemonic sentence with the highlighted text.
- Ctrl+Win+:: Speak the highlighted Japanese text.
- Ctrl+Win+F12: Speak the highlighted English text.
- Ctrl+Win+[: Change the copy mode. This is remembered next time you run the script. See copy_modes.py to see the modes. The mokuro mode works for reading manga btw.
- Word Explanations: The program uses the gpt-4o-mini language model API to provide context-specific definitions of Japanese words within sentences including furigana (furigana may contain inaccuracies)
- Kanji Breakdown: It can also explain the meaning of individual kanji characters within a word.
- Translation: OpenAI API is used to translate Japanese text into English.
- Text-to-Speech: With the Azure Speech service, you can hear the pronunciation of highlighted Japanese or English text.
- Mnemonic Generation: The program can create mnemonic sentences using highlighted text to aid in memorization for JPDB.io Kanji pages.
- AI language model Interaction: You can directly ask questions to the gpt-4o-mini language model API for a more conversational experience. The AI will forget everything for each question, so no back and forth chats.
- Ensure you have the required Python libraries installed (openai, keyboard, pyautogui, pyperclip, pywin32, win32console, azure-cognitiveservices-speech, json, os, time, mecab-python3).
- The program relies on keyboard shortcuts for activation.
- It utilizes the clipboard for text input.
- The terminal window is used for displaying outputs.
- "pip install mecab-python3" should resolve the mecab dependency. Windows binary is here if that doesn't work: https://github.com/ikegami-yukino/mecab/releases . If it's still not working on Windows, see: https://stackoverflow.com/questions/63197703/mecab-importerror-dll-load-failed-the-specified-module-could-not-be-found
- For MPV realtime subtitles analysis, add the lua scripts to your MPV scripts folder. (eg. "%APPDATA%\mpv\scripts")