# Persian Language Tutor

A Gradio-based Persian (Farsi) language learning app for English speakers, built around GCSE Persian vocabulary (Pearson specification).

## Features

- **Vocabulary Study** — Search, browse, and study 918 GCSE Persian words across 39 categories
- **Flashcards with FSRS** — Spaced repetition scheduling (same algorithm as Anki)
- **Idioms & Expressions** — 25 Persian social conventions with cultural context
- **AI Tutor** — Conversational Persian lessons by GCSE theme (via Ollama)
- **Essay Marking** — Write Persian essays, get AI feedback and grading (via Claude)
- **Dashboard** — Track progress, streaks, and mastery
- **Anki Export** — Generate .apkg decks for offline study
- **Voice Input** — Speak Persian via microphone (Whisper STT) in the Tutor tab

## Prerequisites

- `whisper-ollama` conda environment with Python 3.10+
- Ollama running locally with `qwen2.5:7b` (or another model)
- Claude CLI installed (for essay marking / smart mode)

## Setup

```bash
/home/ys/miniforge3/envs/whisper-ollama/bin/pip install gradio genanki fsrs
```

## Running the app

```bash
cd /home/ys/family-repo/Code/python/persian-tutor
/home/ys/miniforge3/envs/whisper-ollama/bin/python app.py
```

Then open http://localhost:7860 in your browser.

## Running tests

```bash
cd /home/ys/family-repo/Code/python/persian-tutor
/home/ys/miniforge3/envs/whisper-ollama/bin/python -m pytest tests/ -v
```

41 tests covering db, vocab, ai, and anki_export modules.

## Expanding vocabulary

The vocabulary can be expanded by editing `data/vocabulary.json` directly or by updating `scripts/build_vocab.py` and re-running it:

```bash
/home/ys/miniforge3/envs/whisper-ollama/bin/python scripts/build_vocab.py
```

## TODO

- [ ] Voice-based vocabulary testing — answer flashcard prompts by speaking Persian
- [ ] Improved UI theme and layout polish