feat(export): add Markdown and JSON task export with GUI dialog

.github/workflows/ci.yml

@@ -0,0 +1,59 @@
+name: CI
+
+on:
+  push:
+    branches: [ main, development ]
+  pull_request:
+    branches: [ main, development ]
+
+jobs:
+  test:
+    name: test (${{ matrix.os }}, py${{ matrix.python-version }})
+    runs-on: ${{ matrix.os }}
+    strategy:
+      fail-fast: false
+      matrix:
+        os: [ ubuntu-latest, windows-latest ]
+        python-version: [ "3.11", "3.12" ]
+
+    steps:
+      - name: Checkout Code
+        uses: actions/checkout@v4
+
+      - name: Set up Python ${{ matrix.python-version }}
+        uses: actions/setup-python@v5
+        with:
+          python-version: ${{ matrix.python-version }}
+          cache: "pip"
+          cache-dependency-path: "pyproject.toml"
+
+      - name: Install Linux dependencies
+        if: matrix.os == 'ubuntu-latest'
+        run: |
+          sudo apt-get update
+          sudo apt-get install -y libportaudio2 libasound2-dev
+
+      - name: Install dependencies
+        run: |
+          python -m pip install --upgrade pip
+          pip install -e .[dev,nlp,whisper]
+
+      - name: Run Ruff Linter
+        run: ruff check src/
+
+      - name: Run Mypy Type Checker
+        run: mypy src/
+
+      - name: Run Safety Check
+        run: safety check
+
+      - name: Run Pytest with Coverage
+        # GUI tests might require mock display on Ubuntu. Let's use xvfb on Ubuntu runner.
+        if: matrix.os == 'ubuntu-latest'
+        run: |
+          sudo apt-get install -y xvfb
+          xvfb-run --server-args="-screen 0 1024x768x24" pytest tests/ -v --cov=src --cov-fail-under=80 --cov-report=xml
+
+      - name: Run Pytest with Coverage (Windows)
+        if: matrix.os == 'windows-latest'
+        run: pytest tests/ -v --cov=src --cov-fail-under=80 --cov-report=xml

.gitignore

@@ -42,3 +42,8 @@ models/
 logs/
 *.log
 .gemini/
+.coverage
+htmlcov/
+.pytest_cache/
+.mypy_cache/
+.ruff_cache/

Makefile

@@ -0,0 +1,18 @@
+# Taskify Makefile target shortcuts
+
+.PHONY: dist clean test lint typecheck
+
+dist:
+	python packaging/build.py
+
+clean:
+	python -c "import shutil; [shutil.rmtree(p, ignore_errors=True) for p in ('build', 'dist')]"
+
+test:
+	pytest tests/ -v
+
+lint:
+	ruff check src/ --fix
+
+typecheck:
+	mypy src/

docs/developer_guide.md

@@ -0,0 +1,140 @@
+# Taskify Developer Guide
+
+This guide describes the architecture, codebase organization, database schema, and test suites for developers contributing to Taskify.
+
+---
+
+## 1. Directory Structure & Architecture
+
+Taskify's architecture is modular, separating audio capture, speech recognition, database storage, extraction scheduling, and the user interface.
+
+```
+src/taskify/
+├── __init__.py
+├── __main__.py          # Program execution launcher
+├── cli/                 # Click command-line interface
+├── audio/               # PyAudio / sounddevice capture streams
+├── stt/                 # Vosk & Whisper speech transcription backends
+├── storage/             # SQLite migrations and CRUD data access
+├── llm/                 # Ollama and rule-based NLP extraction engines
+├── pipeline/            # Background extraction worker scheduler
+└── ui/                  # customtkinter graphical widgets
+```
+
+### Core Architecture Flowchart
+
+```mermaid
+graph TD
+    A[Microphone Stream] --> B[Audio Capture Queue]
+    B --> C[STT Engine: Vosk/Whisper]
+    C --> D[Transcript Database & Rotated Logs]
+    D --> E[Extraction Scheduler]
+    E --> F[LLM / NLP Extraction Engine]
+    F --> G[Tasks & Matrix Database Tables]
+    G --> H[Eisenhower Matrix GUI Dashboard]
+```
+
+---
+
+## 2. Codebase Components
+
+### 1. Configuration (`src/taskify/config/`)
+- Uses `Settings` dataclasses mapped from default settings TOML and user-customizable overrides.
+
+### 2. Audio Capture (`src/taskify/audio/`)
+- Uses `sounddevice` to stream raw float32 array samples to a thread-safe Queue.
+- Implements energy-based silence detection algorithms to pause downstream transcribing during quiet periods.
+
+### 3. Speech-to-Text (`src/taskify/stt/`)
+- Implements the `STTEngine` interface:
+  - `VoskEngine`: Primary offline, light model.
+  - `WhisperEngine`: Heavy model wrapper utilizing `faster-whisper`.
+- Models are lazy loaded and resolved via the `get_stt_engine` factory.
+
+### 4. Storage Engine (`src/taskify/storage/`)
+- `DatabaseManager`: Exposes thread-safe connections using SQLite stdlib. Run database migration scripts sequentially from `storage/migrations/`.
+- `TranscriptWriter`: Background worker flushing transcripts to SQLite and rotating logs at midnight.
+
+### 5. Extraction Pipeline (`src/taskify/pipeline/`)
+- `ExtractionScheduler`: Periodically triggers an extraction cycle in a daemon thread.
+- `EventBus`: Decoupled publisher-subscriber hub matching pipeline updates to frontend dashboard refreshes.
+
+---
+
+## 3. SQLite Database Schema
+
+Taskify uses four relational tables tracking capture sessions, segments, task lists, and categorization entries:
+
+```sql
+-- Schema version tracking
+CREATE TABLE schema_migrations (
+    version INTEGER PRIMARY KEY,
+    applied_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP
+);
+
+-- Voice capture sessions
+CREATE TABLE sessions (
+    id TEXT PRIMARY KEY,
+    start_time REAL NOT NULL,
+    end_time REAL,
+    created_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP
+);
+
+-- Transcribed audio segment lines
+CREATE TABLE transcripts (
+    id INTEGER PRIMARY KEY AUTOINCREMENT,
+    session_id TEXT REFERENCES sessions(id) ON DELETE CASCADE,
+    text TEXT NOT NULL,
+    confidence REAL NOT NULL,
+    start_time REAL NOT NULL,
+    end_time REAL NOT NULL,
+    processed INTEGER DEFAULT 0 CHECK(processed IN (0, 1)),
+    created_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP
+);
+
+-- Extracted tasks
+CREATE TABLE tasks (
+    id INTEGER PRIMARY KEY AUTOINCREMENT,
+    title TEXT NOT NULL,
+    notes TEXT,
+    due_date TEXT,
+    completed INTEGER DEFAULT 0 CHECK(completed IN (0, 1)),
+    user_override INTEGER DEFAULT 0 CHECK(user_override IN (0, 1)),
+    created_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP,
+    updated_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP
+);
+
+-- Eisenhower matrix quadrant entries
+CREATE TABLE matrix_entries (
+    task_id INTEGER PRIMARY KEY REFERENCES tasks(id) ON DELETE CASCADE,
+    quadrant TEXT CHECK(quadrant IN ('urgent_important', 'important_not_urgent', 'urgent_not_important', 'neither')),
+    created_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP,
+    updated_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP
+);
+```
+
+---
+
+## 4. Development Setup & Tests
+
+### Installation
+1. Clone the repository and establish a Python virtual environment:
+   ```bash
+   python -m venv .venv
+   .venv\Scripts\activate  # Windows
+   source .venv/bin/activate  # macOS/Linux
+   ```
+2. Install with all development extras:
+   ```bash
+   pip install -e .[dev,nlp,whisper]
+   ```
+
+### Makefile Shortcuts
+Taskify includes automated quality checks and builders.
+
+| Action | Makefile Target | CLI Alternative |
+|---|---|---|
+| Run Complete Test Suite | `make test` | `pytest tests/ -v` |
+| Code Lint Check | `make lint` | `ruff check src/` |
+| Type-Checking | `make typecheck` | `mypy src/` |
+| Package Executable | `make dist` | `python packaging/build.py` |

docs/model_download.md

@@ -0,0 +1,59 @@
+# Offline STT Model Download Guide
+
+Taskify is a **privacy-first application** that runs all speech-to-text (STT) models locally on your system. No audio data or transcriptions are ever sent online.
+
+---
+
+## 1. Storage Directories
+
+Local model files are stored in the following platform-specific application directories:
+
+| OS | Default Download Directory |
+|---|---|
+| **Windows** | `C:\Users\<Username>\.local\share\taskify\models\` |
+| **macOS** | `/Users/<Username>/.local/share/taskify/models/` |
+| **Linux** | `/home/<Username>/.local/share/taskify/models/` |
+
+---
+
+## 2. Vosk Models Setup
+
+### Automatic Download
+If you are connected to the internet, run the CLI utility to download and extract automatically:
+```bash
+# Download the default small english model
+taskify download-model vosk en-small
+```
+
+### Manual Offline Setup (Air-gapped)
+If you are setting up Taskify on an offline machine, follow these steps:
+1. **Download the model ZIP** from the official Vosk repository:
+   - **en-small** (Lightweight, default): [vosk-model-small-en-us-0.15.zip](https://alphacephei.com/vosk/models/vosk-model-small-en-us-0.15.zip) (Approx. 40 MB)
+   - **en-large** (High accuracy): [vosk-model-en-us-0.22-lgraph.zip](https://alphacephei.com/vosk/models/vosk-model-en-us-0.22-lgraph.zip) (Approx. 120 MB)
+2. **Extract the ZIP file** into your local Taskify model directory:
+   - Path for `en-small`: `.../taskify/models/vosk/vosk-model-small-en-us-0.15/`
+3. **Verify directory structure**:
+   - Ensure the directory contains database folders like `am/`, `graph/`, and description files such as `README`.
+
+---
+
+## 3. Whisper Models Setup
+
+### Automatic Download
+To automatically download standard HuggingFace quantized model parameters via `faster-whisper`:
+```bash
+# Downloads and registers the "base" model
+taskify download-model whisper base
+```
+Supported configurations: `tiny`, `base`, `small`.
+
+### Manual Offline Setup (Air-gapped)
+Huggingface Whisper weights are resolved via the `huggingface_hub` caching mechanism:
+1. On an online machine, download the model files using python:
+   ```python
+   from faster_whisper import WhisperModel
+   model = WhisperModel("base", device="cpu", compute_type="int8")
+   ```
+2. Copy the cached Hugging Face hub folders from your online machine to the offline machine:
+   - Cache Directory (default): `~/.cache/huggingface/hub/`
+3. Ensure the target directory layout is replicated exactly. Taskify will automatically detect the cache on start.

docs/privacy.md

@@ -0,0 +1,47 @@
+# Taskify Privacy Statement & Network Inventory
+
+Taskify is built from the ground up as a **fully offline, privacy-first task manager**. We believe your private spoken conversations, thoughts, plans, and task details belong entirely to you and should never leave your physical device.
+
+---
+
+## 1. Zero Network Interactions
+
+Taskify compiles a strict **Network Interaction Inventory** containing exactly zero runtime remote operations:
+
+- **No Remote Speech Recognition**: All audio captured via your microphone is transcribed locally using `vosk` or `faster-whisper`.
+- **No Cloud LLM Processing**: Task extractions and Eisenhower classifications are performed locally via Ollama (using models running on your system like `phi4` or `gemma`) or via our rule-based `NLPBackend` regex system.
+- **No Telemetry or Tracking**: The application does not collect usage stats, metrics, crash dumps, or run analytics trackers.
+- **Zero Remote Dependencies**: The application does not download runtime assets. (The only remote network requests are initiated explicitly by the user when running `taskify download-model` to fetch voice packs).
+
+---
+
+## 2. Local File & Data Storage Inventory
+
+All data generated by Taskify is stored in plain sight, locally on your machine. You can inspect, modify, back up, or delete it at any time.
+
+### Database
+- **Path**: `~/.local/share/taskify/taskify.db` (SQLite format)
+- **Contents**: Stores capture sessions, transcribed speech text segments, task lists, and Eisenhower Matrix categories.
+
+### Rotation Logs & Transcripts
+- **Path**: `~/.local/share/taskify/logs/`
+- **Markdown Logs (`YYYY-MM-DD.md`)**: A daily human-readable record of all conversations transcribed during that day.
+- **JSON Lines Logs (`YYYY-MM-DD.json`)**: Machine-readable logs of session transcripts for debugging.
+- **Application Diagnostics Log (`app.log`)**: Unified application operation logs (rotated automatically at 10 MB).
+
+### Configuration Files
+- **Default Package Configuration**: `src/taskify/config/config.default.toml`
+- **User Preference Configuration overrides**: `~/.config/taskify/config.toml` (TOML format)
+- **Contents**: Local audio rate, silence detection threshold, selected STT model parameters, and Ollama host server configurations.
+
+### Local Machine Models
+- **Path**: `~/.local/share/taskify/models/`
+- **Contents**: Extracted local speech recognition models (`vosk` directories).
+
+---
+
+## 3. Transparency & Verification
+
+Developers can inspect the codebase to verify our privacy assertions:
+1. Check [ollama_backend.py](file:///c:/Users/nihal/Desktop/Taskify/src/taskify/llm/ollama_backend.py): Outgoing requests are sent exclusively to the user's custom configured `ollama_host` (defaulting to local port `http://localhost:11434`).
+2. Check [capture.py](file:///c:/Users/nihal/Desktop/Taskify/src/taskify/audio/capture.py) and [transcript_writer.py](file:///c:/Users/nihal/Desktop/Taskify/src/taskify/storage/transcript_writer.py): Audio bytes and transcripts are directed exclusively to the local database and rotation log writers.