feat: dockerize application, #14

.dockerignore

@@ -0,0 +1,31 @@
+.git
+.gitignore
+.venv
+venv
+env
+__pycache__
+*.pyc
+.pytest_cache
+.mypy_cache
+.ruff_cache
+.matplotlib_cache
+notebooks
+mlruns
+data/raw
+data/processed
+logs
+docs
+tests
+*.md
+!README.md
+.env
+.env.*
+frontend/dist
+models/checkpoints
+models/**/checkpoints
+models/experiments
+models/roberta_hate_results
+models/distilbert_results
+models/best_distilbert
+models/nb08_*
+models/*_frozen

.env.example

@@ -0,0 +1,15 @@
+# Copy to .env for local development:  cp env.example .env
+# Docker Compose reads these via environment (optional).
+
+# YouTube Data API v3 (optional — /predict-video and scraping)
+# https://console.cloud.google.com/apis/credentials
+YOUTUBE_API_KEY=
+
+# Active model (must match a key in ModelService.AVAILABLE_MODELS)
+MODEL_NAME=LR + TF-IDF (local)
+
+# development | production
+ENV=production
+
+# Used by Streamlit when calling the API from another host (Docker sets this automatically)
+API_URL=http://localhost:8000

Dockerfile

@@ -0,0 +1,40 @@
+# youtube_hate_detector — shared image for FastAPI + Streamlit services
+FROM python:3.12-slim-bookworm
+
+ENV PYTHONDONTWRITEBYTECODE=1 \
+    PYTHONUNBUFFERED=1 \
+    PYTHONPATH=/app \
+    NLTK_DATA=/app/nltk_data \
+    MODEL_NAME="LR + TF-IDF (local)" \
+    ENV=production
+
+WORKDIR /app
+
+# System deps for spaCy / sklearn wheels
+RUN apt-get update \
+    && apt-get install -y --no-install-recommends build-essential curl \
+    && rm -rf /var/lib/apt/lists/*
+
+COPY requirements.txt .
+
+# CPU-only PyTorch keeps the image smaller; sufficient for the default local LR model
+RUN pip install --no-cache-dir --upgrade pip \
+    && pip install --no-cache-dir torch --index-url https://download.pytorch.org/whl/cpu \
+    && pip install --no-cache-dir -r requirements.txt \
+    && python -m spacy download en_core_web_sm
+
+# NLTK corpora used by TextPreprocessor
+RUN python - <<'PY'
+import nltk
+for pkg in ("stopwords", "punkt"):
+    nltk.download(pkg, download_dir="/app/nltk_data")
+PY
+
+COPY configs/ configs/
+COPY src/ src/
+COPY models/final_model.joblib models/final_model.joblib
+
+# Default env template (overridden by docker-compose)
+COPY env.example .env.example
+
+EXPOSE 8000 8501

README.md

@@ -1,26 +1,77 @@
-# Project_9_Equipo3
+# YouTube Toxic Comment Detector (SignalMod)
 
-## 🏗️  Arquitectura
+Binary **Safe vs Toxic** comment moderation assistant with a **FastAPI** backend and a **Streamlit** UI.
+
+## Quick start (Docker)
+
+No manual setup beyond Docker. The image bundles the default model (`models/final_model.joblib`), configs, and NLP assets (spaCy + NLTK).
+
+```bash
+docker compose up --build
+```
+
+| Service   | URL |
+|-----------|-----|
+| Streamlit UI | http://localhost:8501 |
+| FastAPI      | http://localhost:8000 |
+| API docs     | http://localhost:8000/docs |
+
+Optional: set `YOUTUBE_API_KEY` for live comment scraping on `/predict-video`:
+
+```bash
+export YOUTUBE_API_KEY=your_key_here
+docker compose up --build
+```
+
+Stop containers:
+
+```bash
+docker compose down
+```
+
+Docker image and containers use the project name `youtube_hate_detector` (e.g. `youtube_hate_detector-api`). If you previously built `ai-nlp-app:latest`, remove it once: `docker rmi ai-nlp-app:latest`.
+
+## Architecture
 
 ```
 youtube_hate_detector/
+├── configs/           # YAML hyperparameters (non-secret)
 ├── data/
-│   ├── raw/                # Dataset original (no se sube al repo)
-│   └── processed/          # Datos preprocesados
-├── notebooks/              # EDA y experimentos
-├── models/                 # Modelos entrenados (.joblib)
-├── reports/                # Métricas y resultados por experimento
+│   ├── raw/           # Original dataset (gitignored)
+│   └── processed/
+├── models/            # Serialized models (e.g. final_model.joblib)
 ├── src/
-│   ├── data/               # Carga y scraping de datos
-│   ├── features/           # Preprocesamiento y vectorización
-│   ├── models/             # Baseline, ensemble, deep learning
-│   ├── evaluation/         # Métricas y análisis de errores
-│   ├── pipeline/           # Pipeline end-to-end
-│   ├── api/                # FastAPI
-│   ├── app/                # Streamlit
-│   └── utils/              # Logger, config loader
-├── configs/                # Configuración YAML
-├── tests/                  # Tests unitarios
-└── logs/                   # Logs de ejecución
-
-```
+│   ├── api/           # FastAPI (REST)
+│   ├── app/           # Streamlit UI
+│   ├── data/
+│   ├── features/
+│   ├── models/
+│   ├── pipeline/
+│   ├── service/       # ModelService (inference)
+│   └── utils/
+├── tests/
+├── Dockerfile
+└── docker-compose.yml
+```
+
+## Local development (without Docker)
+
+```bash
+python -m venv .venv && source .venv/bin/activate
+pip install -r requirements.txt
+python -m spacy download en_core_web_sm
+
+# Terminal 1 — API
+uvicorn src.api.main:app --reload --host 0.0.0.0 --port 8000
+
+# Terminal 2 — Streamlit
+streamlit run src/app/app.py --server.port 8501
+```
+
+Copy `env.example` to `.env` if you need a YouTube API key or custom `MODEL_NAME`.
+
+## Tests
+
+```bash
+pytest tests/ -v
+```

docker-compose.yml

@@ -0,0 +1,57 @@
+# youtube_hate_detector — API + Streamlit UI
+# Start everything:  docker compose up --build
+# Stop:             docker compose down
+
+name: youtube_hate_detector
+
+services:
+  api:
+    build: .
+    image: youtube_hate_detector:latest
+    container_name: youtube_hate_detector-api
+    command:
+      - uvicorn
+      - src.api.main:app
+      - --host
+      - "0.0.0.0"
+      - --port
+      - "8000"
+    ports:
+      - "8000:8000"
+    environment:
+      MODEL_NAME: "LR + TF-IDF (local)"
+      ENV: production
+      YOUTUBE_API_KEY: ${YOUTUBE_API_KEY:-}
+      NLTK_DATA: /app/nltk_data
+    healthcheck:
+      test: ["CMD", "curl", "-f", "http://localhost:8000/"]
+      interval: 10s
+      timeout: 5s
+      retries: 12
+      start_period: 40s
+    restart: unless-stopped
+
+  streamlit:
+    # Reuses the image built by `api` — do not add `build:` here (parallel builds race on the same tag)
+    image: youtube_hate_detector:latest
+    container_name: youtube_hate_detector-streamlit
+    command:
+      - streamlit
+      - run
+      - src/app/app.py
+      - --server.port=8501
+      - --server.address=0.0.0.0
+      - --server.headless=true
+      - --browser.gatherUsageStats=false
+    ports:
+      - "8501:8501"
+    environment:
+      MODEL_NAME: "LR + TF-IDF (local)"
+      ENV: production
+      API_URL: http://api:8000
+      YOUTUBE_API_KEY: ${YOUTUBE_API_KEY:-}
+      NLTK_DATA: /app/nltk_data
+    depends_on:
+      api:
+        condition: service_healthy
+    restart: unless-stopped

requirements.txt

@@ -0,0 +1,14 @@
+# Runtime dependencies for API + Streamlit (Docker and local installs)
+fastapi==0.136.1
+uvicorn[standard]==0.47.0
+streamlit>=1.41.0,<2
+scikit-learn==1.8.0
+spacy==3.8.14
+nltk==3.9.4
+pandas==3.0.2
+PyYAML==6.0.3
+python-dotenv==1.2.2
+joblib==1.5.3
+pydantic==2.13.4
+transformers==5.9.0
+httpx==0.28.1