Fix Dockerfile for proper model loading and Ollama setup

Dockerfile

@@ -9,38 +9,58 @@ RUN useradd -m -u 1000 user
 
 WORKDIR /app
 
-COPY --chown=user ./requirements.txt requirements.txt
-RUN pip install --no-cache-dir --upgrade -r requirements.txt
-
 # Copy application files including templates
 COPY --chown=user . /app
 
-# Create startup script
-RUN echo '#!/bin/bash' > /app/start.sh && \
-    echo 'echo "Iniciando servidor Ollama..."' >> /app/start.sh && \
-    echo 'ollama serve &' >> /app/start.sh && \
-    echo 'echo "Esperando a que Ollama esté listo..."' >> /app/start.sh && \
-    echo 'timeout=60' >> /app/start.sh && \
-    echo 'while ! nc -z localhost 11434; do' >> /app/start.sh && \
-    echo '  if [ "$timeout" -le "0" ]; then' >> /app/start.sh && \
-    echo '    echo "Tiempo de espera agotado para Ollama"' >> /app/start.sh && \
-    echo '    exit 1' >> /app/start.sh && \
-    echo '  fi' >> /app/start.sh && \
-    echo '  echo "Esperando a Ollama... $timeout segundos restantes"' >> /app/start.sh && \
-    echo '  timeout=$((timeout-1))' >> /app/start.sh && \
-    echo '  sleep 1' >> /app/start.sh && \
-    echo 'done' >> /app/start.sh && \
-    echo 'echo "Descargando modelo desde Hugging Face..."' >> /app/start.sh && \
-    echo 'wget https://huggingface.co/andresdegante/llama3-papalia-nuevo/resolve/main/model.tar.gz' >> /app/start.sh && \
-    echo 'ollama import llama3-papalia-nuevo model.tar.gz' >> /app/start.sh && \
-    echo 'rm model.tar.gz' >> /app/start.sh && \
-    echo 'echo "Iniciando API..."' >> /app/start.sh && \
-    echo 'uvicorn app:app --host 0.0.0.0 --port 7860' >> /app/start.sh
+RUN pip install --no-cache-dir --upgrade -r requirements.txt
+
+# Create directories
+RUN mkdir -p /home/user/.ollama/models && \
+    chown -R user:user /home/user/.ollama
+
+# Create startup script with better error handling
+RUN echo '#!/bin/bash\n\
+set -e\n\
+\n\
+echo "Iniciando servidor Ollama..."\n\
+ollama serve &\n\
+\n\
+echo "Esperando a que Ollama esté disponible..."\n\
+for i in $(seq 1 60); do\n\
+    if nc -z localhost 11434; then\n\
+        echo "Ollama está listo"\n\
+        break\n\
+    fi\n\
+    echo "Esperando a Ollama... $i/60"\n\
+    sleep 1\n\
+done\n\
+\n\
+echo "Descargando modelo desde Hugging Face..."\n\
+cd /tmp\n\
+wget -q --show-progress https://huggingface.co/andresdegante/llama3-papalia-nuevo/resolve/main/model.tar.gz\n\
+\n\
+echo "Importando modelo a Ollama..."\n\
+tar xzf model.tar.gz -C /home/user/.ollama/models\n\
+rm model.tar.gz\n\
+\n\
+echo "Verificando modelo..."\n\
+if ! ollama list | grep -q "llama3-papalia-nuevo"; then\n\
+    echo "Error: Modelo no encontrado después de la importación"\n\
+    exit 1\n\
+fi\n\
+\n\
+echo "Iniciando API..."\n\
+cd /app\n\
+exec uvicorn app:app --host 0.0.0.0 --port 7860\n\
+' > /app/start.sh
 
 RUN chmod +x /app/start.sh
 
 USER user
 ENV PATH="/home/user/.local/bin:$PATH"
 
-# Start both Ollama and the FastAPI app
+# Health check
+HEALTHCHECK --interval=30s --timeout=10s --start-period=60s --retries=3 \
+    CMD curl -f http://localhost:7860/health || exit 1
+
 CMD ["/app/start.sh"]