Dockerfile new style

Dockerfile

@@ -1,84 +1,94 @@
-# Multi-stage build to get pgvector files
-FROM pgvector/pgvector:pg16 as pgvector
+# Start with PostgreSQL image with pgvector
+FROM pgvector/pgvector:pg16
 
-# Main image
-FROM pspdfkit/ai-assistant:nightly
+# Set environment variables
+ENV POSTGRES_USER=db-user \
+    POSTGRES_PASSWORD=password \
+    POSTGRES_DB=ai_assistant \
+    POSTGRES_INITDB_ARGS=--data-checksums \
+    PGDATA=/var/lib/postgresql/data/pgdata \
+    OPENAI_API_KEY="dumy-key" \
+    PGUSER=db-user \
+    PGPASSWORD=password \
+    PGDATABASE=ai_assistant \
+    PGHOST=localhost \
+    PGPORT=5432 \
+    API_AUTH_TOKEN=secret \
+    JWT_PUBLIC_KEY="-----BEGIN PUBLIC KEY-----\nMIIBIjANBgkqhkiG9w0BAQEFAAOCAQ8AMIIBCgKCAQEA2gzhmJ9TDanEzWdP1WG+\n0Ecwbe7f3bv6e5UUpvcT5q68IQJKP47AQdBAnSlFVi4X9SaurbWoXdS6jpmPpk24\nQvitzLNFphHdwjFBelTAOa6taZrSusoFvrtK9x5xsW4zzt/bkpUraNx82Z8MwLwr\nt6HlY7dgO9+xBAabj4t1d2t+0HS8O/ed3CB6T2lj6S8AbLDSEFc9ScO6Uc1XJlSo\nrgyJJSPCpNhSq3AubEZ1wMS1iEtgAzTPRDsQv50qWIbn634HLWxTP/UH6YNJBwzt\n3O6q29kTtjXlMGXCvin37PyX4Jy1IiPFwJm45aWJGKSfVGMDojTJbuUtM+8P9Rrn\nAwIDAQAB\n-----END PUBLIC KEY-----" \
+    JWT_ALGORITHM=RS256 \
+    DASHBOARD_USERNAME=dashboard \
+    DASHBOARD_PASSWORD=secret \
+    SECRET_KEY_BASE=secret-key-base
 
-# Install PostgreSQL 17
-RUN echo "https://dl-cdn.alpinelinux.org/alpine/edge/main" >> /etc/apk/repositories && \
-    echo "https://dl-cdn.alpinelinux.org/alpine/edge/community" >> /etc/apk/repositories && \
-    apk update && \
-    apk add --no-cache \
-    postgresql17 \
-    postgresql17-contrib \
-    postgresql17-dev
+# Install required packages including nginx
+RUN apt-get update && \
+    apt-get install -y nodejs npm python3-venv python3-pip wget curl unzip nginx-light procps && \
+    mkdir -p /tmp/llamafile && \
+    cd /tmp/llamafile && \
+    wget --no-check-certificate https://github.com/Mozilla-Ocho/llamafile/releases/download/0.8.17/llamafile-0.8.17.zip && \
+    unzip llamafile-0.8.17.zip && \
+    cp llamafile-0.8.17/bin/llamafiler /usr/local/bin/ && \
+    chmod +x /usr/local/bin/llamafiler && \
+    rm -rf /tmp/llamafile && \
+    # Create all required nginx directories and files with correct ownership and permissions
+    mkdir -p /var/run/nginx && \
+    mkdir -p /var/lib/nginx/body && \
+    mkdir -p /var/lib/nginx/proxy && \
+    mkdir -p /var/lib/nginx/fastcgi && \
+    mkdir -p /var/lib/nginx/uwsgi && \
+    mkdir -p /var/lib/nginx/scgi && \
+    mkdir -p /var/log/nginx && \
+    touch /var/log/nginx/access.log && \
+    touch /var/log/nginx/error.log && \
+    chmod 755 /var/lib && \
+    chmod 755 /var/log && \
+    chmod 755 /var/run && \
+    chown -R postgres:postgres /var/run/nginx /var/log/nginx /var/lib/nginx && \
+    chmod -R 755 /var/lib/nginx /var/log/nginx && \
+    chmod 777 /var/run/nginx && \
+    chmod 666 /var/log/nginx/access.log && \
+    chmod 666 /var/log/nginx/error.log
 
-# Get PostgreSQL lib directory
-RUN pg_config --pkglibdir > /tmp/pg_libdir
+# Create and activate virtual environment
+RUN python3 -m venv /venv
+ENV PATH="/venv/bin:$PATH"
 
-# Create directories
-RUN mkdir -p $(cat /tmp/pg_libdir) && \
-    mkdir -p /usr/share/postgresql17/extension
+# Install Flask in the virtual environment
+RUN /venv/bin/pip install flask
 
-# Copy files from pgvector image
-COPY --from=pgvector /usr/lib/postgresql/16/lib/vector.so $(cat /tmp/pg_libdir)/
-COPY --from=pgvector /usr/share/postgresql/16/extension/vector--0.8.0.sql /usr/share/postgresql17/extension/
-COPY --from=pgvector /usr/share/postgresql/16/extension/vector.control /usr/share/postgresql17/extension/
+# Copy the Flask app
+COPY app.py /app.py
+COPY service-config.yml /service-config.yml
 
-# Create a modified vector.control file with the correct version
-RUN echo "# vector extension" > /usr/share/postgresql17/extension/vector.control && \
-    echo "comment = 'vector data type and ivfflat and hnsw access methods'" >> /usr/share/postgresql17/extension/vector.control && \
-    echo "default_version = '0.8.0'" >> /usr/share/postgresql17/extension/vector.control && \
-    echo "module_pathname = 'vector'" >> /usr/share/postgresql17/extension/vector.control && \
-    echo "relocatable = false" >> /usr/share/postgresql17/extension/vector.control
+# Create a healthcheck script
+COPY healthcheck.sh /usr/local/bin/healthcheck.sh
+RUN chmod +x /usr/local/bin/healthcheck.sh
 
-# Environment variables
-ENV PGUSER=node \
-    PGPASSWORD=password \
-    PGDATABASE=ai_assistant \
-    PGHOST=localhost \
-    PGPORT=5432 \
-    POSTGRES_USER=node \
-    POSTGRES_PASSWORD=password \
-    POSTGRES_DB=ai_assistant \
-    PGDATA=/tmp/postgresql
+# Pull and extract the AI assistant image
+COPY --from=pspdfkit/ai-assistant:nightly / /ai-assistant
+COPY service-config.yml /ai-assistant/base/config/service-config.yml
+
+# Copy and prepare start script
+COPY start-services.sh /start-services.sh
+RUN chmod +x /start-services.sh
+
+# Copy Nginx configuration
+COPY nginx.conf /etc/nginx/nginx.conf
 
-# Create necessary directories and set permissions
-RUN mkdir -p /run/postgresql && \
-    chown node:node /run/postgresql && \
-    chmod 777 /run/postgresql && \
-    chown -R node:node /usr/share/postgresql17 && \
-    chown -R node:node $(cat /tmp/pg_libdir) && \
-    chmod 755 $(cat /tmp/pg_libdir)/vector.so
+# Create base directory and download models during build
+RUN mkdir -p /base && \
+    wget -q https://huggingface.co/bartowski/gemma-2-2b-it-GGUF/resolve/main/gemma-2-2b-it-Q4_K_M.gguf -O /base/gemma-2b.gguf && \
+    wget -q https://huggingface.co/leliuga/all-MiniLM-L6-v2-GGUF/resolve/main/all-MiniLM-L6-v2.F16.gguf -O /base/embeddings.gguf && \
+    chown -R postgres:postgres /base
 
-# Debug: Show all relevant paths and files
-RUN echo "PostgreSQL paths:" && \
-    pg_config --pkglibdir && \
-    echo "Files in library directory:" && \
-    ls -la $(cat /tmp/pg_libdir) | grep vector && \
-    echo "Files in extension directory:" && \
-    ls -la /usr/share/postgresql17/extension/ | grep vector
+# Expose ports
+EXPOSE 5432 7860 4000 8080 8081 8082
 
-# Create a wrapper script
-RUN echo '#!/bin/sh' > /wrapper.sh && \
-    echo 'set -e' >> /wrapper.sh && \
-    echo 'mkdir -p $PGDATA' >> /wrapper.sh && \
-    echo 'if [ ! -s "$PGDATA/PG_VERSION" ]; then' >> /wrapper.sh && \
-    echo '    initdb -D $PGDATA' >> /wrapper.sh && \
-    echo '    echo "host all all all trust" >> $PGDATA/pg_hba.conf' >> /wrapper.sh && \
-    echo '    echo "listen_addresses = '"'"'*'"'"'" >> $PGDATA/postgresql.conf' >> /wrapper.sh && \
-    echo '    echo "dynamic_library_path = '"'"'\$libdir'"'"'" >> $PGDATA/postgresql.conf' >> /wrapper.sh && \
-    echo 'fi' >> /wrapper.sh && \
-    echo 'pg_ctl -D $PGDATA -o "-c listen_addresses=*" -w start' >> /wrapper.sh && \
-    echo 'createdb $POSTGRES_DB || true' >> /wrapper.sh && \
-    echo 'psql -d $POSTGRES_DB -c "SELECT version();"' >> /wrapper.sh && \
-    echo 'psql -d $POSTGRES_DB -c "SHOW dynamic_library_path;"' >> /wrapper.sh && \
-    echo 'psql -d $POSTGRES_DB -c "CREATE EXTENSION IF NOT EXISTS vector;"' >> /wrapper.sh && \
-    echo 'exec /start-services.sh' >> /wrapper.sh && \
-    chmod +x /wrapper.sh
+# Add a healthcheck
+HEALTHCHECK --interval=3s --timeout=3s --retries=10 \
+  CMD ["healthcheck.sh"]
 
-# Expose necessary ports
-EXPOSE 3000 5432 8080
+USER postgres
 
-# Set the entrypoint
-ENTRYPOINT ["/wrapper.sh"]
+# Start PostgreSQL, Flask, and AI assistant
+CMD ["sh", "-c", "/venv/bin/flask run --host=0.0.0.0 --port=7860 & docker-entrypoint.sh postgres & /start-services.sh"]

app.py

@@ -0,0 +1,10 @@
+from flask import Flask
+
+app = Flask(__name__)
+
+@app.route('/')
+def home():
+    return "PostgreSQL is running with Flask!"
+
+if __name__ == '__main__':
+    app.run(host='0.0.0.0', port=7860)

healthcheck.sh

@@ -0,0 +1,2 @@
+#!/bin/bash
+pg_isready -U db-user -d ai_assistant

nginx.conf

@@ -0,0 +1,30 @@
+pid /var/run/nginx/nginx.pid;  
+
+events {
+    worker_connections 1024;
+}
+
+http {
+    include       /etc/nginx/mime.types;
+    default_type  application/octet-stream;
+
+    server {
+        listen 8080;
+        
+        location /v1/embeddings {
+            proxy_pass http://127.0.0.1:8081;
+            proxy_set_header Host $host;
+            proxy_set_header X-Real-IP $remote_addr;
+            proxy_read_timeout 300;
+            proxy_connect_timeout 300;
+        }
+
+        location /v1 {
+            proxy_pass http://127.0.0.1:8082;
+            proxy_set_header Host $host;
+            proxy_set_header X-Real-IP $remote_addr;
+            proxy_read_timeout 300;
+            proxy_connect_timeout 300;
+        }
+    }
+}

service-config.yml

@@ -4,17 +4,17 @@ aiServices:
   chat:
     provider:
       name: 'openai-compat'
-      baseUrl: http://localhost:8080/v1
+      baseUrl: http://127.0.0.1:8080/v1
     model: 'gemma-2b'
   textEmbeddings:
     provider:
       name: 'openai-compat'
-      baseUrl: http://localhost:8080/v1
+      baseUrl: http://127.0.0.1:8080/v1
     model: 'all-MiniLM-L6-v2'
   headless:
     - provider:
         name: 'openai-compat'
-        baseUrl: http://localhost:8080/v1
+        baseUrl: http://127.0.0.1:8080/v1
       model:
         name: 'gemma-2b'
         id: 'gemma-2b'

start-container.sh

@@ -0,0 +1,28 @@
+#!/bin/sh
+
+# Start nginx
+nginx
+
+# Start llamafiler instances
+/usr/local/bin/llamafiler --model /base/gemma-2b.gguf --listen 0.0.0.0:8082 &
+GEMMA_PID=$!
+
+/usr/local/bin/llamafiler --model /base/embeddings.gguf --listen 0.0.0.0:8081 &
+EMBEDDINGS_PID=$!
+
+# Wait for services
+until curl -s --fail -X POST http://localhost:8080/v1/chat/completions \
+    -H "Content-Type: application/json" \
+    -d '{"model": "gemma-2b", "messages":[{"role":"user","content":"hi"}]}' >/dev/null 2>&1 && \
+    curl -s --fail -X POST http://localhost:8080/v1/embeddings \
+    -H "Content-Type: application/json" \
+    -d '{"input":"test"}' >/dev/null 2>&1
+do
+    sleep 1
+    echo "Waiting for services..."
+done
+
+echo 'Services are ready!'
+
+# Start the main application
+exec ./bin/entrypoint.sh node app/main.bundle.js

start-services.sh

@@ -1,136 +1,72 @@
 #!/bin/sh
 
-# Create directories if they don't exist
-mkdir -p /data/models
-mkdir -p /var/log/llamafiler
-
-# Install required packages if not already installed
-if ! command -v wget >/dev/null 2>&1; then
-    echo "Installing dependencies..."
-    apk add --no-cache wget curl nginx \
-        build-base \
-        gcc \
-        musl-dev \
-        linux-headers \
-        python3 \
-        py3-pip \
-        git \
-        make
-fi
+# Wait for PostgreSQL to be ready
+until pg_isready -h localhost -U db-user
+do
+    echo "Waiting for PostgreSQL to be ready..."
+    sleep 1
+done
 
-# Check if models exist in persistent storage, download only if needed
-if [ ! -f "/data/models/gemma-2b.gguf" ]; then
-    echo "Downloading Gemma model..."
-    wget -q https://huggingface.co/bartowski/gemma-2-2b-it-GGUF/resolve/main/gemma-2-2b-it-Q4_K_M.gguf -O /data/models/gemma-2b.gguf
+# Start nginx with explicit configuration
+echo "Starting nginx..."
+/usr/sbin/nginx -c /etc/nginx/nginx.conf
+if [ $? -ne 0 ]; then
+    echo "Failed to start nginx"
+    exit 1
 fi
 
-if [ ! -f "/data/models/embeddings.gguf" ]; then
-    echo "Downloading embeddings model..."
-    wget -q https://huggingface.co/leliuga/all-MiniLM-L6-v2-GGUF/resolve/main/all-MiniLM-L6-v2.F16.gguf -O /data/models/embeddings.gguf
+# Verify nginx is running
+if ! ps aux | grep nginx | grep -v grep > /dev/null; then
+    echo "Nginx failed to start"
+    exit 1
 fi
 
-# Install llamafiler only if not already installed
-if [ ! -f "/usr/local/bin/llamafiler" ]; then
-    echo "Building llamafiler from source..."
-    mkdir -p /tmp/llamafile
-    cd /tmp/llamafile
-    wget --no-check-certificate https://github.com/Mozilla-Ocho/llamafile/releases/download/0.8.17/llamafile-0.8.17.zip
-    unzip llamafile-0.8.17.zip
-
-    if [ -f llamafile-0.8.17/bin/llamafiler ]; then
-        echo "Installing llamafiler..."
-        cp llamafile-0.8.17/bin/llamafiler /usr/local/bin/
-        chmod +x /usr/local/bin/llamafiler
-    else
-        echo "ERROR: Llamafiler binary not found!"
-        exit 1
-    fi
-    cd - # Return to previous directory
-    rm -rf /tmp/llamafile
-fi
+echo "Nginx started successfully"
 
-# Setup nginx reverse proxy
-cat > /etc/nginx/nginx.conf <<EOF
-events { worker_connections 1024; }
-http {
-    server {
-        listen 8080;
-        location /v1/embeddings {
-            proxy_pass http://localhost:8081;
-            proxy_read_timeout 300;
-            proxy_connect_timeout 300;
-            proxy_send_timeout 300;
-        }
-        location /v1 {
-            proxy_pass http://localhost:8082;
-            proxy_read_timeout 300;
-            proxy_connect_timeout 300;
-            proxy_send_timeout 300;
-        }
-    }
-}
-EOF
+# Start llamafiler instances
+echo "Starting llamafiler services..."
 
-# Start nginx
-echo "Starting nginx..."
-nginx
-echo "Nginx started"
+# Create TMPDIR if it doesn't exist
+mkdir -p /tmp/llamafiler
 
-# Start llamafiler instances with logging
-echo "Starting llamafiler for Gemma model..."
-/usr/local/bin/llamafiler --model /data/models/gemma-2b.gguf --listen 0.0.0.0:8082 &
+echo "Starting chat model..."
+TMPDIR=/tmp/llamafiler /usr/local/bin/llamafiler --model /base/gemma-2b.gguf --listen 0.0.0.0:8082 &
 GEMMA_PID=$!
-echo "Gemma PID: $GEMMA_PID"
 
-echo "Starting llamafiler for embeddings model..."
-/usr/local/bin/llamafiler --model /data/models/embeddings.gguf --listen 0.0.0.0:8081 &
+echo "Starting embedding model..."
+TMPDIR=/tmp/llamafiler /usr/local/bin/llamafiler --model /base/embeddings.gguf --listen 0.0.0.0:8081 &
 EMBEDDINGS_PID=$!
-echo "Embeddings PID: $EMBEDDINGS_PID"
-
-# More robust health check
-echo 'Waiting for servers to start...'
-TIMEOUT=300  # 5 minutes timeout
-START_TIME=$(date +%s)
 
-while true; do
-    CURRENT_TIME=$(date +%s)
-    ELAPSED_TIME=$((CURRENT_TIME - START_TIME))
-    
-    if [ $ELAPSED_TIME -gt $TIMEOUT ]; then
-        echo "Timeout waiting for services to start"
-        exit 1
-    fi
+# Wait for the models to start loading
+echo "Waiting for initial model loading..."
+sleep 15  # Give more time for the models to load into memory
 
-    # Check if processes are still running
-    if ! kill -0 $GEMMA_PID 2>/dev/null; then
-        echo "Gemma model process died."    
-        exit 1
-    fi
-    if ! kill -0 $EMBEDDINGS_PID 2>/dev/null; then
-        echo "Embeddings model process died."
-        exit 1
-    fi
-
-    # Try to contact both services
-    if curl -s --fail -X POST http://localhost:8080/v1/chat/completions \
-        -H "Content-Type: application/json" \
-        -d '{"model": "gemma-2b", "messages":[{"role":"user","content":"hi"}]}' >/dev/null 2>&1 && \
-        curl -s --fail -X POST http://localhost:8080/v1/embeddings \
-        -H "Content-Type: application/json" \
-        -d '{"input":"test"}' >/dev/null 2>&1; then
-        break  # Both services are responding
-    fi
-    
-    sleep 5
-    echo "Waiting for services... (Elapsed time: ${ELAPSED_TIME}s)"
+echo 'Waiting for servers to start...'
+until curl -s --fail -X POST http://localhost:8080/v1/chat/completions \
+    -H "Content-Type: application/json" \
+    -d '{"model": "gemma-2b", "messages":[{"role":"user","content":"hi"}]}' >/dev/null 2>&1 && \
+      curl -s --fail -X POST http://localhost:8080/v1/embeddings \
+    -H "Content-Type: application/json" \
+    -d '{"input":"test"}' >/dev/null 2>&1
+do
+  # Check if processes are still running
+  if ! kill -0 $GEMMA_PID 2>/dev/null; then
+    echo "Gemma model process died."    
+    exit 1
+  fi
+  if ! kill -0 $EMBEDDINGS_PID 2>/dev/null; then
+    echo "Embeddings model process died."
+    exit 1
+  fi
+  
+  sleep 2
+  echo "Waiting for services... (PIDs: Gemma=$GEMMA_PID, Embeddings=$EMBEDDINGS_PID)"
 done
 
 echo 'Servers are ready!'
 
-# Start the main application
-if [ -f "./bin/entrypoint.sh" ]; then
-    exec ./bin/entrypoint.sh node app/main.bundle.js
-else
-    echo "Warning: entrypoint.sh not found, starting node directly"
-    exec node app/main.bundle.js
-fi
+# Start the AI assistant application from the correct directory
+cd /ai-assistant/base && \
+NODE_PATH=/ai-assistant/base/node_modules \
+PATH=/ai-assistant/base/node_modules/.bin:/ai-assistant/base/bin:$PATH \
+     exec /ai-assistant/base/bin/entrypoint.sh node app/main.bundle.js