Add ffmpeg cuda support and download nltk tokenizer (#27)

Dockerfile

@@ -31,9 +31,17 @@ RUN pip3 install --no-cache-dir -r requirements.txt
 # Setup scripts and execute them
 COPY scripts scripts
 RUN bash scripts/install_redis.sh && \
-    bash scripts/install_ffmpeg.sh && \
     bash scripts/install_punkt_tokenizers.sh
 
+# Install ffmpeg based on the architecture
+RUN if [ "${TARGETPLATFORM}" = "linux/arm64" ]; then \
+      bash /scripts/install_ffmpeg.sh; \
+    elif [ "${TARGETPLATFORM}" = "linux/amd64" ]; then \
+      bash /scripts/install_ffmpeg_cuda.sh; \
+    else \
+      echo "Unsupported platform: ${TARGETARCH}" && exit 1; \
+    fi
+
 # Install Vespa and pin the version. All versions can be found using `dns list vespa`
 # This is installed as a separate docker layer since we need to upgrade vespa regularly
 RUN dnf config-manager --add-repo https://raw.githubusercontent.com/vespa-engine/vespa/master/dist/vespa-engine.repo && \
@@ -48,3 +56,6 @@ ENV VESPA_LOG_STDOUT="true"
 ENV VESPA_LOG_FORMAT="vespa"
 ENV VESPA_CLI_HOME=/tmp/.vespa
 ENV VESPA_CLI_CACHE_DIR=/tmp/.cache/vespa
+ENV NVIDIA_DRIVER_CAPABILITIES=utility,compute,video
+# expose nltk data to all users
+ENV NLTK_DATA=/root/nltk_data

scripts/install_ffmpeg_cuda.sh

@@ -0,0 +1,80 @@
+#!/bin/bash
+
+set -euo pipefail
+set -x
+
+# Step 1: Install CUDA Toolkit
+# Add the NVIDIA repository for CUDA and install CUDA toolkit version 12.6
+dnf config-manager --add-repo https://developer.download.nvidia.com/compute/cuda/repos/rhel8/x86_64/cuda-rhel8.repo
+dnf clean all
+dnf -y install cuda-toolkit-12-6
+# Set CUDA environment variables for PATH and LD_LIBRARY_PATH
+export PATH=/usr/local/cuda/bin:${PATH}
+export LD_LIBRARY_PATH=/usr/local/cuda/lib64
+# Verify nvcc installation (CUDA compiler)
+nvcc --version
+# Step 2: Install dependencies required for FFmpeg
+# Install libraries and tools required for FFmpeg compilation
+dnf install -y libtool \
+    glibc \
+    glibc-devel \
+    numactl \
+    numactl-devel \
+    openssl \
+    yasm \
+    pkg-config \
+    openssl-devel \
+    git \
+    gcc \
+    make \
+    gcc-c++ \
+    kernel-headers \
+    automake
+# Step 3: Install x264
+# Add the RPM Fusion repository and install x264 and its development files
+dnf install -y https://download1.rpmfusion.org/free/el/rpmfusion-free-release-8.noarch.rpm
+dnf install -y x264 x264-devel
+# Step 4: Install NVENC codec headers
+# Install the necessary development tools and clone the nv-codec-headers repository
+# Clone the nv-codec-headers repository and install it
+git clone https://git.videolan.org/git/ffmpeg/nv-codec-headers.git
+cd nv-codec-headers
+git checkout 9934f17316b66ce6de12f3b82203a298bc9351d8 # Fix the version
+make
+make install
+cd ..
+# Set PKG_CONFIG_PATH for pkg-config to find the newly installed nv-codec-headers
+export PKG_CONFIG_PATH=/usr/local/lib/pkgconfig:$PKG_CONFIG_PATH
+# Add /usr/local/lib to the library search path and update the dynamic linker cache
+echo "/usr/local/lib" | tee -a /etc/ld.so.conf
+ldconfig
+# Step 5: Install FFmpeg
+# Clone the FFmpeg repository
+git clone https://git.ffmpeg.org/ffmpeg.git
+# Configure and compile FFmpeg with necessary flags for NVIDIA, x264, and other libraries
+cd ffmpeg
+git checkout faa366003b58ba26484070ca408be4b9d5473a73 # Fix the version
+./configure --enable-nonfree \
+    --enable-cuda-nvcc \
+    --enable-libnpp \
+    --enable-libx264 \
+    --enable-openssl \
+    --enable-nvenc \
+    --enable-gpl \
+    --extra-cflags=-I/usr/local/cuda/include \
+    --extra-ldflags=-L/usr/local/cuda/lib64 \
+    --disable-static \
+    --enable-shared
+# Compile and install FFmpeg
+make -j $(nproc)
+make install
+# Do some cleanup
+rm -rf /nv-codec-headers
+rm -rf /ffmpeg
+dnf remove -y \
+    make \
+    git \
+    automake
+dnf clean all
+set +x
+

scripts/install_punkt_tokenizers.sh

@@ -1,7 +1,14 @@
 #!/bin/bash
 # This script is meant to be run at buildtime.
+set -euo pipefail
+set -x
 
 mkdir -p /root/nltk_data/tokenizers
 curl https://raw.githubusercontent.com/nltk/nltk_data/gh-pages/packages/tokenizers/punkt.zip -o /root/nltk_data/tokenizers/punkt.zip
 unzip /root/nltk_data/tokenizers/punkt.zip -d /root/nltk_data/tokenizers/
-rm /root/nltk_data/tokenizers/punkt.zip
+rm /root/nltk_data/tokenizers/punkt.zip
+
+curl https://raw.githubusercontent.com/nltk/nltk_data/gh-pages/packages/tokenizers/punkt_tab.zip -o /root/nltk_data/tokenizers/punkt_tab.zip
+unzip /root/nltk_data/tokenizers/punkt_tab.zip -d /root/nltk_data/tokenizers/
+rm /root/nltk_data/tokenizers/punkt_tab.zip
+set +x