volknode 0.2.5__tar.gz

volknode-0.2.5/.gitignore

@@ -0,0 +1,21 @@
+node_modules
+*.env
+ai-suite/llm-node/.venv
+!*.example.env
+!devcontainer.config.env
+temp
+ai-suite/nexus/src/temp/*
+!ai-suite/nexus/src/temp
+!ai-suite/nexus/src/temp/temp
+
+ai-suite/bin
+secrets
+inventory.ini
+
+.claude/
+dist/
+gcloud_admin_service_account_key.json
+
+__pycache__/
+
+.pytest_cache

volknode-0.2.5/.netlify/netlify.toml

@@ -0,0 +1,26 @@
+plugins = []
+headers = []
+redirects = []
+
+[functions]
+
+[functions."*"]
+
+[build]
+publish = "/monorepo/ai-suite/llm-node"
+publishOrigin = "default"
+base = "/monorepo/ai-suite/llm-node"
+
+[build.environment]
+
+[build.processing]
+
+[build.processing.css]
+
+[build.processing.html]
+
+[build.processing.images]
+
+[build.processing.js]
+
+[build.services]

volknode-0.2.5/.python-version

@@ -0,0 +1 @@
+3.12

volknode-0.2.5/PKG-INFO

@@ -0,0 +1,11 @@
+Metadata-Version: 2.4
+Name: volknode
+Version: 0.2.5
+Summary: Add your description here
+Requires-Python: >=3.12
+Requires-Dist: attrs>=22.2.0
+Requires-Dist: cryptography>=44.0.0
+Requires-Dist: httpx<0.29.0,>=0.23.0
+Requires-Dist: huggingface-hub<1.0,>=0.34.0
+Requires-Dist: python-dateutil>=2.8.0
+Requires-Dist: typer>=0.20.0

volknode-0.2.5/Taskfile.yml

@@ -0,0 +1,44 @@
+version: "3"
+vars:
+  GITLAB_PROJECT_ID:
+    sh: cat ../../global-configuration.json | jq -r ".gitlabProjectId"
+
+tasks:
+  test:nexus:integration:
+    desc: |
+      Run integration tests against the nexus backend.
+      Requires the nexus backend to be running (use task ai-suite:nexus:containers:keep-up-with-fake-llm-real-ca).
+    dir: ./
+    cmds:
+      - uv run pytest volknode/src/nexus/ -v
+
+  test:unit:
+    desc: Run unit tests
+    dir: ./
+    cmds:
+      - uv run pytest volknode/src/usecases/ -v
+
+  volknode-with-fake-llm:
+    desc: Run volknode with a fake LLM runner
+    dir: ./
+    cmds:
+      - USE_FAKE_LLM=true uv run python -m volknode.main {{.CLI_ARGS}}
+
+  volknode-with-fake-llm-and-nexus:
+    desc: Run volknode with a fake LLM runner
+    dir: ./
+    cmds:
+      - USE_FAKE_LLM=true USE_FAKE_NEXUS=true uv run python -m volknode.main {{.CLI_ARGS}}
+
+  volknode:
+    desc: Run volknode
+    dir: ./
+    cmds:
+      - USE_FAKE_LLM=false uv run python -m volknode.main {{.CLI_ARGS}}
+
+  build:
+    desc: Build volknode executable with PyInstaller
+    dir: ./
+    cmds:
+      - mkdir -p ./installsh/dist
+      - cp ./installsh/install.sh ./installsh/dist/

volknode-0.2.5/gitlab/ci.yml

@@ -0,0 +1,121 @@
+stages:
+  - commit-stage
+  - deploy-to-test
+
+variables:
+  ENVIRONMENT: CI
+
+llm-node-nexus-integration-tests:
+  stage: commit-stage
+  image:
+    name: gitlab.internal.volksai.io:5050/volksai/monorepo/generic-ci-image:latest
+    entrypoint: [""]
+    pull_policy: if-not-present
+  variables:
+    TESTCONTAINERS_HOST_OVERRIDE: host.docker.internal
+    TESTCONTAINERS_RYUK_DISABLED: "true"
+    NEXUS_HOST: host.docker.internal
+    NEXUS_CONTAINER_NAME: nexus-$CI_JOB_ID
+    LOCAL_ENV_NEXUS_DB_HOST_INTERNALLY_DOCKER: dev-postgres-$CI_JOB_ID
+    LOCAL_ENV_NEXUS_DB_PORT_INTERNALLY_DOCKER: "5432"
+    LOCAL_ENV_NEXUS_DB_USER: "username"
+    LOCAL_ENV_NEXUS_DB_PASSWORD: "password"
+    LOCAL_ENV_NEXUS_DB_NAME: "database"
+    POSTGRES_VERSION: "17.4"
+  rules:
+    - if: '$RUN_ALL_TESTS == "true"'
+    - if: '$CI_PIPELINE_SOURCE == "push"'
+      changes:
+        - ai-suite/llm-node/**/*
+        - ai-suite/nexus/**/*
+        - ai-suite/openapi/**/*
+  tags:
+    - docker-executor
+  before_script:
+    - ln -sfn $CI_PROJECT_DIR /monorepo
+    - cd $CI_PROJECT_DIR/ai-suite/nexus/local-development/devcontainers && go mod download
+    - cd $CI_PROJECT_DIR
+    - cd $CI_PROJECT_DIR/ai-suite/llm-node && uv sync & UV_SYNC_PID=$!
+    - nohup task ai-suite:nexus:containers:keep-up-with-fake-llm-real-ca > backend-dev.log 2>&1 &
+    - $CI_PROJECT_DIR/scripts/wait-for-container.sh $NEXUS_CONTAINER_NAME
+    - nexus_port=$(docker inspect --format='{{(index (index .NetworkSettings.Ports "8080/tcp") 0).HostPort}}' $NEXUS_CONTAINER_NAME)
+    - $CI_PROJECT_DIR/scripts/wait-for-http-status-200.sh http://host.docker.internal:$nexus_port/chat-backend/is-live
+    - echo "NEXUS_PORT=$nexus_port" >> $CI_PROJECT_DIR/nexus.env
+    - wait $UV_SYNC_PID
+  script:
+    - source $CI_PROJECT_DIR/nexus.env
+    - NEXUS_PORT=$NEXUS_PORT task ai-suite:llm-node:test:nexus:integration
+  after_script:
+    - echo "--- backend-dev.log ---" && cat $CI_PROJECT_DIR/backend-dev.log 2>/dev/null || true
+    - echo "--- docker ps -a ---" && docker ps -a 2>/dev/null || true
+    - echo "Cleaning up job containers"
+    - docker stop $NEXUS_CONTAINER_NAME $LOCAL_ENV_NEXUS_DB_HOST_INTERNALLY_DOCKER 2>/dev/null || true
+    - docker rm $NEXUS_CONTAINER_NAME $LOCAL_ENV_NEXUS_DB_HOST_INTERNALLY_DOCKER 2>/dev/null || true
+    - docker network prune -f 2>/dev/null || true
+
+deploy-volksnode-to-pypi:
+  stage: deploy-to-test
+  needs: []
+  allow_failure: true
+  retry: 2
+  image:
+    name: gitlab.internal.volksai.io:5050/volksai/monorepo/generic-ci-image:latest
+    entrypoint: [""]
+    pull_policy: if-not-present
+  rules:
+    - if: '$RUN_ALL_TESTS == "true"'
+    - if: '$CI_PIPELINE_SOURCE == "push"'
+      changes:
+        - ai-suite/llm-node/**/*
+  tags:
+    - docker-executor
+  variables:
+    UV_PUBLISH_TOKEN: $PYPI_TOKEN
+  script:
+    - cd $CI_PROJECT_DIR/ai-suite/llm-node
+    - uv build
+    - uv publish
+
+test-llm-node-installsh:
+  stage: slow-tests
+  resource_group: test-llm-node-installsh
+  needs: []
+  image:
+    name: gitlab.internal.volksai.io:5050/volksai/monorepo/generic-ci-image:latest
+    entrypoint: [""]
+    pull_policy: if-not-present
+  rules:
+    - if: '$RUN_ALL_TESTS == "true"'
+    - if: '$CI_PIPELINE_SOURCE == "push"'
+      changes:
+        - ai-suite/llm-node/**/*
+  tags:
+    - docker-executor
+  variables:
+    NETLIFY_AUTH_TOKEN: $NETLIFY_AUTH_TOKEN
+    NETLIFY_LLM_NODE_INSTALLSH_TEST_SITE_ID: $NETLIFY_LLM_NODE_INSTALLSH_TEST_SITE_ID
+    GCLOUD_SERVICE_ACCOUNT_KEY: $GCLOUD_SERVICE_ACCOUNT_KEY
+  script:
+    - $CI_PROJECT_DIR/gitlab/scripts/configure-resource-group-newest-first.sh test-llm-node-installsh
+    - chmod +x $CI_PROJECT_DIR/ai-suite/llm-node/gitlab/scripts/test-installsh.sh
+    - $CI_PROJECT_DIR/ai-suite/llm-node/gitlab/scripts/test-installsh.sh
+
+llm-node-vllm-integration-tests:
+  stage: slow-tests
+  resource_group: llm-node-vllm-integration-tests
+  needs: []
+  image:
+    name: gitlab.internal.volksai.io:5050/volksai/monorepo/generic-ci-image:latest
+    entrypoint: [""]
+    pull_policy: if-not-present
+  rules:
+    - if: '$RUN_ALL_TESTS == "true"'
+    - if: '$CI_PIPELINE_SOURCE == "push"'
+      changes:
+        - ai-suite/llm-node/**/*
+  tags:
+    - docker-executor
+  script:
+    - $CI_PROJECT_DIR/gitlab/scripts/configure-resource-group-newest-first.sh llm-node-vllm-integration-tests
+    - chmod +x $CI_PROJECT_DIR/ai-suite/llm-node/gitlab/scripts/run-tests-on-gcloud.sh
+    - $CI_PROJECT_DIR/ai-suite/llm-node/gitlab/scripts/run-tests-on-gcloud.sh

volknode-0.2.5/gitlab/scripts/run-tests-on-gcloud.sh

@@ -0,0 +1,252 @@
+#!/bin/bash
+set -e
+
+# GCLOUD_PROJECT=${GCLOUD_PROJECT:?"\$GCLOUD_PROJECT must be set"}
+GCLOUD_PROJECT=${GCLOUD_PROJECT:-"llm-node-testing-479012-f6"}
+GCLOUD_ZONES=${GCLOUD_ZONES:-"us-central1-a us-central1-b us-central1-c us-central1-f us-east1-b us-east1-c us-east4-a us-east4-b us-east4-c us-west1-a us-west1-b us-west4-a us-west4-b europe-west1-b europe-west1-c europe-west4-a europe-west4-b europe-west4-c"}
+GCLOUD_MACHINE_TYPE=${GCLOUD_MACHINE_TYPE:-"g2-standard-8"}
+GCLOUD_GPU_TYPE=${GCLOUD_GPU_TYPE:-"nvidia-l4"}
+GCLOUD_GPU_COUNT=${GCLOUD_GPU_COUNT:-1}
+GCLOUD_DISK_SIZE_GB=${GCLOUD_DISK_SIZE_GB:-200}
+GCLOUD_IMAGE_FAMILY=${GCLOUD_IMAGE_FAMILY:-"ubuntu-2204-lts"}
+GCLOUD_IMAGE_PROJECT=${GCLOUD_IMAGE_PROJECT:-"ubuntu-os-cloud"}
+
+GCLOUD_SERVICE_ACCOUNT_KEY=${GCLOUD_SERVICE_ACCOUNT_KEY:-""}
+
+SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+REPO_ROOT="${CI_PROJECT_DIR:-$(cd "$SCRIPT_DIR/../../../.." && pwd)}"
+
+# Authenticate with gcloud
+echo "=== Authenticating with gcloud ==="
+if [[ -n "$GCLOUD_SERVICE_ACCOUNT_KEY" ]]; then
+    echo "$GCLOUD_SERVICE_ACCOUNT_KEY" > /tmp/gcloud-sa-key.json
+    gcloud auth activate-service-account --key-file=/tmp/gcloud-sa-key.json
+    rm -f /tmp/gcloud-sa-key.json
+else
+    echo "Warning: \$GCLOUD_SERVICE_ACCOUNT_KEY not set, assuming gcloud is already authenticated"
+fi
+gcloud config set project "$GCLOUD_PROJECT" --quiet
+
+instance_name=""
+GCLOUD_ZONE=""
+
+cleanup() {
+    echo "=== Cleaning up ==="
+    if [[ -n "$instance_name" && -n "$GCLOUD_ZONE" ]]; then
+        echo "Deleting GCP instance: $instance_name (zone: $GCLOUD_ZONE)"
+        gcloud compute instances delete "$instance_name" \
+            --project="$GCLOUD_PROJECT" \
+            --zone="$GCLOUD_ZONE" \
+            --quiet || true
+        echo "Instance deleted"
+    fi
+    if [[ -n "$gpu_instance_provision_tmp_dir" ]]; then
+        rm -rf "$gpu_instance_provision_tmp_dir" || true
+    fi
+}
+
+trap cleanup EXIT
+
+random_unique_tmp_dir_name=$(tr -dc 'a-zA-Z' </dev/urandom | head -c 16)
+gpu_instance_provision_tmp_dir=/tmp/vllm-integration-tests/$random_unique_tmp_dir_name
+gpu_instance_user_private_key_path=$gpu_instance_provision_tmp_dir/gpu_instance_key
+gpu_instance_user_public_key_path=$gpu_instance_provision_tmp_dir/gpu_instance_key.pub
+
+rm -rf $gpu_instance_provision_tmp_dir
+mkdir -p $gpu_instance_provision_tmp_dir
+
+# Generate SSH key pair
+ssh-keygen -b 2048 -t rsa -f $gpu_instance_user_private_key_path -q -N ""
+llmnode_runner_user_public_key=$(cat $gpu_instance_user_public_key_path)
+chmod 600 $gpu_instance_user_private_key_path
+
+instance_name="vllm-integration-test-$(tr -dc 'a-z0-9' </dev/urandom | head -c 8)"
+
+# Create startup script that installs NVIDIA drivers, CUDA, and sets up the user.
+# GCP startup scripts run on every boot. After installing NVIDIA drivers, a reboot
+# is required to load the kernel module. The script uses a marker file to track
+# whether the driver install + reboot has already happened.
+startup_script=$(cat <<'STARTUP'
+#!/bin/bash
+exec > /var/log/startup-script.log 2>&1
+set -x
+
+# Phase 1: Install NVIDIA drivers and reboot
+if [[ ! -f /var/log/nvidia-driver-installed ]]; then
+    apt-get update
+    apt-get install -y build-essential linux-headers-$(uname -r) openssh-server sudo curl python3 python3-pip python3-venv
+
+    curl -fsSL https://nvidia.github.io/libnvidia-container/gpgkey | gpg --dearmor -o /usr/share/keyrings/nvidia-container-toolkit-keyring.gpg
+    distribution=$(. /etc/os-release; echo $ID$VERSION_ID)
+    curl -sL "https://nvidia.github.io/libnvidia-container/$distribution/libnvidia-container.list" | \
+        sed 's#deb https://#deb [signed-by=/usr/share/keyrings/nvidia-container-toolkit-keyring.gpg] https://#g' | \
+        tee /etc/apt/sources.list.d/nvidia-container-toolkit.list
+
+    apt-get update
+    apt-get install -y nvidia-driver-550
+    apt-get install -y cuda-toolkit-12-8
+
+    echo 'export PATH=/usr/local/cuda/bin:$PATH' >> /etc/profile.d/cuda.sh
+    echo 'export LD_LIBRARY_PATH=/usr/local/cuda/lib64:$LD_LIBRARY_PATH' >> /etc/profile.d/cuda.sh
+
+    touch /var/log/nvidia-driver-installed
+    echo "NVIDIA driver installed, rebooting to load kernel module..."
+    reboot
+    exit 0
+fi
+
+# Phase 2: After reboot — set up user and signal readiness
+if ! id llmnode-runner &>/dev/null; then
+    useradd -m -s /bin/bash llmnode-runner
+fi
+echo 'llmnode-runner ALL=(ALL) NOPASSWD:ALL' >> /etc/sudoers
+
+mkdir -p /home/llmnode-runner/.ssh
+echo 'PUBLIC_KEY_PLACEHOLDER' >> /home/llmnode-runner/.ssh/authorized_keys
+chmod 700 /home/llmnode-runner/.ssh
+chmod 600 /home/llmnode-runner/.ssh/authorized_keys
+chown -R llmnode-runner:llmnode-runner /home/llmnode-runner/.ssh
+
+nvidia-smi
+echo "GPU setup complete"
+touch /var/log/startup-script-done
+STARTUP
+)
+
+startup_script="${startup_script//PUBLIC_KEY_PLACEHOLDER/$llmnode_runner_user_public_key}"
+
+echo "$startup_script" > $gpu_instance_provision_tmp_dir/startup-script.sh
+
+# Create the instance, trying each zone until one succeeds
+echo "=== Creating GCP GPU instance ==="
+echo "  Project: $GCLOUD_PROJECT"
+echo "  Machine type: $GCLOUD_MACHINE_TYPE"
+echo "  GPU: ${GCLOUD_GPU_COUNT}x $GCLOUD_GPU_TYPE"
+echo "  Disk size: ${GCLOUD_DISK_SIZE_GB}GB"
+echo "  Instance name: $instance_name"
+echo "  Zones to try: $GCLOUD_ZONES"
+
+for zone in $GCLOUD_ZONES; do
+    echo "Trying zone: $zone..."
+    if gcloud compute instances create "$instance_name" \
+        --project="$GCLOUD_PROJECT" \
+        --zone="$zone" \
+        --machine-type="$GCLOUD_MACHINE_TYPE" \
+        --accelerator="type=$GCLOUD_GPU_TYPE,count=$GCLOUD_GPU_COUNT" \
+        --maintenance-policy=TERMINATE \
+        --boot-disk-size="${GCLOUD_DISK_SIZE_GB}GB" \
+        --image-family="$GCLOUD_IMAGE_FAMILY" \
+        --image-project="$GCLOUD_IMAGE_PROJECT" \
+        --metadata-from-file=startup-script=$gpu_instance_provision_tmp_dir/startup-script.sh 2>&1; then
+        GCLOUD_ZONE="$zone"
+        echo "Created instance: $instance_name in zone: $GCLOUD_ZONE"
+        break
+    else
+        echo "Zone $zone unavailable, trying next..."
+    fi
+done
+
+if [[ -z "$GCLOUD_ZONE" ]]; then
+    echo "Error: Could not create instance in any zone"
+    exit 1
+fi
+
+# Wait for instance to be running
+echo "=== Waiting for instance to be running ==="
+max_attempts=60
+attempt=0
+while [ $attempt -lt $max_attempts ]; do
+    status=$(gcloud compute instances describe "$instance_name" \
+        --project="$GCLOUD_PROJECT" \
+        --zone="$GCLOUD_ZONE" \
+        --format="value(status)")
+
+    if [[ "$status" == "RUNNING" ]]; then
+        echo "Instance is running"
+        break
+    fi
+
+    echo "Instance status: $status (attempt $((attempt+1))/$max_attempts)"
+    sleep 10
+    attempt=$((attempt+1))
+done
+
+if [[ "$status" != "RUNNING" ]]; then
+    echo "Error: Instance did not become ready"
+    exit 1
+fi
+
+ssh_host=$(gcloud compute instances describe "$instance_name" \
+    --project="$GCLOUD_PROJECT" \
+    --zone="$GCLOUD_ZONE" \
+    --format="value(networkInterfaces[0].accessConfigs[0].natIP)")
+
+echo "Instance external IP: $ssh_host"
+
+# Wait for SSH
+echo "=== Waiting for SSH to be available ==="
+attempt=0
+while [ $attempt -lt 30 ]; do
+    if ssh -o StrictHostKeyChecking=no -o UserKnownHostsFile=/dev/null \
+        -o ConnectTimeout=5 -i $gpu_instance_user_private_key_path \
+        llmnode-runner@$ssh_host "echo 'SSH ready'" 2>/dev/null; then
+        echo "SSH is ready"
+        break
+    fi
+    echo "Waiting for SSH... (attempt $((attempt+1))/30)"
+    sleep 10
+    attempt=$((attempt+1))
+done
+
+# Wait for startup script to finish (NVIDIA drivers, CUDA)
+echo "=== Waiting for startup script to complete ==="
+attempt=0
+while [ $attempt -lt 60 ]; do
+    if ssh -o StrictHostKeyChecking=no -o UserKnownHostsFile=/dev/null \
+        -o ConnectTimeout=5 -i $gpu_instance_user_private_key_path \
+        llmnode-runner@$ssh_host "test -f /var/log/startup-script-done" 2>/dev/null; then
+        echo "Startup script completed"
+        break
+    fi
+    echo "Waiting for startup script to finish... (attempt $((attempt+1))/60)"
+    sleep 15
+    attempt=$((attempt+1))
+done
+
+if ! ssh -o StrictHostKeyChecking=no -o UserKnownHostsFile=/dev/null \
+    -o ConnectTimeout=5 -i $gpu_instance_user_private_key_path \
+    llmnode-runner@$ssh_host "test -f /var/log/startup-script-done" 2>/dev/null; then
+    echo "Error: Startup script did not complete"
+    echo "=== Startup script log ==="
+    ssh -o StrictHostKeyChecking=no -o UserKnownHostsFile=/dev/null \
+        -i $gpu_instance_user_private_key_path \
+        llmnode-runner@$ssh_host "sudo cat /var/log/startup-script.log" 2>/dev/null || true
+    exit 1
+fi
+
+# Copy test files to instance
+echo "=== Copying test files to GPU instance ==="
+rsync -az --progress \
+    -e "ssh -o StrictHostKeyChecking=no -o UserKnownHostsFile=/dev/null -i $gpu_instance_user_private_key_path" \
+    --exclude='.venv' \
+    --exclude='__pycache__' \
+    --exclude='.pytest_cache' \
+    --exclude='.mypy_cache' \
+    --exclude='gitlab' \
+    $REPO_ROOT/ai-suite/llm-node/ \
+    llmnode-runner@$ssh_host:~/llm-node
+
+# Run integration tests
+echo "=== Running vLLM integration tests ==="
+ssh -o StrictHostKeyChecking=no -o UserKnownHostsFile=/dev/null \
+    -i $gpu_instance_user_private_key_path \
+    llmnode-runner@$ssh_host \
+    "pip install uv && \
+     export PATH=\$HOME/.local/bin:/usr/local/cuda/bin:\$PATH && \
+     export LD_LIBRARY_PATH=/usr/local/cuda/lib64:\$LD_LIBRARY_PATH && \
+     cd ~/llm-node && \
+     uv sync && \
+     uv pip install vllm && \
+     uv run pytest volknode/src/llm/llm_integration_test.py -v"
+
+echo "=== vLLM integration tests completed successfully ==="