foxhunt/archive/scripts/deploy_hyperopt_direct.sh

#!/bin/bash
set -e

# Load RunPod credentials
source .env.runpod

TIMESTAMP=$(date +%Y%m%d_%H%M%S)

echo "========================================="
echo "RunPod Hyperopt Deployment (Direct REST API)"
echo "========================================="
echo ""
echo "Deploying 2 pods:"
echo "  1. DQN Hyperopt"
echo "  2. PPO Hyperopt"
echo ""

# Deploy DQN Hyperopt Pod
echo "========================================="
echo "1. DEPLOYING DQN HYPEROPT POD"
echo "========================================="

DQN_OUTPUT_DIR="dqn_hyperopt_${TIMESTAMP}"

DQN_PAYLOAD=$(cat <<EOF
{
  "cloudType": "SECURE",
  "computeType": "GPU",
  "dataCenterIds": ["EUR-IS-1"],
  "dataCenterPriority": "availability",
  "gpuTypeIds": ["NVIDIA RTX A4000"],
  "gpuCount": 1,
  "name": "foxhunt-dqn-hyperopt-${TIMESTAMP}",
  "imageName": "jgrusewski/foxhunt-hyperopt:latest",
  "containerDiskInGb": 50,
  "networkVolumeId": "${RUNPOD_VOLUME_ID}",
  "volumeMountPath": "/runpod-volume",
  "dockerStartCmd": [
    "hyperopt_dqn_demo",
    "--parquet-file", "/runpod-volume/test_data/ES_FUT_180d.parquet",
    "--trials", "50",
    "--epochs", "100",
    "--base-dir", "/runpod-volume/ml_training/${DQN_OUTPUT_DIR}"
  ],
  "containerRegistryAuthId": "${RUNPOD_CONTAINER_REGISTRY_AUTH_ID}",
  "ports": ["8888/http", "22/tcp"],
  "interruptible": false
}
EOF
)

echo "Deploying DQN pod..."
DQN_RESPONSE=$(curl -s -X POST https://rest.runpod.io/v1/pods \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer ${RUNPOD_API_KEY}" \
  -d "$DQN_PAYLOAD")

DQN_POD_ID=$(echo "$DQN_RESPONSE" | jq -r '.id // "ERROR"')

if [ "$DQN_POD_ID" = "ERROR" ] || [ "$DQN_POD_ID" = "null" ]; then
  echo "ERROR: Failed to deploy DQN pod"
  echo "Response: $DQN_RESPONSE"
  exit 1
fi

echo "✅ DQN Pod Deployed Successfully"
echo "  Pod ID: $DQN_POD_ID"
echo "  GPU: $(echo "$DQN_RESPONSE" | jq -r '.machine.gpuType.displayName // "RTX A4000"')"
echo "  Datacenter: $(echo "$DQN_RESPONSE" | jq -r '.machine.dataCenterId // "EUR-IS-1"')"
echo "  Cost: \$$(echo "$DQN_RESPONSE" | jq -r '.costPerHr // "0.25"')/hr"
echo "  Output: /runpod-volume/ml_training/${DQN_OUTPUT_DIR}"
echo ""

# Deploy PPO Hyperopt Pod
echo "========================================="
echo "2. DEPLOYING PPO HYPEROPT POD"
echo "========================================="

PPO_OUTPUT_DIR="ppo_hyperopt_${TIMESTAMP}"

PPO_PAYLOAD=$(cat <<EOF
{
  "cloudType": "SECURE",
  "computeType": "GPU",
  "dataCenterIds": ["EUR-IS-1"],
  "dataCenterPriority": "availability",
  "gpuTypeIds": ["NVIDIA RTX A4000"],
  "gpuCount": 1,
  "name": "foxhunt-ppo-hyperopt-${TIMESTAMP}",
  "imageName": "jgrusewski/foxhunt-hyperopt:latest",
  "containerDiskInGb": 50,
  "networkVolumeId": "${RUNPOD_VOLUME_ID}",
  "volumeMountPath": "/runpod-volume",
  "dockerStartCmd": [
    "hyperopt_ppo_demo",
    "--parquet-file", "/runpod-volume/test_data/ES_FUT_180d.parquet",
    "--trials", "50",
    "--episodes", "2000",
    "--base-dir", "/runpod-volume/ml_training/${PPO_OUTPUT_DIR}",
    "--early-stopping-min-epochs", "50"
  ],
  "containerRegistryAuthId": "${RUNPOD_CONTAINER_REGISTRY_AUTH_ID}",
  "ports": ["8888/http", "22/tcp"],
  "interruptible": false
}
EOF
)

echo "Deploying PPO pod..."
PPO_RESPONSE=$(curl -s -X POST https://rest.runpod.io/v1/pods \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer ${RUNPOD_API_KEY}" \
  -d "$PPO_PAYLOAD")

PPO_POD_ID=$(echo "$PPO_RESPONSE" | jq -r '.id // "ERROR"')

if [ "$PPO_POD_ID" = "ERROR" ] || [ "$PPO_POD_ID" = "null" ]; then
  echo "ERROR: Failed to deploy PPO pod"
  echo "Response: $PPO_RESPONSE"
  exit 1
fi

echo "✅ PPO Pod Deployed Successfully"
echo "  Pod ID: $PPO_POD_ID"
echo "  GPU: $(echo "$PPO_RESPONSE" | jq -r '.machine.gpuType.displayName // "RTX A4000"')"
echo "  Datacenter: $(echo "$PPO_RESPONSE" | jq -r '.machine.dataCenterId // "EUR-IS-1"')"
echo "  Cost: \$$(echo "$PPO_RESPONSE" | jq -r '.costPerHr // "0.25"')/hr"
echo "  Output: /runpod-volume/ml_training/${PPO_OUTPUT_DIR}"
echo ""

# Summary
echo "========================================="
echo "DEPLOYMENT SUMMARY"
echo "========================================="
echo ""
echo "DQN Hyperopt:"
echo "  Pod ID: $DQN_POD_ID"
echo "  Output: /runpod-volume/ml_training/${DQN_OUTPUT_DIR}"
echo "  Trials: 50"
echo "  Epochs/trial: 100"
echo "  Expected duration: 12-25 min"
echo "  Expected cost: \$0.05-\$0.10"
echo ""
echo "PPO Hyperopt:"
echo "  Pod ID: $PPO_POD_ID"
echo "  Output: /runpod-volume/ml_training/${PPO_OUTPUT_DIR}"
echo "  Trials: 50"
echo "  Episodes/trial: 2000"
echo "  Expected duration: 10-20 min"
echo "  Expected cost: \$0.04-\$0.08"
echo ""
echo "TOTAL ESTIMATED COST: \$0.09-\$0.18"
echo ""
echo "========================================="
echo "MONITORING COMMANDS"
echo "========================================="
echo ""
echo "Monitor DQN logs:"
echo "  ./monitor_dqn_hyperopt_pod.sh $DQN_POD_ID"
echo ""
echo "Monitor PPO logs:"
echo "  ./monitor_ppo_hyperopt_pod.sh $PPO_POD_ID"
echo ""
echo "View pods in dashboard:"
echo "  https://www.runpod.io/console/pods"
echo ""
echo "Check S3 results (after completion):"
echo "  aws s3 ls s3://se3zdnb5o4/ml_training/${DQN_OUTPUT_DIR}/ --profile runpod --recursive"
echo "  aws s3 ls s3://se3zdnb5o4/ml_training/${PPO_OUTPUT_DIR}/ --profile runpod --recursive"
echo ""
echo "Terminate pods (when complete):"
echo "  curl -X POST https://rest.runpod.io/v1/pods/${DQN_POD_ID}/terminate \\"
echo "    -H \"Authorization: Bearer \$RUNPOD_API_KEY\""
echo "  curl -X POST https://rest.runpod.io/v1/pods/${PPO_POD_ID}/terminate \\"
echo "    -H \"Authorization: Bearer \$RUNPOD_API_KEY\""
echo ""