Merge pull request #656 from aws-samples/twelvelabs-sample

twelvelabs video embedding search sample update - add 3.0 sample, upd…

Author: sharonxiaohanli

…ock-twelvelabs-embedding-s3vectors.ipynb …bs-embedding-s3vectors-marengo-2.7.ipynbmulti-modal/TwelveLabs/bedrock-twelvelabs-embedding-s3vectors.ipynb renamed to multi-modal/TwelveLabs/bedrock-twelvelabs-embedding-s3vectors-marengo-2.7.ipynb multi-modal/TwelveLabs/bedrock-twelvelabs-embedding-s3vectors.ipynb renamed to multi-modal/TwelveLabs/bedrock-twelvelabs-embedding-s3vectors-marengo-2.7.ipynb

@@ -5,7 +5,7 @@
    "id": "3a0b2a41-bebe-4f3a-b803-eadc275a916f",
    "metadata": {},
    "source": [
-    "# TwelveLabs Multimodal Embedding Search using Amazon Bedrock and Amazon S3 Vectors\n",
+    "# TwelveLabs Multimodal Embedding (Marengo Embed 2.7) Search using Amazon Bedrock and Amazon S3 Vectors\n",
     "Work with TwelveLabs Marengo Embed 2.7 Model and Amazon S3 Vectores\n",
     "![TwelveLabs Embedding](./images/12labs-embed-s3vectors.png)"
    ]
@@ -82,7 +82,8 @@
     "aws_account_id = '<YOUR_AWS_ACCOUNT_ID>'\n",
     "\n",
     "s3vector_bucket = \"<S3_VECTOR_BUCKET_NAME_TO_CREATE>\"\n",
-    "s3vector_index = \"<S3_VECTOR_INDEX_NAME_TO_CREATE>\""
+    "s3vector_index = \"<S3_VECTOR_INDEX_NAME_TO_CREATE>\"\n",
+    "dimension = 1024"
    ]
   },
   {
@@ -235,8 +236,11 @@
    "outputs": [],
    "source": [
     "# Create a S3 vector bucket\n",
-    "s3vectors.create_vector_bucket(vectorBucketName=s3vector_bucket)\n",
-    "print(f\"Vector bucket '{s3vector_bucket}' created successfully.\")"
+    "try:\n",
+    "    s3vectors.create_vector_bucket(vectorBucketName=s3vector_bucket)\n",
+    "    print(f\"Vector bucket '{s3vector_bucket}' created successfully.\")\n",
+    "except Exception as ex:\n",
+    "    print(ex)"
    ]
   },
   {
@@ -247,17 +251,19 @@
    "outputs": [],
    "source": [
     "# Create an index in the vector store\n",
-    "vector_dimension = 1024\n",
-    "distance_metric = 'cosine' # or 'euclidean'\n",
-    "\n",
-    "s3vectors.create_index(\n",
-    "    vectorBucketName=s3vector_bucket,\n",
-    "    indexName=s3vector_index,\n",
-    "    dataType='float32',  # Common data type for vector embeddings\n",
-    "    dimension=vector_dimension,\n",
-    "    distanceMetric=distance_metric\n",
-    ")\n",
-    "print(f\"Vector index '{s3vector_index}' created successfully in bucket '{s3vector_bucket}'.\")\n"
+    "try:\n",
+    "    distance_metric = 'cosine' # or 'euclidean'\n",
+    "    \n",
+    "    s3vectors.create_index(\n",
+    "        vectorBucketName=s3vector_bucket,\n",
+    "        indexName=s3vector_index,\n",
+    "        dataType='float32',  # Common data type for vector embeddings\n",
+    "        dimension=dimension,\n",
+    "        distanceMetric=distance_metric\n",
+    "    )\n",
+    "    print(f\"Vector index '{s3vector_index}' created successfully in bucket '{s3vector_bucket}'.\")\n",
+    "except Exception as ex:\n",
+    "    print(ex)"
    ]
   },
   {
@@ -312,8 +318,10 @@
   {
    "cell_type": "code",
    "execution_count": null,
-   "id": "795ad337-ab73-4676-a2fc-93d504f6dda0",
-   "metadata": {},
+   "id": "c2d15aa6-6924-47d7-abc6-d78d6c252068",
+   "metadata": {
+    "scrolled": true
+   },
    "outputs": [],
    "source": [
     "# Read image\n",
@@ -322,32 +330,23 @@
     "with open('./images/meridian-car.png', \"rb\") as image_file:\n",
     "    base64_string = base64.b64encode(image_file.read()).decode(\"utf-8\")\n",
     "\n",
-    "import uuid\n",
-    "query_prefix = f'{s3_prefix}/input/{uuid.uuid4()}'\n",
-    "\n",
-    "# Create an input embedding\n",
-    "response = bedrock.start_async_invoke(\n",
-    "    modelId=model_id,\n",
-    "    modelInput = {\n",
-    "        \"inputType\": \"image\",\n",
-    "        \"mediaSource\": {\n",
-    "            \"base64String\": base64_string\n",
-    "        }\n",
-    "    },\n",
-    "    outputDataConfig={\n",
-    "        \"s3OutputDataConfig\": {\n",
-    "            \"s3Uri\": f's3://{s3_bucket}/{query_prefix}'\n",
-    "        }\n",
-    "    }\n",
+    "request = {\n",
+    "   \"inputType\": \"image\",\n",
+    "   \"mediaSource\": {\n",
+    "      \"base64String\": base64_string\n",
+    "   }\n",
+    "}\n",
+    "\n",
+    "# Make the request\n",
+    "response = bedrock.invoke_model(\n",
+    "    modelId=f'us.{model_id}',\n",
+    "    body=json.dumps(request)\n",
     ")\n",
     "\n",
-    "# Print Job ID\n",
-    "invocation_arn = response[\"invocationArn\"]\n",
-    "print(\"Async Job Started\")\n",
-    "print(\"Invocation Arn:\", invocation_arn)\n",
-    "\n",
-    "query = wait_for_output_file(s3_bucket, query_prefix, invocation_arn)\n",
-    "display(JSON(query))"
+    "# Print the response body\n",
+    "response_body = json.loads(response['body'].read().decode('utf-8'))\n",
+    "embedding = response_body[\"data\"][0][\"embedding\"]\n",
+    "display(JSON(embedding))"
    ]
   },
   {
@@ -368,8 +367,6 @@
    },
    "outputs": [],
    "source": [
-    "embedding = query[0][\"embedding\"]\n",
-    "\n",
     "# Query vector index.\n",
     "response = s3vectors.query_vectors(\n",
     "    vectorBucketName=s3vector_bucket,\n",