Camel Components

AWS S3 Vectors

Since Camel 4.17

Both producer and consumer are supported

The AWS S3 Vectors component stores and queries vector embeddings using Amazon S3 Vectors.

Prerequisites

You need an AWS account with access to S3 Vectors. See Amazon S3 Vectors.

URI Format

aws2-s3-vectors://vectorBucketName[?options]

You can append query options to the URI:

?option1=value&option2=value&…​

Configuring Options

Camel components are configured on two separate levels:

  • component level

  • endpoint level

Configuring Component Options

At the component level, you set general and shared configurations that are, then, inherited by the endpoints. It is the highest configuration level.

For example, a component may have security settings, credentials for authentication, urls for network connection and so forth.

Some components only have a few options, and others may have many. Because components typically have pre-configured defaults that are commonly used, then you may often only need to configure a few options on a component; or none at all.

You can configure components using:

  • the Component DSL.

  • in a configuration file (application.properties, *.yaml files, etc).

  • directly in the Java code.

Configuring Endpoint Options

You usually spend more time setting up endpoints because they have many options. These options help you customize what you want the endpoint to do. The options are also categorized into whether the endpoint is used as a consumer (from), as a producer (to), or both.

Configuring endpoints is most often done directly in the endpoint URI as path and query parameters. You can also use the Endpoint DSL and DataFormat DSL as a type safe way of configuring endpoints and data formats in Java.

A good practice when configuring options is to use Property Placeholders.

Property placeholders provide a few benefits:

  • They help prevent using hardcoded urls, port numbers, sensitive information, and other settings.

  • They allow externalizing the configuration from the code.

  • They help the code to become more flexible and reusable.

The following two sections list all the options, firstly for the component followed by the endpoint.

Component Options

The AWS S3 Vectors component supports the following options which are listed below.

Name Description Default Type

configuration (common)

The component configuration.

AWS2S3VectorsConfiguration

dataType (common)

The data type of the vectors of the index to create. S3 Vectors supports float32 only.

Enum values:

  • float32

float32

String

distanceMetric (common)

The distance metric of the index to create, which the similarity search uses. Options: cosine, euclidean.

Enum values:

  • cosine

  • euclidean

cosine

String

overrideEndpoint (common)

Set the need for overriding the endpoint. This option needs to be used in combination with uriEndpointOverride option.

false

boolean

region (common)

The region in which S3 Vectors client needs to work. When using this parameter, the configuration will expect the lowercase name of the region (for example ap-east-1).

Enum values:

  • ap-south-2

  • ap-south-1

  • eu-south-1

  • eu-south-2

  • us-gov-east-1

  • me-central-1

  • il-central-1

  • ca-central-1

  • eu-central-1

  • us-iso-west-1

  • eu-central-2

  • eu-isoe-west-1

  • us-west-1

  • us-west-2

  • af-south-1

  • eu-north-1

  • eu-west-3

  • eu-west-2

  • eu-west-1

  • ap-northeast-3

  • ap-northeast-2

  • ap-northeast-1

  • me-south-1

  • sa-east-1

  • ap-east-1

  • cn-north-1

  • ca-west-1

  • us-gov-west-1

  • ap-southeast-1

  • ap-southeast-2

  • us-iso-east-1

  • ap-southeast-3

  • ap-southeast-4

  • us-east-1

  • us-east-2

  • cn-northwest-1

  • us-isob-east-1

  • aws-global

  • aws-cn-global

  • aws-us-gov-global

  • aws-iso-global

  • aws-iso-b-global

String

similarityThreshold (common)

Deprecated (deprecated) Never used: S3 Vectors has no similarity threshold. The consumer sets the CamelAwsS3VectorsDistance header to the distance of each vector from the query vector. Deprecation note: S3 Vectors has no similarity threshold, the component never used this option. Filter the results on their CamelAwsS3VectorsDistance header instead.

Float

topK (common)

The number of top similar vectors to return in a query.

10

Integer

uriEndpointOverride (common)

Set the overriding uri endpoint. This option needs to be used in combination with overrideEndpoint option.

String

vectorDimensions (common)

The dimensions of the vector embeddings (default: 1536, which is the dimension for OpenAI text-embedding-3-small).

1536

Integer

vectorIndexName (common)

The name of the vector index.

String

bridgeErrorHandler (consumer)

Allows for bridging the consumer to the Camel routing Error Handler, which mean any exceptions (if possible) occurred while the Camel consumer is trying to pickup incoming messages, or the likes, will now be processed as a message and handled by the routing Error Handler. Important: This is only possible if the 3rd party component allows Camel to be alerted if an exception was thrown. Some components handle this internally only, and therefore bridgeErrorHandler is not possible. In other situations we may improve the Camel component to hook into the 3rd party component and make this possible for future releases. By default the consumer will use the org.apache.camel.spi.ExceptionHandler to deal with exceptions, that will be logged at WARN or ERROR level and ignored.

false

boolean

consumerMetadataFilter (consumer)

The metadata filter of the similarity search of the consumer, as a JSON object with the filter operators of S3 Vectors, such as \{genre: \{$eq: drama}}. Only the vectors whose metadata match the filter are consumed.

String

consumerQueryVector (consumer)

The query vector to use for the consumer to poll for similar vectors. Specified as comma-separated float values (e.g., 0.1,0.2,0.3). If not specified, the consumer will not poll.

String

delay (consumer)

Milliseconds before the next poll for the consumer.

500

long

deleteAfterRead (consumer)

Delete vectors after they have been consumed.

false

boolean

maxMessagesPerPoll (consumer)

The maximum number of messages to consume per poll for the consumer.

10

int

lazyStartProducer (producer)

Whether the producer should be started lazy (on the first message). By starting lazy you can use this to allow CamelContext and routes to startup in situations where a producer may otherwise fail during starting and cause the route to fail being started. By deferring this startup to be lazy then the startup failure can be handled during routing messages via Camel’s routing error handlers. Beware that when the first message is processed then creating and starting the producer may take a little time and prolong the total processing time of the processing.

false

boolean

operation (producer)

The operation to perform.

Enum values:

  • putVectors

  • queryVectors

  • deleteVectors

  • getVectors

  • createVectorBucket

  • deleteVectorBucket

  • listVectorBuckets

  • describeVectorBucket

  • createVectorIndex

  • deleteVectorIndex

  • listVectorIndexes

  • describeVectorIndex

AWS2S3VectorsOperations

autowiredEnabled (advanced)

Whether autowiring is enabled. This is used for automatic autowiring options (the option must be marked as autowired) by looking up in the registry to find if there is a single instance of matching type, which then gets configured on the component. This can be used for automatic configuring JDBC data sources, JMS connection factories, AWS Clients, etc.

true

boolean

s3VectorsClient (advanced)

Autowired Reference to a software.amazon.awssdk.services.s3vectors.S3VectorsClient in the registry.

S3VectorsClient

healthCheckConsumerEnabled (health)

Used for enabling or disabling all consumer based health checks from this component.

true

boolean

healthCheckProducerEnabled (health)

Used for enabling or disabling all producer based health checks from this component. Notice: Camel has by default disabled all producer based health-checks. You can turn on producer checks globally by setting camel.health.producersEnabled=true.

true

boolean

proxyHost (proxy)

To define a proxy host when instantiating the S3 Vectors client.

String

proxyPort (proxy)

To define a proxy port when instantiating the S3 Vectors client.

Integer

proxyProtocol (proxy)

To define a proxy protocol when instantiating the S3 Vectors client.

Enum values:

  • HTTP

  • HTTPS

HTTPS

Protocol

accessKey (security)

Amazon AWS Access Key.

String

profileCredentialsName (security)

If using a profile credentials provider, this parameter will set the profile name.

String

secretKey (security)

Amazon AWS Secret Key.

String

sessionToken (security)

Amazon AWS Session Token used when the user needs to assume an IAM role.

String

trustAllCertificates (security)

If we want to trust all certificates in case of overriding the endpoint.

false

boolean

useDefaultCredentialsProvider (security)

Set whether the S3 Vectors client should expect to load credentials through a default credentials provider.

false

boolean

useProfileCredentialsProvider (security)

Set whether the S3 Vectors client should expect to load credentials through a profile credentials provider.

false

boolean

useSessionCredentials (security)

Set whether the S3 Vectors client should expect to use Session Credentials. This is useful in a situation in which the user needs to assume an IAM role for doing operations in S3 Vectors.

false

boolean

Endpoint Options

The AWS S3 Vectors endpoint is configured using URI syntax:

aws2-s3-vectors://vectorBucketName

With the following path and query parameters:

Path Parameters

Name Description Default Type

vectorBucketName (common)

Required Vector bucket name or ARN.

String

Query Parameters

Name Description Default Type

dataType (common)

The data type of the vectors of the index to create. S3 Vectors supports float32 only.

Enum values:

  • float32

float32

String

distanceMetric (common)

The distance metric of the index to create, which the similarity search uses. Options: cosine, euclidean.

Enum values:

  • cosine

  • euclidean

cosine

String

overrideEndpoint (common)

Set the need for overriding the endpoint. This option needs to be used in combination with uriEndpointOverride option.

false

boolean

region (common)

The region in which S3 Vectors client needs to work. When using this parameter, the configuration will expect the lowercase name of the region (for example ap-east-1).

Enum values:

  • ap-south-2

  • ap-south-1

  • eu-south-1

  • eu-south-2

  • us-gov-east-1

  • me-central-1

  • il-central-1

  • ca-central-1

  • eu-central-1

  • us-iso-west-1

  • eu-central-2

  • eu-isoe-west-1

  • us-west-1

  • us-west-2

  • af-south-1

  • eu-north-1

  • eu-west-3

  • eu-west-2

  • eu-west-1

  • ap-northeast-3

  • ap-northeast-2

  • ap-northeast-1

  • me-south-1

  • sa-east-1

  • ap-east-1

  • cn-north-1

  • ca-west-1

  • us-gov-west-1

  • ap-southeast-1

  • ap-southeast-2

  • us-iso-east-1

  • ap-southeast-3

  • ap-southeast-4

  • us-east-1

  • us-east-2

  • cn-northwest-1

  • us-isob-east-1

  • aws-global

  • aws-cn-global

  • aws-us-gov-global

  • aws-iso-global

  • aws-iso-b-global

String

similarityThreshold (common)

Deprecated (deprecated) Never used: S3 Vectors has no similarity threshold. The consumer sets the CamelAwsS3VectorsDistance header to the distance of each vector from the query vector. Deprecation note: S3 Vectors has no similarity threshold, the component never used this option. Filter the results on their CamelAwsS3VectorsDistance header instead.

Float

topK (common)

The number of top similar vectors to return in a query.

10

Integer

uriEndpointOverride (common)

Set the overriding uri endpoint. This option needs to be used in combination with overrideEndpoint option.

String

vectorDimensions (common)

The dimensions of the vector embeddings (default: 1536, which is the dimension for OpenAI text-embedding-3-small).

1536

Integer

vectorIndexName (common)

The name of the vector index.

String

consumerMetadataFilter (consumer)

The metadata filter of the similarity search of the consumer, as a JSON object with the filter operators of S3 Vectors, such as \{genre: \{$eq: drama}}. Only the vectors whose metadata match the filter are consumed.

String

consumerQueryVector (consumer)

The query vector to use for the consumer to poll for similar vectors. Specified as comma-separated float values (e.g., 0.1,0.2,0.3). If not specified, the consumer will not poll.

String

delay (consumer)

Milliseconds before the next poll for the consumer.

500

long

deleteAfterRead (consumer)

Delete vectors after they have been consumed.

false

boolean

maxMessagesPerPoll (consumer)

The maximum number of messages to consume per poll for the consumer.

10

int

sendEmptyMessageWhenIdle (consumer)

If the polling consumer did not poll any files, you can enable this option to send an empty message (no body) instead.

false

boolean

bridgeErrorHandler (consumer (advanced))

Allows for bridging the consumer to the Camel routing Error Handler, which mean any exceptions (if possible) occurred while the Camel consumer is trying to pickup incoming messages, or the likes, will now be processed as a message and handled by the routing Error Handler. Important: This is only possible if the 3rd party component allows Camel to be alerted if an exception was thrown. Some components handle this internally only, and therefore bridgeErrorHandler is not possible. In other situations we may improve the Camel component to hook into the 3rd party component and make this possible for future releases. By default the consumer will use the org.apache.camel.spi.ExceptionHandler to deal with exceptions, that will be logged at WARN or ERROR level and ignored.

false

boolean

exceptionHandler (consumer (advanced))

To let the consumer use a custom ExceptionHandler. Notice if the option bridgeErrorHandler is enabled then this option is not in use. By default the consumer will deal with exceptions, that will be logged at WARN or ERROR level and ignored.

ExceptionHandler

exchangePattern (consumer (advanced))

Sets the exchange pattern when the consumer creates an exchange.

Enum values:

  • InOnly

  • InOut

ExchangePattern

pollStrategy (consumer (advanced))

A pluggable org.apache.camel.PollingConsumerPollingStrategy allowing you to provide your custom implementation to control error handling usually occurred during the poll operation before an Exchange have been created and being routed in Camel.

PollingConsumerPollStrategy

operation (producer)

The operation to perform.

Enum values:

  • putVectors

  • queryVectors

  • deleteVectors

  • getVectors

  • createVectorBucket

  • deleteVectorBucket

  • listVectorBuckets

  • describeVectorBucket

  • createVectorIndex

  • deleteVectorIndex

  • listVectorIndexes

  • describeVectorIndex

AWS2S3VectorsOperations

lazyStartProducer (producer (advanced))

Whether the producer should be started lazy (on the first message). By starting lazy you can use this to allow CamelContext and routes to startup in situations where a producer may otherwise fail during starting and cause the route to fail being started. By deferring this startup to be lazy then the startup failure can be handled during routing messages via Camel’s routing error handlers. Beware that when the first message is processed then creating and starting the producer may take a little time and prolong the total processing time of the processing.

false

boolean

s3VectorsClient (advanced)

Autowired Reference to a software.amazon.awssdk.services.s3vectors.S3VectorsClient in the registry.

S3VectorsClient

proxyHost (proxy)

To define a proxy host when instantiating the S3 Vectors client.

String

proxyPort (proxy)

To define a proxy port when instantiating the S3 Vectors client.

Integer

proxyProtocol (proxy)

To define a proxy protocol when instantiating the S3 Vectors client.

Enum values:

  • HTTP

  • HTTPS

HTTPS

Protocol

backoffErrorThreshold (scheduler)

The number of subsequent error polls (failed due some error) that should happen before the backoffMultipler should kick-in.

int

backoffIdleThreshold (scheduler)

The number of subsequent idle polls that should happen before the backoffMultipler should kick-in.

int

backoffMultiplier (scheduler)

To let the scheduled polling consumer backoff if there has been a number of subsequent idles/errors in a row. The multiplier is then the number of polls that will be skipped before the next actual attempt is happening again. When this option is in use then backoffIdleThreshold and/or backoffErrorThreshold must also be configured.

int

greedy (scheduler)

If greedy is enabled, then the ScheduledPollConsumer will run immediately again, if the previous run polled 1 or more messages.

false

boolean

initialDelay (scheduler)

Milliseconds before the first poll starts.

1000

long

repeatCount (scheduler)

Specifies a maximum limit of number of fires. So if you set it to 1, the scheduler will only fire once. If you set it to 5, it will only fire five times. A value of zero or negative means fire forever.

0

long

runLoggingLevel (scheduler)

The consumer logs a start/complete log line when it polls. This option allows you to configure the logging level for that.

Enum values:

  • TRACE

  • DEBUG

  • INFO

  • WARN

  • ERROR

  • OFF

TRACE

LoggingLevel

scheduledExecutorService (scheduler)

Allows for configuring a custom/shared thread pool to use for the consumer. By default each consumer has its own single threaded thread pool.

ScheduledExecutorService

scheduler (scheduler)

To use a cron scheduler from either camel-spring or camel-quartz component. Use value spring or quartz for built in scheduler.

none

Object

schedulerProperties (scheduler)

To configure additional properties when using a custom scheduler or any of the Quartz, Spring based scheduler. This is a multi-value option with prefix: scheduler.

Map

startScheduler (scheduler)

Whether the scheduler should be auto started.

true

boolean

timeUnit (scheduler)

Time unit for initialDelay and delay options.

Enum values:

  • NANOSECONDS

  • MICROSECONDS

  • MILLISECONDS

  • SECONDS

  • MINUTES

  • HOURS

  • DAYS

MILLISECONDS

TimeUnit

useFixedDelay (scheduler)

Controls if fixed delay or fixed rate is used. See ScheduledExecutorService in JDK for details.

true

boolean

accessKey (security)

Amazon AWS Access Key.

String

profileCredentialsName (security)

If using a profile credentials provider, this parameter will set the profile name.

String

secretKey (security)

Amazon AWS Secret Key.

String

sessionToken (security)

Amazon AWS Session Token used when the user needs to assume an IAM role.

String

trustAllCertificates (security)

If we want to trust all certificates in case of overriding the endpoint.

false

boolean

useDefaultCredentialsProvider (security)

Set whether the S3 Vectors client should expect to load credentials through a default credentials provider.

false

boolean

useProfileCredentialsProvider (security)

Set whether the S3 Vectors client should expect to load credentials through a profile credentials provider.

false

boolean

useSessionCredentials (security)

Set whether the S3 Vectors client should expect to use Session Credentials. This is useful in a situation in which the user needs to assume an IAM role for doing operations in S3 Vectors.

false

boolean

Message Headers

The AWS S3 Vectors component supports the following message header(s), which is/are listed below:

Name Description Default Type

CamelAwsS3VectorsOperation (common)

Constant: OPERATION

The operation to perform.

String

CamelAwsS3VectorsVectorBucketName (producer)

Constant: VECTOR_BUCKET_NAME

The name of the vector bucket to use for the current operation instead of the vector bucket of the endpoint. An input only: neither the producer nor the consumer sets it.

String

CamelAwsS3VectorsVectorIndexName (producer)

Constant: VECTOR_INDEX_NAME

The name of the vector index to use for the current operation instead of the vectorIndexName option of the endpoint. An input only: neither the producer nor the consumer sets it.

String

CamelAwsS3VectorsVectorId (common)

Constant: VECTOR_ID

The key of a vector. putVectors puts the vector under this key (the exchange id when it is not set) and sets it to the key it used. getVectors and deleteVectors take their key, or a list of keys, from this header only when the message body holds no keys. The consumer sets it to the key of the consumed vector.

String or List

CamelAwsS3VectorsVectorData (producer)

Constant: VECTOR_DATA

The vector embedding data as a list of floats or float array.

List or float[]

CamelAwsS3VectorsVectorDimensions (producer)

Constant: VECTOR_DIMENSIONS

The dimensions of the vector.

Integer

CamelAwsS3VectorsDataType (producer)

Constant: DATA_TYPE

The data type of the vectors of the index to create (float32).

String

CamelAwsS3VectorsVectorMetadata (producer)

Constant: VECTOR_METADATA

Additional metadata for the vector as a map. The values may be Strings, numbers, booleans or lists of these.

Map

CamelAwsS3VectorsQueryVector (producer)

Constant: QUERY_VECTOR

The query vector for similarity search as a list of floats or float array.

List or float[]

CamelAwsS3VectorsTopK (producer)

Constant: TOP_K

The number of top similar vectors to return.

Integer

CamelAwsS3VectorsDistanceMetric (producer)

Constant: DISTANCE_METRIC

The distance metric of the index to create (cosine or euclidean).

String

CamelAwsS3VectorsSimilarityThreshold (producer)

Constant: SIMILARITY_THRESHOLD

Deprecated (deprecated) Never used: S3 Vectors has no similarity threshold. Deprecation note: S3 Vectors has no similarity threshold, the component never used this header. Filter the results on their CamelAwsS3VectorsDistance header instead.

Float

CamelAwsS3VectorsMetadataFilter (producer)

Constant: METADATA_FILTER

The metadata filter of queryVectors, as a JSON object with the filter operators of S3 Vectors, such as $eq, $in or $and. Only the vectors whose metadata match the filter are returned.

String

CamelAwsS3VectorsSimilarityScore (consumer)

Constant: SIMILARITY_SCORE

Deprecated (deprecated) Never set: S3 Vectors returns a distance, see CamelAwsS3VectorsDistance. Deprecation note: S3 Vectors returns a distance, not a similarity score, and the component never set this header. Use the CamelAwsS3VectorsDistance header instead.

Float

CamelAwsS3VectorsDistance (consumer)

Constant: DISTANCE

The distance of the consumed vector from the query vector, computed with the distance metric of the index: the lower, the closer.

Float

CamelAwsS3VectorsResultCount (consumer)

Constant: RESULT_COUNT

The number of vectors returned in the result.

Integer

CamelAwsS3VectorsIndexStatus (consumer)

Constant: INDEX_STATUS

The status of the vector index.

String

CamelAwsS3VectorsVectorBucketArn (consumer)

Constant: VECTOR_BUCKET_ARN

The ARN of the vector bucket.

String

Message Body

The message body at each side of the AWS S3 Vectors component: Consumer is the body of the messages the consumer creates, Consumer reply the bodies a consumer that answers its caller can reply with, Producer input the bodies the producer can send or write, and Producer output the body after the producer.

Side Body Description

Consumer

pojo (software.amazon.awssdk.services.s3vectors.model.QueryOutputVector)

One message per vector that the similarity search of consumerQueryVector finds and that was not sent before, as a QueryOutputVector of the AWS SDK with its key, its metadata and its distance, the key also in the CamelAwsS3VectorsVectorId header and the distance in the CamelAwsS3VectorsDistance header.

Producer input

any

The vector of putVectors and queryVectors, unless the CamelAwsS3VectorsVectorData or CamelAwsS3VectorsQueryVector header has it: one vector as a List of numbers, a float[] or a double[]. Text such as a JSON array fails, so unmarshal it first. The keys of getVectors and deleteVectors: a String, a String[] or a List of Strings. When the body holds no keys, they come from the CamelAwsS3VectorsVectorId header. The other operations do not use the body.

Producer output

pojo

The response of the AWS SDK for the operation, such as the PutVectorsResponse of putVectors, with the key of the vector in the CamelAwsS3VectorsVectorId header. A List for queryVectors, getVectors, listVectorBuckets and listVectorIndexes, the last two the first page only. The vectors of queryVectors hold their keys, their metadata and their distance, and the vectors of getVectors their keys, their data and their metadata.

  • With operation=queryVectors: list (java.util.List<software.amazon.awssdk.services.s3vectors.model.QueryOutputVector>)

  • With operation=getVectors: list (java.util.List<software.amazon.awssdk.services.s3vectors.model.GetOutputVector>)

  • With operation=listVectorBuckets: list (java.util.List<software.amazon.awssdk.services.s3vectors.model.VectorBucketSummary>)

  • With operation=listVectorIndexes: list (java.util.List<software.amazon.awssdk.services.s3vectors.model.IndexSummary>)

The bodies are the ones with the default options. A side that is not listed is not known, and tooling does not guess it. The carriers are described in the Camel Catalog.

Usage

Producer Operations

  • putVectors - Insert vectors

  • queryVectors - Search similar vectors

  • deleteVectors - Delete vectors

  • getVectors - Retrieve vectors by ID

  • createVectorBucket - Create bucket

  • deleteVectorBucket - Delete bucket

  • listVectorBuckets - List buckets (the first page of the result only)

  • describeVectorBucket - Get bucket info

  • createVectorIndex - Create index

  • deleteVectorIndex - Delete index

  • listVectorIndexes - List indexes (the first page of the result only)

  • describeVectorIndex - Get index info

listVectorBuckets and listVectorIndexes send a single request, so they return the first page of the buckets or indexes that S3 Vectors returns, not all of them.

Bucket and index of an operation

Each operation uses the vector bucket of the endpoint URI and the vectorIndexName option. The CamelAwsS3VectorsVectorBucketName and CamelAwsS3VectorsVectorIndexName headers override them for a single message. These headers are inputs only: neither the producer nor the consumer sets them, so a route that reads from one bucket and writes to another one does not write back into the bucket it reads from.

Vector keys

putVectors puts the vector under the key of the CamelAwsS3VectorsVectorId header, or under the exchange id when the header is not set, and sets the header to the key it used.

getVectors and deleteVectors take their keys from the message body when the body holds keys: a String, or a Collection or an array whose elements are all Strings. Otherwise they take the keys from the CamelAwsS3VectorsVectorId header (a String or a list of Strings). Keys that are not Strings, such as a list of numbers, fail with an IllegalArgumentException. As the body comes first, a getVectors after a putVectors in the same route gets the keys of its body, not the key that the putVectors left in the header.

getVectors returns the vectors with their data and their metadata.

Vector data and metadata

The vector of putVectors and queryVectors is a List of numbers, a float[] or a double[]. A null element fails with an IllegalArgumentException that names its position.

The CamelAwsS3VectorsVectorMetadata header of putVectors is a Map whose values may be Strings, numbers, booleans, or lists of these. Any other value fails with an IllegalArgumentException that names its key.

Metadata filter and distance

The CamelAwsS3VectorsMetadataFilter header of queryVectors and the consumerMetadataFilter option of the consumer are JSON objects with the filter operators of S3 Vectors, such as {"genre": {"$eq": "drama"}} or {"$and": [{"genre": "drama"}, {"year": {"$gte": 2020}}]}. Only the vectors whose metadata match the filter are returned. A filter that is not a valid JSON object fails with an IllegalArgumentException: the producer fails the exchange, and the consumer does not start.

queryVectors and the consumer return the vectors with their metadata and their distance from the query vector. The distance is computed with the distance metric of the index (cosine or euclidean): the lower, the closer. S3 Vectors has no similarity threshold, so the similarityThreshold option and the CamelAwsS3VectorsSimilarityThreshold and CamelAwsS3VectorsSimilarityScore headers are deprecated and do nothing. To keep only the close vectors, filter on the distance, such as .filter(simple("${header.CamelAwsS3VectorsDistance} < 0.3")) after the consumer.

Index data type and distance metric

The dataType option of createVectorIndex is float32, the only data type of S3 Vectors. The distanceMetric option is cosine (the default) or euclidean.

Consumer

The consumer polls for vectors using similarity search. It tracks processed vector IDs to avoid duplicates.

Each consumed vector is a QueryOutputVector in the message body, with its metadata. The consumer sets the CamelAwsS3VectorsVectorId header to the key of the vector and the CamelAwsS3VectorsDistance header to its distance from the consumerQueryVector.

Examples

The vector in the message body is a List of numbers, a float[] or a double[]. The YAML DSL has no way to write such a list as a constant, so the YAML examples build it with a Groovy expression and need camel-groovy on the classpath.

Insert Vectors

  • Java

  • YAML

from("direct:insert")
    .setHeader(AWS2S3VectorsConstants.VECTOR_ID, constant("doc-001"))
    .setBody(constant(Arrays.asList(0.1f, 0.2f, 0.3f)))
    .to("aws2-s3-vectors://my-bucket?operation=putVectors&vectorIndexName=my-index");
- route:
    from:
      uri: direct:insert
      steps:
        - setHeader:
            name: CamelAwsS3VectorsVectorId
            constant: doc-001
        - setBody:
            groovy: "[0.1, 0.2, 0.3]"
        - to:
            uri: aws2-s3-vectors://my-bucket
            parameters:
              operation: putVectors
              vectorIndexName: my-index

Query Similar Vectors

  • Java

  • YAML

from("direct:search")
    .setBody(constant(Arrays.asList(0.15f, 0.25f, 0.35f)))
    .setHeader(AWS2S3VectorsConstants.TOP_K, constant(5))
    .to("aws2-s3-vectors://my-bucket?operation=queryVectors&vectorIndexName=my-index")
    .log("Found ${body.size} similar vectors");
- route:
    from:
      uri: direct:search
      steps:
        - setBody:
            groovy: "[0.15, 0.25, 0.35]"
        - setHeader:
            name: CamelAwsS3VectorsTopK
            constant: 5
        - to:
            uri: aws2-s3-vectors://my-bucket
            parameters:
              operation: queryVectors
              vectorIndexName: my-index
        - log:
            message: "Found ${body.size} similar vectors"

Consumer Polling

  • Java

  • YAML

from("aws2-s3-vectors://my-bucket?"
    + "vectorIndexName=my-index"
    + "&consumerQueryVector=0.1,0.2,0.3"
    + "&delay=5000"
    + "&maxMessagesPerPoll=10")
    .log("Vector ID: ${header.CamelAwsS3VectorsVectorId}")
    .to("direct:process");
- route:
    from:
      uri: aws2-s3-vectors://my-bucket
      parameters:
        vectorIndexName: my-index
        consumerQueryVector: "0.1,0.2,0.3"
        delay: 5000
        maxMessagesPerPoll: 10
      steps:
        - log:
            message: "Vector ID: ${header.CamelAwsS3VectorsVectorId}"
        - to:
            uri: direct:process

Query with a Metadata Filter

  • Java

  • YAML

from("direct:search")
    .setBody(constant(Arrays.asList(0.15f, 0.25f, 0.35f)))
    .setHeader(AWS2S3VectorsConstants.METADATA_FILTER, constant("{\"genre\": {\"$eq\": \"drama\"}}"))
    .to("aws2-s3-vectors://my-bucket?operation=queryVectors&vectorIndexName=my-index")
    .log("Found ${body.size} similar drama vectors");
- route:
    from:
      uri: direct:search
      steps:
        - setBody:
            groovy: "[0.15, 0.25, 0.35]"
        - setHeader:
            name: CamelAwsS3VectorsMetadataFilter
            constant: '{"genre": {"$eq": "drama"}}'
        - to:
            uri: aws2-s3-vectors://my-bucket
            parameters:
              operation: queryVectors
              vectorIndexName: my-index
        - log:
            message: "Found ${body.size} similar drama vectors"

Create Index

  • Java

  • YAML

from("direct:createIndex")
    .setHeader(AWS2S3VectorsConstants.VECTOR_DIMENSIONS, constant(1536))
    .setHeader(AWS2S3VectorsConstants.DATA_TYPE, constant("float32"))
    .setHeader(AWS2S3VectorsConstants.DISTANCE_METRIC, constant("cosine"))
    .to("aws2-s3-vectors://my-bucket?operation=createVectorIndex&vectorIndexName=my-index");
- route:
    from:
      uri: direct:createIndex
      steps:
        - setHeader:
            name: CamelAwsS3VectorsVectorDimensions
            constant: 1536
        - setHeader:
            name: CamelAwsS3VectorsDataType
            constant: float32
        - setHeader:
            name: CamelAwsS3VectorsDistanceMetric
            constant: cosine
        - to:
            uri: aws2-s3-vectors://my-bucket
            parameters:
              operation: createVectorIndex
              vectorIndexName: my-index

Dependencies

Maven users will need to add the following dependency to their pom.xml.

pom.xml

<dependency>
    <groupId>org.apache.camel</groupId>
    <artifactId>camel-aws2-s3-vectors</artifactId>
    <version>x.x.x</version>
    <!-- use the same version as your Camel core version -->
</dependency>

where x.x.x is the version number of the latest Camel release.