Merge pull request #2867 from rockwotj/aws

aws/ai: support bedrock embeddings

docs/modules/components/pages/processors/aws_bedrock_embeddings.adoc

@@ -0,0 +1,199 @@
+= aws_bedrock_embeddings
+:type: processor
+:status: experimental
+:categories: ["AI"]
+
+
+
+////
+     THIS FILE IS AUTOGENERATED!
+
+     To make changes, edit the corresponding source file under:
+
+     https://github.com/redpanda-data/connect/tree/main/internal/impl/<provider>.
+
+     And:
+
+     https://github.com/redpanda-data/connect/tree/main/cmd/tools/docs_gen/templates/plugin.adoc.tmpl
+////
+
+// © 2024 Redpanda Data Inc.
+
+
+component_type_dropdown::[]
+
+
+Computes vector embeddings on text, using the AWS Bedrock API.
+
+Introduced in version 4.37.0.
+
+
+[tabs]
+======
+Common::
++
+--
+
+```yml
+# Common config fields, showing default values
+label: ""
+aws_bedrock_embeddings:
+  model: amazon.titan-embed-text-v1 # No default (required)
+  text: "" # No default (optional)
+```
+
+--
+Advanced::
++
+--
+
+```yml
+# All config fields, showing default values
+label: ""
+aws_bedrock_embeddings:
+  region: ""
+  endpoint: ""
+  credentials:
+    profile: ""
+    id: ""
+    secret: ""
+    token: ""
+    from_ec2_role: false
+    role: ""
+    role_external_id: ""
+  model: amazon.titan-embed-text-v1 # No default (required)
+  text: "" # No default (optional)
+```
+
+--
+======
+
+This processor sends text to your chosen large language model (LLM) and computes vector embeddings, using the AWS Bedrock API.
+For more information, see the https://docs.aws.amazon.com/bedrock/latest/userguide[AWS Bedrock documentation^].
+
+== Fields
+
+=== `region`
+
+The AWS region to target.
+
+
+*Type*: `string`
+
+*Default*: `""`
+
+=== `endpoint`
+
+Allows you to specify a custom endpoint for the AWS API.
+
+
+*Type*: `string`
+
+*Default*: `""`
+
+=== `credentials`
+
+Optional manual configuration of AWS credentials to use. More information can be found in xref:guides:cloud/aws.adoc[].
+
+
+*Type*: `object`
+
+
+=== `credentials.profile`
+
+A profile from `~/.aws/credentials` to use.
+
+
+*Type*: `string`
+
+*Default*: `""`
+
+=== `credentials.id`
+
+The ID of credentials to use.
+
+
+*Type*: `string`
+
+*Default*: `""`
+
+=== `credentials.secret`
+
+The secret for the credentials being used.
+[CAUTION]
+====
+This field contains sensitive information that usually shouldn't be added to a config directly, read our xref:configuration:secrets.adoc[secrets page for more info].
+====
+
+
+
+*Type*: `string`
+
+*Default*: `""`
+
+=== `credentials.token`
+
+The token for the credentials being used, required when using short term credentials.
+
+
+*Type*: `string`
+
+*Default*: `""`
+
+=== `credentials.from_ec2_role`
+
+Use the credentials of a host EC2 machine configured to assume https://docs.aws.amazon.com/IAM/latest/UserGuide/id_roles_use_switch-role-ec2.html[an IAM role associated with the instance^].
+
+
+*Type*: `bool`
+
+*Default*: `false`
+Requires version 4.2.0 or newer
+
+=== `credentials.role`
+
+A role ARN to assume.
+
+
+*Type*: `string`
+
+*Default*: `""`
+
+=== `credentials.role_external_id`
+
+An external ID to provide when assuming a role.
+
+
+*Type*: `string`
+
+*Default*: `""`
+
+=== `model`
+
+The model ID to use. For a full list see the https://docs.aws.amazon.com/bedrock/latest/userguide/model-ids.html[AWS Bedrock documentation^].
+
+
+*Type*: `string`
+
+
+```yml
+# Examples
+
+model: amazon.titan-embed-text-v1
+
+model: amazon.titan-embed-text-v2:0
+
+model: cohere.embed-english-v3
+
+model: cohere.embed-multilingual-v3
+```
+
+=== `text`
+
+The prompt you want to generate a response for. By default, the processor submits the entire payload as a string.
+
+
+*Type*: `string`
+
+
+

internal/impl/aws/enterprise/processor_bedrock.go → internal/impl/aws/enterprise/processor_bedrock_chat.go

@@ -23,111 +23,111 @@ import (
 )
 
 const (
-	bedpFieldModel        = "model"
-	bedpFieldUserPrompt   = "prompt"
-	bedpFieldSystemPrompt = "system_prompt"
-	bedpFieldMaxTokens    = "max_tokens"
-	bedpFieldTemp         = "stop"
-	bedpFieldStop         = "temperature"
-	bedpFieldTopP         = "top_p"
+	bedcpFieldModel        = "model"
+	bedcpFieldUserPrompt   = "prompt"
+	bedcpFieldSystemPrompt = "system_prompt"
+	bedcpFieldMaxTokens    = "max_tokens"
+	bedcpFieldTemp         = "stop"
+	bedcpFieldStop         = "temperature"
+	bedcpFieldTopP         = "top_p"
 )
 
 func init() {
-	err := service.RegisterProcessor("aws_bedrock_chat", newBedrockConfigSpec(), newBedrockProcessor)
+	err := service.RegisterProcessor("aws_bedrock_chat", newBedrockChatConfigSpec(), newBedrockChatProcessor)
 	if err != nil {
 		panic(err)
 	}
 }
 
-func newBedrockConfigSpec() *service.ConfigSpec {
+func newBedrockChatConfigSpec() *service.ConfigSpec {
 	return service.NewConfigSpec().
 		Summary("Generates responses to messages in a chat conversation, using the AWS Bedrock API.").
 		Description(`This processor sends prompts to your chosen large language model (LLM) and generates text from the responses, using the AWS Bedrock API.
 For more information, see the https://docs.aws.amazon.com/bedrock/latest/userguide[AWS Bedrock documentation^].`).
 		Categories("AI").
 		Version("4.34.0").
 		Fields(config.SessionFields()...).
-		Field(service.NewStringField(bedpFieldModel).
+		Field(service.NewStringField(bedcpFieldModel).
 			Examples("amazon.titan-text-express-v1", "anthropic.claude-3-5-sonnet-20240620-v1:0", "cohere.command-text-v14", "meta.llama3-1-70b-instruct-v1:0", "mistral.mistral-large-2402-v1:0").
 			Description("The model ID to use. For a full list see the https://docs.aws.amazon.com/bedrock/latest/userguide/model-ids.html[AWS Bedrock documentation^].")).
-		Field(service.NewStringField(bedpFieldUserPrompt).
+		Field(service.NewStringField(bedcpFieldUserPrompt).
 			Description("The prompt you want to generate a response for. By default, the processor submits the entire payload as a string.").
 			Optional()).
-		Field(service.NewStringField(bedpFieldSystemPrompt).
+		Field(service.NewStringField(bedcpFieldSystemPrompt).
 			Optional().
 			Description("The system prompt to submit to the AWS Bedrock LLM.")).
-		Field(service.NewIntField(bedpFieldMaxTokens).
+		Field(service.NewIntField(bedcpFieldMaxTokens).
 			Optional().
 			Description("The maximum number of tokens to allow in the generated response.").
 			LintRule(`root = this < 1 { ["field must be greater than or equal to 1"] }`)).
-		Field(service.NewFloatField(bedpFieldTemp).
+		Field(service.NewFloatField(bedcpFieldTemp).
 			Optional().
 			Description("The likelihood of the model selecting higher-probability options while generating a response. A lower value makes the model omre likely to choose higher-probability options, while a higher value makes the model more likely to choose lower-probability options.").
 			LintRule(`root = if this < 0 || this > 1 { ["field must be between 0.0-1.0"] }`)).
-		Field(service.NewStringListField(bedpFieldStop).
+		Field(service.NewStringListField(bedcpFieldStop).
 			Optional().
 			Advanced().
 			Description("A list of stop sequences. A stop sequence is a sequence of characters that causes the model to stop generating the response.")).
-		Field(service.NewFloatField(bedpFieldTopP).
+		Field(service.NewFloatField(bedcpFieldTopP).
 			Optional().
 			Advanced().
 			Description("The percentage of most-likely candidates that the model considers for the next token. For example, if you choose a value of 0.8, the model selects from the top 80% of the probability distribution of tokens that could be next in the sequence. ").
 			LintRule(`root = if this < 0 || this > 1 { ["field must be between 0.0-1.0"] }`))
 }
 
-func newBedrockProcessor(conf *service.ParsedConfig, mgr *service.Resources) (service.Processor, error) {
+func newBedrockChatProcessor(conf *service.ParsedConfig, mgr *service.Resources) (service.Processor, error) {
 	aconf, err := aws.GetSession(context.Background(), conf)
 	if err != nil {
 		return nil, err
 	}
 	client := bedrockruntime.NewFromConfig(aconf)
-	model, err := conf.FieldString(bedpFieldModel)
+	model, err := conf.FieldString(bedcpFieldModel)
 	if err != nil {
 		return nil, err
 	}
-	p := &bedrockProcessor{
+	p := &bedrockChatProcessor{
 		client: client,
 		model:  model,
 	}
-	if conf.Contains(bedpFieldUserPrompt) {
-		pf, err := conf.FieldInterpolatedString(bedpFieldUserPrompt)
+	if conf.Contains(bedcpFieldUserPrompt) {
+		pf, err := conf.FieldInterpolatedString(bedcpFieldUserPrompt)
 		if err != nil {
 			return nil, err
 		}
 		p.userPrompt = pf
 	}
-	if conf.Contains(bedpFieldSystemPrompt) {
-		pf, err := conf.FieldInterpolatedString(bedpFieldSystemPrompt)
+	if conf.Contains(bedcpFieldSystemPrompt) {
+		pf, err := conf.FieldInterpolatedString(bedcpFieldSystemPrompt)
 		if err != nil {
 			return nil, err
 		}
 		p.systemPrompt = pf
 	}
-	if conf.Contains(bedpFieldMaxTokens) {
-		v, err := conf.FieldInt(bedpFieldMaxTokens)
+	if conf.Contains(bedcpFieldMaxTokens) {
+		v, err := conf.FieldInt(bedcpFieldMaxTokens)
 		if err != nil {
 			return nil, err
 		}
 		mt := int32(v)
 		p.maxTokens = &mt
 	}
-	if conf.Contains(bedpFieldTemp) {
-		v, err := conf.FieldFloat(bedpFieldTemp)
+	if conf.Contains(bedcpFieldTemp) {
+		v, err := conf.FieldFloat(bedcpFieldTemp)
 		if err != nil {
 			return nil, err
 		}
 		t := float32(v)
 		p.temp = &t
 	}
-	if conf.Contains(bedpFieldStop) {
-		stop, err := conf.FieldStringList(bedpFieldStop)
+	if conf.Contains(bedcpFieldStop) {
+		stop, err := conf.FieldStringList(bedcpFieldStop)
 		if err != nil {
 			return nil, err
 		}
 		p.stop = stop
 	}
-	if conf.Contains(bedpFieldTopP) {
-		v, err := conf.FieldFloat(bedpFieldTopP)
+	if conf.Contains(bedcpFieldTopP) {
+		v, err := conf.FieldFloat(bedcpFieldTopP)
 		if err != nil {
 			return nil, err
 		}
@@ -137,7 +137,7 @@ func newBedrockProcessor(conf *service.ParsedConfig, mgr *service.Resources) (se
 	return p, nil
 }
 
-type bedrockProcessor struct {
+type bedrockChatProcessor struct {
 	client *bedrockruntime.Client
 	model  string
 
@@ -149,7 +149,7 @@ type bedrockProcessor struct {
 	topP         *float32
 }
 
-func (b *bedrockProcessor) Process(ctx context.Context, msg *service.Message) (service.MessageBatch, error) {
+func (b *bedrockChatProcessor) Process(ctx context.Context, msg *service.Message) (service.MessageBatch, error) {
 	prompt, err := b.computePrompt(msg)
 	if err != nil {
 		return nil, err
@@ -176,7 +176,7 @@ func (b *bedrockProcessor) Process(ctx context.Context, msg *service.Message) (s
 	if b.systemPrompt != nil {
 		prompt, err := b.systemPrompt.TryString(msg)
 		if err != nil {
-			return nil, fmt.Errorf("unable to interpolate `%s`: %w", bedpFieldSystemPrompt, err)
+			return nil, fmt.Errorf("unable to interpolate `%s`: %w", bedcpFieldSystemPrompt, err)
 		}
 		input.System = []bedrocktypes.SystemContentBlock{
 			&bedrocktypes.SystemContentBlockMemberText{Value: prompt},
@@ -204,7 +204,7 @@ func (b *bedrockProcessor) Process(ctx context.Context, msg *service.Message) (s
 	return service.MessageBatch{out}, nil
 }
 
-func (b *bedrockProcessor) computePrompt(msg *service.Message) (string, error) {
+func (b *bedrockChatProcessor) computePrompt(msg *service.Message) (string, error) {
 	if b.userPrompt != nil {
 		return b.userPrompt.TryString(msg)
 	}
@@ -218,6 +218,6 @@ func (b *bedrockProcessor) computePrompt(msg *service.Message) (string, error) {
 	return string(buf), nil
 }
 
-func (b *bedrockProcessor) Close(ctx context.Context) error {
+func (b *bedrockChatProcessor) Close(ctx context.Context) error {
 	return nil
 }