addressed comments and cleaned up defaults/templates

Signed-off-by: Amit Galitzky <[email protected]>

src/main/java/org/opensearch/flowframework/common/DefaultUseCases.java

@@ -28,7 +28,7 @@ public enum DefaultUseCases {
     COHERE_EMBEDDING_MODEL_DEPLOY(
         "cohere-embedding_model_deploy",
         "defaults/cohere-embedding-defaults.json",
-        "substitutionTemplates/deploy-remote-model-template-extra-params.json"
+        "substitutionTemplates/deploy-remote-model-extra-params-template.json"
     ),
     /** defaults file and substitution ready template for Bedrock Titan embedding model */
     BEDROCK_TITAN_EMBEDDING_MODEL_DEPLOY(

src/main/java/org/opensearch/flowframework/util/ParseUtils.java

@@ -59,6 +59,7 @@ public class ParseUtils {
     // Matches ${{ foo.bar }} (whitespace optional) with capturing groups 1=foo, 2=bar
     // private static final Pattern SUBSTITUTION_PATTERN = Pattern.compile("\\$\\{\\{\\s*(.+)\\.(.+?)\\s*\\}\\}");
     private static final Pattern SUBSTITUTION_PATTERN = Pattern.compile("\\$\\{\\{\\s*([\\w_]+)\\.([\\w_]+)\\s*\\}\\}");
+    private static final Pattern JSON_ARRAY_DOUBLE_QUOTES_PATTERN = Pattern.compile("\"\\[(.*?)]\"");
 
     private ParseUtils() {}
 
@@ -70,7 +71,7 @@ private ParseUtils() {}
      * @param json the json string
      * @return The XContent parser for the json string
      * @throws IOException on failure to create the parser
-    */
+     */
     public static XContentParser jsonToParser(String json) throws IOException {
         XContentParser parser = JsonXContent.jsonXContent.createParser(
             NamedXContentRegistry.EMPTY,
@@ -104,7 +105,7 @@ public static String resourceToString(String path) throws IOException {
      * Builds an XContent object representing a map of String keys to String values.
      *
      * @param xContentBuilder An XContent builder whose position is at the start of the map object to build
-     * @param map A map as key-value String pairs.
+     * @param map             A map as key-value String pairs.
      * @throws IOException on a build failure
      */
     public static void buildStringToStringMap(XContentBuilder xContentBuilder, Map<?, ?> map) throws IOException {
@@ -119,7 +120,7 @@ public static void buildStringToStringMap(XContentBuilder xContentBuilder, Map<?
      * Builds an XContent object representing a map of String keys to Object values.
      *
      * @param xContentBuilder An XContent builder whose position is at the start of the map object to build
-     * @param map A map as key-value String to Object.
+     * @param map             A map as key-value String to Object.
      * @throws IOException on a build failure
      */
     public static void buildStringToObjectMap(XContentBuilder xContentBuilder, Map<?, ?> map) throws IOException {
@@ -138,7 +139,7 @@ public static void buildStringToObjectMap(XContentBuilder xContentBuilder, Map<?
      * Builds an XContent object representing a LLMSpec.
      *
      * @param xContentBuilder An XContent builder whose position is at the start of the map object to build
-     * @param llm LLMSpec
+     * @param llm             LLMSpec
      * @throws IOException on a build failure
      */
     public static void buildLLMMap(XContentBuilder xContentBuilder, LLMSpec llm) throws IOException {
@@ -171,6 +172,7 @@ public static Map<String, String> parseStringToStringMap(XContentParser parser)
      * Parses an XContent object representing a map of String keys to Object values.
      * The Object value here can either be a string or a map
      * If an array is found in the given parser we conver the array to a string representation of the array
+     *
      * @param parser An XContent parser whose position is at the start of the map object to parse
      * @return A map as identified by the key-value pairs in the XContent
      * @throws IOException on a parse failure
@@ -189,10 +191,13 @@ public static Map<String, Object> parseStringToObjectMap(XContentParser parser)
                 // Handle array: convert it to a string representation
                 List<String> elements = new ArrayList<>();
                 while (parser.nextToken() != XContentParser.Token.END_ARRAY) {
-                    elements.add("\"" + parser.text() + "\"");  // Adding escaped quotes around each element
+                    if (parser.currentToken().equals(XContentParser.Token.VALUE_NUMBER)) {
+                        elements.add(String.valueOf(parser.numberValue()));  // If number value don't add escaping quotes
+                    } else {
+                        elements.add("\"" + parser.text() + "\"");  // Adding escaped quotes around each element
+                    }
                 }
-                String arrayString = "[" + String.join(", ", elements) + "]";
-                map.put(fieldName, arrayString);
+                map.put(fieldName, elements.toString());
             } else {
                 // Otherwise, parse it as a string
                 map.put(fieldName, parser.text());
@@ -220,6 +225,7 @@ public static Instant parseInstant(XContentParser parser) throws IOException {
      * (e.g., john||own_index,testrole|__user__, no backend role so you see two verticle line after john.).
      * This is the user string format used internally in the OPENSEARCH_SECURITY_USER_INFO_THREAD_CONTEXT and may be
      * parsed using User.parse(string).
+     *
      * @param client Client containing user info. A public API request will fill in the user info in the thread context.
      * @return parsed user object
      */
@@ -233,7 +239,7 @@ public static User getUserContext(Client client) {
      * Creates a XContentParser from a given Registry
      *
      * @param xContentRegistry main registry for serializable content
-     * @param bytesReference given bytes to be parsed
+     * @param bytesReference   given bytes to be parsed
      * @return bytesReference of {@link java.time.Instant}
      * @throws IOException IOException if content can't be parsed correctly
      */
@@ -244,7 +250,8 @@ public static XContentParser createXContentParserFromRegistry(NamedXContentRegis
 
     /**
      * Generates a string to string Map
-     * @param map content map
+     *
+     * @param map       content map
      * @param fieldName fieldName
      * @return instance of the map
      */
@@ -260,15 +267,15 @@ public static Map<String, String> getStringToStringMap(Object map, String fieldN
      * Creates a map containing the specified input keys, with values derived from template data or previous node
      * output.
      *
-     * @param requiredInputKeys A set of keys that must be present, or will cause an exception to be thrown
-     * @param optionalInputKeys A set of keys that may be present, or will be absent in the returned map
-     * @param currentNodeInputs Input params and content for this node, from workflow parsing
-     * @param outputs WorkflowData content of previous steps
+     * @param requiredInputKeys  A set of keys that must be present, or will cause an exception to be thrown
+     * @param optionalInputKeys  A set of keys that may be present, or will be absent in the returned map
+     * @param currentNodeInputs  Input params and content for this node, from workflow parsing
+     * @param outputs            WorkflowData content of previous steps
      * @param previousNodeInputs Input params for this node that come from previous steps
-     * @param params Params that came from REST path
+     * @param params             Params that came from REST path
      * @return A map containing the requiredInputKeys with their corresponding values,
-     *         and optionalInputKeys with their corresponding values if present.
-     *         Throws a {@link FlowFrameworkException} if a required key is not present.
+     * and optionalInputKeys with their corresponding values if present.
+     * Throws a {@link FlowFrameworkException} if a required key is not present.
      */
     public static Map<String, Object> getInputsFromPreviousSteps(
         Set<String> requiredInputKeys,
@@ -357,9 +364,10 @@ public static Map<String, Object> getInputsFromPreviousSteps(
 
     /**
      * Executes substitution on the given value by looking at any matching values in either the ouputs or params map
-     * @param value the Object that will have the substitution done on
+     *
+     * @param value   the Object that will have the substitution done on
      * @param outputs potential location of values to be substituted in
-     * @param params potential location of values to be subsituted in
+     * @param params  potential location of values to be subsituted in
      * @return the substituted object back
      */
     public static Object conditionallySubstitute(Object value, Map<String, WorkflowData> outputs, Map<String, String> params) {
@@ -403,6 +411,7 @@ public static Object conditionallySubstitute(Object value, Map<String, WorkflowD
 
     /**
      * Generates a string based on an arbitrary String to object map using Jackson
+     *
      * @param map content map
      * @return instance of the string
      * @throws JsonProcessingException JsonProcessingException from Jackson for issues processing map
@@ -415,6 +424,7 @@ public static String parseArbitraryStringToObjectMapToString(Map<String, Object>
 
     /**
      * Generates a String to String map based on a Json File
+     *
      * @param path file path
      * @return instance of the string
      * @throws JsonProcessingException JsonProcessingException from Jackson for issues processing map
@@ -430,15 +440,21 @@ public static Map<String, String> parseJsonFileToStringToStringMap(String path)
      * (e.g.  "[\"text\", \"hello\"]" to "["text", "hello"]"), this is needed for processors that take in string arrays,
      * This also removes the quotations around the array making the array valid to consume
      * (e.g. "weights": "[0.7, 0.3]" to "weights": [0.7, 0.3])
+     *
      * @param input The inputString given to be transformed
      * @return the transformed string
      */
     public static String removingBackslashesAndQuotesInArrayInJsonString(String input) {
-        return Pattern.compile("\"\\[(.*?)]\"").matcher(input).replaceAll(matchResult -> {
+        Matcher matcher = JSON_ARRAY_DOUBLE_QUOTES_PATTERN.matcher(input);
+        StringBuffer result = new StringBuffer();
+        while (matcher.find()) {
             // Extract matched content and remove backslashes before quotes
-            String withoutEscapes = matchResult.group(1).replaceAll("\\\\\"", "\"");
+            String withoutEscapes = matcher.group(1).replaceAll("\\\\\"", "\"");
             // Return the transformed string with the brackets but without the outer quotes
-            return "[" + withoutEscapes + "]";
-        });
+            matcher.appendReplacement(result, "[" + withoutEscapes + "]");
+        }
+        // Append remaining input after the last match
+        matcher.appendTail(result);
+        return result.toString();
     }
 }

src/main/resources/defaults/bedrock-titan-embedding-defaults.json

@@ -4,7 +4,6 @@
     "create_connector.name": "Amazon Bedrock Connector: embedding",
     "create_connector.description": "The connector to bedrock Titan embedding model",
     "create_connector.region": "us-east-1",
-    "create_connector.endpoint": "api.openai.com",
     "create_connector.credential.access_key": "123",
     "create_connector.credential.secret_key": "123",
     "create_connector.credential.session_token": "123",

src/main/resources/defaults/bedrock-titan-mulitmodal-defaults.json → src/main/resources/defaults/bedrock-titan-multimodal-defaults.json

@@ -1,11 +1,10 @@
 {
     "template.name": "deploy-bedrock-titan-multimodal-embedding-model",
-    "template.description": "deploying Amazon Bedrock Titan multimodal embedding model ",
+    "template.description": "Deploying Amazon Bedrock Titan multimodal embedding model ",
     "create_connector.name": "Amazon Bedrock Connector: multi-modal embedding",
     "create_connector.description": "The connector to bedrock Titan multi-modal embedding model",
     "create_connector.region": "us-east-1",
-    "create_connector.input_docs_processed_step_size": 2,
-    "create_connector.endpoint": "api.openai.com",
+    "create_connector.input_docs_processed_step_size": "2",
     "create_connector.credential.access_key": "123",
     "create_connector.credential.secret_key": "123",
     "create_connector.credential.session_token": "123",

src/main/resources/defaults/cohere-chat-defaults.json

@@ -1,6 +1,6 @@
 {
     "template.name": "deploy-cohere-chat-model",
-    "template.description": "deploying cohere chat model",
+    "template.description": "Deploying a Cohere chat model",
     "create_connector.name": "Cohere Chat Model",
     "create_connector.description": "The connector to Cohere's public chat API",
     "create_connector.protocol": "http",

src/main/resources/defaults/cohere-embedding-defaults.json

@@ -1,6 +1,6 @@
 {
     "template.name": "deploy-cohere-model",
-    "template.description": "deploying cohere embedding model",
+    "template.description": "Deploying a Cohere embedding model",
     "create_connector.name": "cohere-embedding-connector",
     "create_connector.description": "The connector to Cohere's public embed API",
     "create_connector.protocol": "http",

src/main/resources/defaults/cohere-embedding-semantic-search-defaults.json

@@ -1,6 +1,6 @@
 {
     "template.name": "semantic search with cohere embedding",
-    "template.description": "Setting up semantic search, with cohere embedding model",
+    "template.description": "Setting up semantic search, with a Cohere embedding model",
     "create_connector.name": "cohere-embedding-connector",
     "create_connector.description": "The connector to Cohere's public embed API",
     "create_connector.protocol": "http",

src/main/resources/defaults/conversational-search-defaults.json

@@ -1,6 +1,6 @@
 {
     "template.name": "deploy-cohere-chat-model",
-    "template.description": "deploying cohere chat model",
+    "template.description": "A template to deploy a Cohere chat model",
     "create_connector.name": "Cohere Chat Model",
     "create_connector.description": "The connector to Cohere's public chat API",
     "create_connector.protocol": "http",
@@ -13,7 +13,7 @@
     "register_remote_model.description": "cohere-chat-model",
     "create_search_pipeline.pipeline_id": "rag-pipeline",
     "create_search_pipeline.retrieval_augmented_generation.tag": "openai_pipeline_demo",
-    "create_search_pipeline.retrieval_augmented_generation.description": "Demo pipeline Using cohere Connector",
+    "create_search_pipeline.retrieval_augmented_generation.description": "Demo pipeline using a Cohere chat model",
     "create_search_pipeline.retrieval_augmented_generation.context_field_list": "[\"text\"]",
     "create_search_pipeline.retrieval_augmented_generation.system_prompt": "You are a helpful assistant",
     "create_search_pipeline.retrieval_augmented_generation.user_instructions": "Generate a concise and informative answer in less than 100 words for the given question"

src/main/resources/defaults/local-sparse-search-biencoder-defaults.json

@@ -1,6 +1,6 @@
 {
     "template.name": "local-model-neural-sparse-search",
-    "template.description": "setting up neural sparse search with local model",
+    "template.description": "Setting up neural sparse search with pretrained local model",
     "register_local_sparse_encoding_model.name": "amazon/neural-sparse/opensearch-neural-sparse-encoding-v1",
     "register_local_sparse_encoding_model.description": "This is a neural sparse encoding model",
     "register_local_sparse_encoding_model.model_format": "TORCH_SCRIPT",

src/main/resources/defaults/multi-modal-search-defaults.json

@@ -4,12 +4,12 @@
     "create_ingest_pipeline.pipeline_id": "nlp-multimodal-ingest-pipeline",
     "create_ingest_pipeline.description": "A text/image embedding pipeline",
     "create_ingest_pipeline.model_id": "123",
-    "create_ingest_pipeline.embedding": "vector_embedding",
+    "text_image_embedding.embedding": "vector_embedding",
     "text_image_embedding.field_map.text": "image_description",
     "text_image_embedding.field_map.image": "image_binary",
     "create_index.name": "my-multimodal-nlp-index",
-    "create_index.settings.number_of_shards": 2,
-    "text_image_embedding.field_map.output.dimension": 1024,
+    "create_index.settings.number_of_shards": "2",
+    "text_image_embedding.field_map.output.dimension": "1024",
     "create_index.mappings.method.engine": "lucene",
     "create_index.mappings.method.name": "hnsw"
 }

src/main/resources/defaults/multimodal-search-bedrock-titan-defaults.json

@@ -17,7 +17,7 @@
     "register_remote_model.description": "bedrock-multi-modal-embedding-model",
     "create_ingest_pipeline.pipeline_id": "nlp-multimodal-ingest-pipeline",
     "create_ingest_pipeline.description": "A text/image embedding pipeline",
-    "text_image_embedding.create_ingest_pipeline.embedding": "vector_embedding",
+    "text_image_embedding.embedding": "vector_embedding",
     "text_image_embedding.field_map.text": "image_description",
     "text_image_embedding.field_map.image": "image_binary",
     "create_index.name": "my-multimodal-nlp-index",

src/main/resources/defaults/openai-chat-defaults.json

@@ -1,6 +1,6 @@
 {
       "template.name": "deploy-openai-chat-model",
-      "template.description": "deploying openAI chat model",
+      "template.description": "Deploying an OpenAI chat model",
       "create_connector.name": "OpenAI Chat Connector",
       "create_connector.description": "Connector to public OpenAI model",
       "create_connector.protocol": "http",

src/main/resources/defaults/openai-embedding-defaults.json

@@ -1,6 +1,6 @@
 {
     "template.name": "deploy-openai-model",
-    "template.description": "deploying openAI embedding model",
+    "template.description": "Deploying an OpenAI embedding model",
     "create_connector.name": "OpenAI-embedding-connector",
     "create_connector.description": "Connector to public OpenAI model",
     "create_connector.protocol": "http",

src/main/resources/substitutionTemplates/deploy-remote-bedrock-model-template.json

@@ -26,7 +26,7 @@
               "input_docs_processed_step_size": "${{create_connector.input_docs_processed_step_size}}"
             },
             "credential": {
-              "access_ key": "${{create_connector.credential.access_key}}",
+              "access_key": "${{create_connector.credential.access_key}}",
               "secret_key": "${{create_connector.credential.secret_key}}",
               "session_token": "${{create_connector.credential.session_token}}"
             },

src/main/resources/substitutionTemplates/deploy-remote-model-chat-template.json

@@ -44,7 +44,7 @@
           "id": "register_model",
           "type": "register_remote_model",
           "previous_node_inputs": {
-            "create_connector_step_1": "parameters"
+            "create_connector": "parameters"
           },
           "user_inputs": {
             "name": "${{register_remote_model.name}}",
@@ -56,7 +56,7 @@
           "id": "deploy_model",
           "type": "deploy_model",
           "previous_node_inputs": {
-            "register_model_1": "model_id"
+            "register_model": "model_id"
           }
         }
       ],

src/main/resources/substitutionTemplates/deploy-remote-model-template.json

@@ -46,7 +46,7 @@
           "id": "register_model",
           "type": "register_remote_model",
           "previous_node_inputs": {
-            "create_connector_step_1": "parameters"
+            "create_connector": "parameters"
           },
           "user_inputs": {
             "name": "${{register_remote_model.name}}",
@@ -58,7 +58,7 @@
           "id": "deploy_model",
           "type": "deploy_model",
           "previous_node_inputs": {
-            "register_model_1": "model_id"
+            "register_model": "model_id"
           }
         }
       ],

src/main/resources/substitutionTemplates/multi-modal-search-template.json

@@ -23,7 +23,7 @@
                 {
                   "text_image_embedding": {
                     "model_id": "${{create_ingest_pipeline.model_id}}",
-                    "embedding": "${{create_ingest_pipeline.embedding}}",
+                    "embedding": "${{text_image_embedding.embedding}}",
                     "field_map": {
                       "text": "${{text_image_embedding.field_map.text}}",
                       "image": "${{text_image_embedding.field_map.image}}"
@@ -53,7 +53,7 @@
                     "id": {
                       "type": "text"
                     },
-                    "${{text_embedding.field_map.output}}": {
+                    "${{text_image_embedding.embedding}}": {
                       "type": "knn_vector",
                       "dimension": "${{text_image_embedding.field_map.output.dimension}}",
                       "method": {