[Backport main] Simplify RAG presets; add bulk API details

common/constants.ts

@@ -250,6 +250,8 @@ export const BULK_API_DOCS_LINK =
   'https://opensearch.org/docs/latest/api-reference/document-apis/bulk/';
 export const SEARCH_PIPELINE_DOCS_LINK =
   'https://opensearch.org/docs/latest/search-plugins/search-pipelines/using-search-pipeline/';
+export const ML_RESPONSE_PROCESSOR_EXAMPLE_DOCS_LINK =
+  'https://opensearch.org/docs/latest/search-plugins/search-pipelines/ml-inference-search-response/#example-externally-hosted-text-embedding-model';
 
 /**
  * Text chunking algorithm constants

public/general_components/results/ml_response.tsx

@@ -0,0 +1,60 @@
+/*
+ * Copyright OpenSearch Contributors
+ * SPDX-License-Identifier: Apache-2.0
+ */
+
+import React from 'react';
+import { isEmpty } from 'lodash';
+import {
+  EuiCode,
+  EuiCodeEditor,
+  EuiEmptyPrompt,
+  EuiLink,
+  EuiSpacer,
+  EuiText,
+} from '@elastic/eui';
+import {
+  customStringify,
+  ML_RESPONSE_PROCESSOR_EXAMPLE_DOCS_LINK,
+} from '../../../common';
+
+interface MLResponseProps {
+  mlResponse: {};
+}
+
+/**
+ * Small component to render the ML response within a raw search response.
+ */
+export function MLResponse(props: MLResponseProps) {
+  return (
+    <>
+      <EuiSpacer size="s" />
+      <EuiText size="s">
+        Showing results stored in <EuiCode>ext.ml_inference</EuiCode> from the
+        search response.{' '}
+        <EuiLink href={ML_RESPONSE_PROCESSOR_EXAMPLE_DOCS_LINK} target="_blank">
+          See an example
+        </EuiLink>
+      </EuiText>
+      <EuiSpacer size="m" />
+      {isEmpty(props.mlResponse) ? (
+        <EuiEmptyPrompt title={<h2>No response found</h2>} titleSize="s" />
+      ) : (
+        <EuiCodeEditor
+          mode="json"
+          theme="textmate"
+          width="100%"
+          height="100%"
+          value={customStringify(props.mlResponse)}
+          readOnly={true}
+          setOptions={{
+            fontSize: '12px',
+            autoScrollEditorIntoView: true,
+            wrap: true,
+          }}
+          tabSize={2}
+        />
+      )}
+    </>
+  );
+}

public/general_components/results/results.tsx

@@ -4,6 +4,7 @@
  */
 
 import React, { useState } from 'react';
+import { get } from 'lodash';
 import {
   EuiPanel,
   EuiFlexGroup,
@@ -13,6 +14,7 @@ import {
 import { SearchResponse } from '../../../common';
 import { ResultsTable } from './results_table';
 import { ResultsJSON } from './results_json';
+import { MLResponse } from './ml_response';
 
 interface ResultsProps {
   response: SearchResponse;
@@ -21,6 +23,7 @@ interface ResultsProps {
 enum VIEW {
   HITS_TABLE = 'hits_table',
   RAW_JSON = 'raw_json',
+  ML_RESPONSE = 'ml_response',
 }
 
 /**
@@ -55,6 +58,10 @@ export function Results(props: ResultsProps) {
                 id: VIEW.RAW_JSON,
                 label: 'Raw JSON',
               },
+              {
+                id: VIEW.ML_RESPONSE,
+                label: 'ML response',
+              },
             ]}
             idSelected={selectedView}
             onChange={(id) => setSelectedView(id as VIEW)}
@@ -69,9 +76,18 @@ export function Results(props: ResultsProps) {
             {selectedView === VIEW.RAW_JSON && (
               <ResultsJSON response={props.response} />
             )}
+            {selectedView === VIEW.ML_RESPONSE && (
+              <MLResponse
+                mlResponse={getMLResponseFromSearchResponse(props.response)}
+              />
+            )}
           </>
         </EuiFlexItem>
       </EuiFlexGroup>
     </EuiPanel>
   );
 }
+
+function getMLResponseFromSearchResponse(searchResponse: SearchResponse): {} {
+  return get(searchResponse, 'ext.ml_inference', {});
+}

public/pages/workflow_detail/workflow_inputs/ingest_inputs/bulk_popover_content.tsx

@@ -0,0 +1,49 @@
+/*
+ * Copyright OpenSearch Contributors
+ * SPDX-License-Identifier: Apache-2.0
+ */
+
+import React from 'react';
+import {
+  EuiText,
+  EuiFlexGroup,
+  EuiFlexItem,
+  EuiLink,
+  EuiCodeBlock,
+  EuiPopoverTitle,
+} from '@elastic/eui';
+import { BULK_API_DOCS_LINK } from '../../../../../common';
+
+interface BulkPopoverContentProps {
+  indexName: string;
+}
+
+/**
+ * A basic component containing details about the bulk API and link to documentation.
+ * Provides a partially-complete example, dynamically populated based on an index name.
+ */
+export function BulkPopoverContent(props: BulkPopoverContentProps) {
+  return (
+    <EuiFlexItem style={{ width: '40vw' }}>
+      <EuiPopoverTitle>Ingest additional data</EuiPopoverTitle>
+      <EuiFlexGroup direction="column">
+        <EuiFlexItem grow={false}>
+          <EuiText>
+            You can ingest additional bulk data into the same index using the
+            Bulk API.{' '}
+            <EuiLink href={BULK_API_DOCS_LINK} target="_blank">
+              Learn more
+            </EuiLink>
+          </EuiText>
+        </EuiFlexItem>
+        <EuiFlexItem grow={false}>
+          <EuiCodeBlock fontSize="m" isCopyable={true}>
+            {`POST ${props.indexName}/_bulk
+{ "index": { "_index": "${props.indexName}", "_id": //YOUR DOC ID// } }
+{ //INSERT YOUR DOCUMENTS// }`}
+          </EuiCodeBlock>
+        </EuiFlexItem>
+      </EuiFlexGroup>
+    </EuiFlexItem>
+  );
+}

public/pages/workflow_detail/workflow_inputs/ingest_inputs/source_data.tsx

@@ -14,6 +14,7 @@ import {
   EuiCodeBlock,
   EuiSmallButtonEmpty,
   EuiEmptyPrompt,
+  EuiPopover,
 } from '@elastic/eui';
 import {
   MapEntry,
@@ -25,6 +26,7 @@ import {
   toFormattedDate,
 } from '../../../../../common';
 import { SourceDataModal } from './source_data_modal';
+import { BulkPopoverContent } from './bulk_popover_content';
 
 interface SourceDataProps {
   workflow: Workflow | undefined;
@@ -54,6 +56,9 @@ export function SourceData(props: SourceDataProps) {
   // edit modal state
   const [isEditModalOpen, setIsEditModalOpen] = useState<boolean>(false);
 
+  // bulk API popover state
+  const [bulkPopoverOpen, setBulkPopoverOpen] = useState<boolean>(false);
+
   // hook to listen when the docs form value changes.
   useEffect(() => {
     if (values?.ingest?.docs) {
@@ -133,11 +138,49 @@ export function SourceData(props: SourceDataProps) {
           </EuiFlexGroup>
         </EuiFlexItem>
         {props.lastIngested !== undefined && (
-          <EuiFlexItem grow={false}>
-            <EuiText size="s" color="subdued">
-              {`Last ingested: ${toFormattedDate(props.lastIngested)}`}
-            </EuiText>
-          </EuiFlexItem>
+          <>
+            <EuiFlexItem grow={false}>
+              <EuiText size="s">
+                {`Last ingested: ${toFormattedDate(props.lastIngested)}`}
+              </EuiText>
+            </EuiFlexItem>
+            <EuiFlexItem>
+              <EuiFlexGroup
+                direction="row"
+                gutterSize="none"
+                justifyContent="flexStart"
+                style={{ marginTop: '-8px' }}
+              >
+                <EuiFlexItem grow={false}>
+                  <EuiText size="s">
+                    Ingest additional data with the bulk API.
+                  </EuiText>
+                </EuiFlexItem>
+                <EuiFlexItem grow={false}>
+                  <EuiPopover
+                    isOpen={bulkPopoverOpen}
+                    initialFocus={false}
+                    anchorPosition="downCenter"
+                    closePopover={() => setBulkPopoverOpen(false)}
+                    button={
+                      <EuiSmallButtonEmpty
+                        style={{ marginTop: '-4px' }}
+                        onClick={() => {
+                          setBulkPopoverOpen(!bulkPopoverOpen);
+                        }}
+                      >
+                        Learn more
+                      </EuiSmallButtonEmpty>
+                    }
+                  >
+                    <BulkPopoverContent
+                      indexName={getIn(values, 'ingest.index.name')}
+                    />
+                  </EuiPopover>
+                </EuiFlexItem>
+              </EuiFlexGroup>
+            </EuiFlexItem>
+          </>
         )}
         {docsPopulated ? (
           <>

public/pages/workflows/new_workflow/quick_configure_modal.tsx

@@ -741,7 +741,8 @@ function updateRAGSearchResponseProcessors(
   llmInterface: ModelInterface | undefined
 ): WorkflowConfig {
   config.search.enrichResponse.processors.forEach((processor, idx) => {
-    // prefill ML inference
+    // prefill ML inference. By default, store the inference results
+    // under the `ext.ml_inference` response body.
     if (processor.type === PROCESSOR_TYPE.ML) {
       config.search.enrichResponse.processors[idx].fields.forEach((field) => {
         if (field.id === 'model' && fields.llmId) {
@@ -784,7 +785,7 @@ function updateRAGSearchResponseProcessors(
                 ...outputMap[0],
                 value: {
                   transformType: TRANSFORM_TYPE.FIELD,
-                  value: fields.llmResponseField,
+                  value: `ext.ml_inference.${fields.llmResponseField}`,
                 },
               };
             } else {

public/pages/workflows/new_workflow/utils.ts

@@ -4,7 +4,6 @@
  */
 
 import {
-  CollapseProcessor,
   MLIngestProcessor,
   MLSearchRequestProcessor,
   MLSearchResponseProcessor,
@@ -253,7 +252,6 @@ export function fetchRAGMetadata(version: string): UIState {
   baseState.config.search.request.value = customStringify(FETCH_ALL_QUERY);
   baseState.config.search.enrichResponse.processors = [
     new MLSearchResponseProcessor().toObj(),
-    new CollapseProcessor().toObj(),
   ];
   return baseState;
 }
@@ -278,7 +276,6 @@ export function fetchVectorSearchWithRAGMetadata(version: string): UIState {
   ];
   baseState.config.search.enrichResponse.processors = [
     new MLSearchResponseProcessor().toObj(),
-    new CollapseProcessor().toObj(),
   ];
   return baseState;
 }