Facet counts and aggregations for Vector search

frontends/main/src/app-pages/SearchPage/SearchPage.test.tsx

@@ -152,50 +152,6 @@ describe("SearchPage", () => {
     },
   )
 
-  test("Vector Hybrid Search passes correct params and hides count", async () => {
-    setMockApiResponses({
-      search: {
-        count: 700,
-        metadata: {
-          aggregations: {
-            resource_type_group: [{ key: "course", doc_count: 100 }],
-          },
-          suggestions: [],
-        },
-        results: factories.learningResources.resources({ count: 5 }).results,
-      },
-    })
-
-    // Authenticate as path editor (admin)
-    setMockResponse.get(urls.userMe.get(), {
-      is_learning_path_editor: true,
-      is_authenticated: true,
-    })
-
-    renderWithProviders(<SearchPage />, { url: "?vector_search=true&q=test" })
-
-    await waitFor(() => {
-      const call = makeRequest.mock.calls.find(([_method, url]) => {
-        return url.includes(urls.search.vectorResources())
-      })
-      expect(call).toBeDefined()
-    })
-
-    const call = makeRequest.mock.calls.find(([_method, url]) =>
-      url.includes(urls.search.vectorResources()),
-    )
-    invariant(call)
-    const fullUrl = new URL(call[1], "http://mit.edu")
-    const apiSearchParams = fullUrl.searchParams
-
-    expect(apiSearchParams.get("hybrid_search")).toBe("true")
-    expect(apiSearchParams.get("q")).toBe("test")
-
-    // Ensure count is hidden
-    const hideCountText = screen.queryByText("700 results")
-    expect(hideCountText).toBeNull()
-  })
-
   test("Toggling facets", async () => {
     setMockApiResponses({
       search: {
@@ -1060,4 +1016,61 @@ describe("UniversalAIBanner", () => {
     expect(screen.queryByText("Universal AI")).not.toBeInTheDocument()
     expect(screen.queryByText("New on MIT Learn")).not.toBeInTheDocument()
   })
+
+  test("Vector Hybrid Search passes correct params and renders expected count/facets", async () => {
+    setMockApiResponses({
+      search: {
+        count: 700,
+        metadata: {
+          aggregations: {
+            resource_type_group: [{ key: "course", doc_count: 100 }],
+          },
+          suggestions: [],
+        },
+        results: factories.learningResources.resources({ count: 5 }).results,
+      },
+    })
+
+    // Authenticate as path editor (admin)
+    setMockResponse.get(urls.userMe.get(), {
+      is_learning_path_editor: true,
+      is_authenticated: true,
+    })
+
+    renderWithProviders(<SearchPage />, { url: "?vector_search=true&q=test" })
+
+    await waitFor(() => {
+      const call = makeRequest.mock.calls.find(([_method, url]) => {
+        return url.includes(urls.search.vectorResources())
+      })
+      expect(call).toBeDefined()
+    })
+
+    const call = makeRequest.mock.calls.find(([_method, url]) =>
+      url.includes(urls.search.vectorResources()),
+    )
+    invariant(call)
+    const fullUrl = new URL(call[1], "http://mit.edu")
+    const apiSearchParams = fullUrl.searchParams
+
+    expect(apiSearchParams.get("hybrid_search")).toBe("true")
+    expect(apiSearchParams.get("q")).toBe("test")
+
+    // Ensure count is visible
+    const countText = await screen.findByText("700 results")
+    expect(countText).toBeVisible()
+
+    // Ensure facets are visible
+    await waitFor(() => {
+      const tabs = screen.getAllByRole("tab")
+      expect(
+        tabs.map((tab) => (tab.textContent || "").replace(/\s/g, "")),
+      ).toEqual([
+        "All(100)",
+        "Courses(100)",
+        "Programs(0)",
+        "LearningMaterials(0)",
+      ])
+    })
+  })
 })

frontends/main/src/page-components/SearchDisplay/SearchDisplay.tsx

@@ -516,8 +516,8 @@ const searchModeDropdownOptions = Object.entries(
 
 /**
  * Extracts only the fields supported by the vector search API from a broader
- * search params object, dropping admin-only params (e.g., aggregations,
- * content_file_score_weight) that the vector endpoint does not accept.
+ * search params object, dropping admin-only params (e.g., content_file_score_weight)
+ * that the vector endpoint does not accept.
  *
  * The `as` casts for enum arrays are safe because the v0 and v1 generated
  * clients define separate (but structurally identical) enum types for the same
@@ -526,6 +526,7 @@ const searchModeDropdownOptions = Object.entries(
 const toVectorSearchParams = (
   params: ReturnType<typeof getSearchParams>,
 ): VectorSearchRequest => ({
+  aggregations: params.aggregations as VectorSearchRequest["aggregations"],
   certification: params.certification,
   certification_type:
     params.certification_type as VectorSearchRequest["certification_type"],
@@ -625,10 +626,13 @@ const SearchDisplay: React.FC<SearchDisplayProps> = ({
   const wantsVectorSearch = searchParams.get("vector_search") === "true"
   const isVectorSearch = wantsVectorSearch && user?.is_learning_path_editor
 
+  const queryOptions = isVectorSearch
+    ? learningResourceQueries.vectorSearch(toVectorSearchParams(allParams))
+    : learningResourceQueries.search(allParams as LRSearchRequest)
+
+  // @ts-expect-error Typescript has trouble unifying the different query key types
   const { data, isLoading, isFetching } = useQuery({
-    ...(isVectorSearch
-      ? learningResourceQueries.vectorSearch(toVectorSearchParams(allParams))
-      : learningResourceQueries.search(allParams as LRSearchRequest)),
+    ...queryOptions,
     enabled: !wantsVectorSearch || !isUserLoading,
     placeholderData: keepPreviousData,
     select: (timedData: {
@@ -985,9 +989,7 @@ const SearchDisplay: React.FC<SearchDisplayProps> = ({
                * the count when data is loaded even if count is same as previous
                * count.
                */}
-              {isFetching || isLoading || isVectorSearch
-                ? ""
-                : `${data?.count} results`}
+              {isFetching || isLoading ? "" : `${data?.count} results`}
             </VisuallyHidden>
             <UniversalAIBanner searchParams={searchParams} />
             <Stack direction="row" justifyContent="space-between">

main/settings.py

@@ -822,10 +822,10 @@ def get_all_config_keys():
 QDRANT_CLIENT_TIMEOUT = get_int(name="QDRANT_CLIENT_TIMEOUT", default=10)
 
 VECTOR_HYBRID_SEARCH_PREFETCH_MULTIPLIER = get_int(
-    name="VECTOR_HYBRID_SEARCH_PREFETCH_MULTIPLIER", default=20
+    name="VECTOR_HYBRID_SEARCH_PREFETCH_MULTIPLIER", default=5
 )
 VECTOR_HYBRID_SEARCH_PREFETCH_MAX_LIMIT = get_int(
-    name="VECTOR_HYBRID_SEARCH_PREFETCH_MAX_LIMIT", default=10000
+    name="VECTOR_HYBRID_SEARCH_PREFETCH_MAX_LIMIT", default=500
 )
 # toggle to use requests (default for local) or webdriver which renders js elements
 EMBEDDINGS_EXTERNAL_FETCH_USE_WEBDRIVER = get_bool(

openapi/specs/v0.yaml

@@ -827,6 +827,58 @@ paths:
       description: Vector Search for content
       summary: Content File Vector Search
       parameters:
+      - in: query
+        name: aggregations
+        schema:
+          type: array
+          items:
+            enum:
+            - key
+            - course_number
+            - platform
+            - offered_by
+            - file_extension
+            - content_feature_type
+            - run_readable_id
+            - resource_readable_id
+            - run_title
+            - edx_module_id
+            - content_type
+            - description
+            - title
+            - url
+            - file_type
+            - summary
+            - flashcards
+            - checksum
+            type: string
+            description: |-
+              * `key` - Key
+              * `course_number` - Course Number
+              * `platform` - Platform
+              * `offered_by` - Offered By
+              * `file_extension` - File Extension
+              * `content_feature_type` - Content Feature Type
+              * `run_readable_id` - Run Readable Id
+              * `resource_readable_id` - Resource Readable Id
+              * `run_title` - Run Title
+              * `edx_module_id` - Edx Module Id
+              * `content_type` - Content Type
+              * `description` - Description
+              * `title` - Title
+              * `url` - Url
+              * `file_type` - File Type
+              * `summary` - Summary
+              * `flashcards` - Flashcards
+              * `checksum` - Checksum
+        description: "aggregations for facet counts             \n\n* `key` - Key\n\
+          * `course_number` - Course Number\n* `platform` - Platform\n* `offered_by`\
+          \ - Offered By\n* `file_extension` - File Extension\n* `content_feature_type`\
+          \ - Content Feature Type\n* `run_readable_id` - Run Readable Id\n* `resource_readable_id`\
+          \ - Resource Readable Id\n* `run_title` - Run Title\n* `edx_module_id` -\
+          \ Edx Module Id\n* `content_type` - Content Type\n* `description` - Description\n\
+          * `title` - Title\n* `url` - Url\n* `file_type` - File Type\n* `summary`\
+          \ - Summary\n* `flashcards` - Flashcards\n* `checksum` - Checksum"
       - in: query
         name: collection_name
         schema:
@@ -961,6 +1013,61 @@ paths:
       description: Vector Search for learning resources
       summary: Vector Search
       parameters:
+      - in: query
+        name: aggregations
+        schema:
+          type: array
+          items:
+            enum:
+            - readable_id
+            - resource_type
+            - certification
+            - certification_type
+            - professional
+            - free
+            - course_feature
+            - topic
+            - ocw_topic
+            - level
+            - department
+            - platform
+            - offered_by
+            - delivery
+            - title
+            - url
+            - resource_type_group
+            - resource_category
+            - published
+            type: string
+            description: |-
+              * `readable_id` - Readable Id
+              * `resource_type` - Resource Type
+              * `certification` - Certification
+              * `certification_type` - Certification Type
+              * `professional` - Professional
+              * `free` - Free
+              * `course_feature` - Course Feature
+              * `topic` - Topic
+              * `ocw_topic` - Ocw Topic
+              * `level` - Level
+              * `department` - Department
+              * `platform` - Platform
+              * `offered_by` - Offered By
+              * `delivery` - Delivery
+              * `title` - Title
+              * `url` - Url
+              * `resource_type_group` - Resource Type Group
+              * `resource_category` - Resource Category
+              * `published` - Published
+        description: "aggregations for facet counts             \n\n* `readable_id`\
+          \ - Readable Id\n* `resource_type` - Resource Type\n* `certification` -\
+          \ Certification\n* `certification_type` - Certification Type\n* `professional`\
+          \ - Professional\n* `free` - Free\n* `course_feature` - Course Feature\n\
+          * `topic` - Topic\n* `ocw_topic` - Ocw Topic\n* `level` - Level\n* `department`\
+          \ - Department\n* `platform` - Platform\n* `offered_by` - Offered By\n*\
+          \ `delivery` - Delivery\n* `title` - Title\n* `url` - Url\n* `resource_type_group`\
+          \ - Resource Type Group\n* `resource_category` - Resource Category\n* `published`\
+          \ - Published"
       - in: query
         name: certification
         schema:
@@ -1255,6 +1362,13 @@ paths:
         schema:
           type: boolean
           nullable: true
+      - in: query
+        name: published
+        schema:
+          type: boolean
+          default: true
+        description: If the resource is published. We default to True unless passed
+          in
       - in: query
         name: q
         schema:

vector_search/constants.py

@@ -45,6 +45,8 @@
     "title": "title",
     "url": "url",
     "resource_type_group": "resource_type_group",
+    "resource_category": "resource_category",
+    "published": "published",
 }
 
 
@@ -71,6 +73,7 @@
     "url": models.PayloadSchemaType.KEYWORD,
     "title": models.PayloadSchemaType.KEYWORD,
     "resource_type_group": models.PayloadSchemaType.KEYWORD,
+    "resource_category": models.PayloadSchemaType.KEYWORD,
 }
 
 """
@@ -92,3 +95,14 @@
 QDRANT_TOPIC_INDEXES = {
     "name": models.PayloadSchemaType.KEYWORD,
 }
+
+
+CONTENT_FILES_RETRIEVE_PAYLOAD = True
+RESOURCES_RETRIEVE_PAYLOAD = ["readable_id"]
+
+
+COLLECTION_PARAM_MAP = {
+    RESOURCES_COLLECTION_NAME: QDRANT_RESOURCE_PARAM_MAP,
+    TOPICS_COLLECTION_NAME: QDRANT_TOPICS_PARAM_MAP,
+    CONTENT_FILES_COLLECTION_NAME: QDRANT_CONTENT_FILE_PARAM_MAP,
+}