devflowinc · cdxker · Dec 19, 2024 · Dec 11, 2024 · Dec 19, 2024 · Dec 19, 2024
diff --git a/.env.server b/.env.server
@@ -50,3 +50,4 @@ FIRECRAWL_API_KEY=fc-abdef**************
 PDF2MD_URL="http://localhost:8081"
 BATCH_CHUNK_LIMIT=120
 CHAT_COMPLETION_TIMEOUT_SECS=10
+YOUTUBE_API_KEY=""
diff --git a/.github/workflows/push-server.yml b/.github/workflows/push-server.yml
@@ -802,3 +802,53 @@ jobs:
           push: true
           tags: ${{ steps.meta.outputs.tags }}
           labels: ${{ steps.meta.outputs.labels }}
+
+
+  pagefind-worker:
+    name: Push Pagefind Index worker
+    runs-on: ${{ matrix.runner }}
+    strategy:
+      matrix:
+        runner: [blacksmith-8vcpu-ubuntu-2204]
+        platform: [linux/amd64]
+        exclude:
+          - runner: blacksmith-8vcpu-ubuntu-2204
+            platform: linux/arm64
+          - runner: blacksmith-8vcpu-ubuntu-2204-arm
+            platform: linux/amd64
+    steps:
+      - name: Checkout the repo
+        uses: actions/checkout@v4
+
+      # - name: Set up QEMU
+      #   uses: docker/setup-qemu-action@v3
+
+      - name: Setup buildx
+        uses: docker/setup-buildx-action@v3
+
+      - name: Login to Docker Hub
+        uses: docker/login-action@v3
+        with:
+          username: ${{ secrets.DOCKER_USERNAME }}
+          password: ${{ secrets.DOCKER_PASSWORD }}
+
+      - name: Docker meta
+        id: meta
+        uses: docker/metadata-action@v5
+        with:
+          # list of Docker images to use as base name for tags
+          images: |
+            trieve/pagefind-worker
+          tags: |
+            type=raw,latest
+            type=sha
+
+      - name: Build and push Docker image
+        uses: useblacksmith/[email protected]
+        with:
+          platforms: ${{ matrix.platform }}
+          context: server/
+          file: ./server/Dockerfile.pagefind-worker
+          push: true
+          tags: ${{ steps.meta.outputs.tags }}
+          labels: ${{ steps.meta.outputs.labels }}
diff --git a/clients/search-component/example/.env.dist b/clients/search-component/example/.env.dist
@@ -6,4 +6,8 @@ VITE_BRAND_NAME="Trieve"
 VITE_BRAND_COLOR="#CB53EB"
 VITE_PROBLEM_LINK="mailto:[email protected]?subject="
 VITE_SHOW_FLOATING_BTN="true"
-VITE_FLOATING_BTN_POSITION="bottom-right"
+VITE_FLOATING_BTN_POSITION="bottom-right"
+VITE_USE_PAGEFIND="false"
+VITE_PAGEFIND_URL="https://pagefind-testing-index.trieve.ai/pagfind-index-west/pagefind"
+VITE_USE_GROUP_SEARCH="false"
+VITE_DEFAULT_TAGS='[{"tag": "pink", "label": "Pink"}]'
diff --git a/clients/search-component/example/index.html b/clients/search-component/example/index.html
@@ -8,13 +8,13 @@
     <link rel="icon" type="image/png" sizes="32x32" href="https://cdn.trieve.ai/favicon-32x32.png" />
     <link rel="icon" type="image/png" sizes="16x16" href="https://cdn.trieve.ai/favicon-16x16.png" />
     <meta name="viewport" content="width=device-width, initial-scale=1.0" />
-    <meta property="og:title" content="Trieve Drop-In RAG and Search Component">
+    <meta property="og:title" content="Search Component By Trieve">
     <meta property="og:site_name" content="advanced relevance semantic search and RAG API">
     <meta property="og:url" content="https://docsearch.trieve.ai">
     <meta property="og:description" content="Build better, faster, and more relevant search and RAG with our open source API. Date recency biasing, re-ranker models, dense vector search, sub-sentence highlighting, and more all on one endpoint that you can host yourself." >
     <meta property="og:type" content="">
     <meta property="og:image" content="https://cdn.trieve.ai/trieve-og.png">
-    <title>Trieve Drop-In RAG and Search Component</title>
+    <title>Search Component By Trieve</title>
   </head>
   <body>
     <div id="root"></div>

diff --git a/clients/search-component/example/src/routes/ecommerce.tsx b/clients/search-component/example/src/routes/ecommerce.tsx
@@ -19,6 +19,8 @@ export default function ECommerce() {
   const useGroupSearch = import.meta.env.VITE_USE_GROUP_SEARCH == "true";
   const showFloatingButton = import.meta.env.VITE_SHOW_FLOATING_BTN == "true";
   const floatingButtonPosition = import.meta.env.VITE_FLOATING_BTN_POSITION;
+  const usePagefind = import.meta.env.VITE_USE_PAGEFIND == "true";
+  const pagefindUrl = import.meta.env.VITE_PAGEFIND_URL;
 
   const defaultSearchQueries: string[] = (
     import.meta.env.VITE_DEFAULT_SEARCH_QUERIES ?? ""
@@ -33,9 +35,8 @@ export default function ECommerce() {
   return (
     <>
       <div
-        className={`p-12 flex flex-col items-center justify-center w-screen h-screen relative ${
-          theme === "dark" ? "bg-zinc-900 text-zinc-50" : ""
-        }`}
+        className={`p-12 flex flex-col items-center justify-center w-screen h-screen relative ${theme === "dark" ? "bg-zinc-900 text-zinc-50" : ""
+          }`}
       >
         <div className="absolute top-6 right-6">
           <ul>
@@ -81,10 +82,21 @@ export default function ECommerce() {
                 use_autocomplete: false,
                 search_type: "fulltext",
               }}
+              buttonTriggers={[
+                {
+                  selector: ".random-trigger-location",
+                  mode: "chat",
+                },
+              ]}
+              pagefindOptions={{
+                usePagefind: usePagefind,
+                cdnBaseUrl: pagefindUrl,
+              }}
               defaultSearchQueries={defaultSearchQueries}
               tags={defaultTags}
               floatingButtonPosition={floatingButtonPosition}
               showFloatingButton={showFloatingButton}
+              debounceMs={500}
             />
           </>
         ) : (

diff --git a/clients/search-component/example/src/routes/index.lazy.tsx b/clients/search-component/example/src/routes/index.lazy.tsx
@@ -17,6 +17,8 @@ export default function Home() {
   const brandFontFamily = import.meta.env.VITE_BRAND_FONT_FAMILY;
   const problemLink = import.meta.env.VITE_PROBLEM_LINK;
   const useGroupSearch = import.meta.env.VITE_USE_GROUP_SEARCH == "true";
+  const usePagefind = import.meta.env.VITE_USE_PAGEFIND == "true";
+  const pagefindUrl = import.meta.env.VITE_PAGEFIND_URL;
   const defaultSearchQueries: string[] = (
     import.meta.env.VITE_DEFAULT_SEARCH_QUERIES ?? ""
   ).split(",");
@@ -127,6 +129,10 @@ export default function Home() {
                 },
               ]}
               useGroupSearch={useGroupSearch}
+              pagefindOptions={{
+                usePagefind: usePagefind,
+                cdnBaseUrl: pagefindUrl,
+              }}
               defaultAiQuestions={[
                 "What is Trieve?",
                 "How to perform autocomplete search?",

diff --git a/clients/search-component/src/utils/hooks/modal-context.tsx b/clients/search-component/src/utils/hooks/modal-context.tsx
@@ -14,7 +14,10 @@ import {
 } from "trieve-ts-sdk";
 import {
   countChunks,
+  countChunksWithPagefind,
+  groupSearchWithPagefind,
   groupSearchWithTrieve,
+  searchWithPagefind,
   searchWithTrieve,
 } from "../trieve";
 
@@ -33,11 +36,19 @@ type customAutoCompleteAddOn = {
   use_autocomplete?: boolean;
 };
 
+// eslint-disable-next-line @typescript-eslint/no-explicit-any
+export type PagefindApi = any;
+
 export type currencyPosition = "before" | "after";
 export type ModalTypes = "ecommerce" | "docs";
 export type SearchModes = "chat" | "search";
 export type searchOptions = simpleSearchReqPayload & customAutoCompleteAddOn;
 
+export interface PagefindOptions {
+  usePagefind: boolean;
+  cdnBaseUrl?: string;
+}
+
 export type ModalProps = {
   datasetId: string;
   apiKey: string;
@@ -66,6 +77,7 @@ export type ModalProps = {
     icon?: () => JSX.Element;
   }[];
   defaultSearchMode?: SearchModes;
+  pagefindOptions?: PagefindOptions;
   type?: ModalTypes;
   useGroupSearch?: boolean;
   allowSwitchingModes?: boolean;
@@ -147,6 +159,7 @@ const ModalContext = createContext<{
   currentGroup: ChunkGroup | null;
   setCurrentGroup: React.Dispatch<React.SetStateAction<ChunkGroup | null>>;
   tagCounts: CountChunkQueryResponseBody[];
+  pagefind?: PagefindApi;
 }>({
   props: defaultProps,
   trieveSDK: (() => {}) as unknown as TrieveSDK,
@@ -170,6 +183,7 @@ const ModalContext = createContext<{
   setCurrentGroup: () => {},
   tagCounts: [],
   setContextProps: () => {},
+  pagefind: null,
 });
 
 const ModalProvider = ({
@@ -197,6 +211,7 @@ const ModalProvider = ({
   const [currentTag, setCurrentTag] = useState(
     props.tags?.find((t) => t.selected)?.tag || "all"
   );
+  const [pagefind, setPagefind] = useState<PagefindApi | null>(null);
 
   const [currentGroup, setCurrentGroup] = useState<ChunkGroup | null>(null);
 
@@ -214,7 +229,7 @@ const ModalProvider = ({
 
     try {
       setLoadingResults(true);
-      if (props.useGroupSearch) {
+      if (props.useGroupSearch && !props.pagefindOptions?.usePagefind) {
         const results = await groupSearchWithTrieve({
           query: query,
           searchOptions: props.searchOptions,
@@ -236,6 +251,33 @@ const ModalProvider = ({
 
         setResults(Array.from(groupMap.values()));
         setRequestID(results.requestID);
+      } else if (props.useGroupSearch && props.pagefindOptions?.usePagefind) {
+
+        const results = await groupSearchWithPagefind(
+          pagefind,
+          query,
+          props.datasetId,
+          currentTag !== "all" ? currentTag : undefined
+        );
+        const groupMap = new Map<string, GroupChunk[]>();
+        results.groups.forEach((group) => {
+          const title = group.chunks[0].chunk.metadata?.title;
+          if (groupMap.has(title)) {
+            groupMap.get(title)?.push(group);
+          } else {
+            groupMap.set(title, [group]);
+          }
+        });
+        setResults(Array.from(groupMap.values()));
+
+      } else if (!props.useGroupSearch && props.pagefindOptions?.usePagefind) {
+        const results = await searchWithPagefind(
+          pagefind,
+          query,
+          props.datasetId,
+          currentTag !== "all" ? currentTag : undefined
+        );
+        setResults(results);
       } else {
         const results = await searchWithTrieve({
           query: query,
@@ -266,24 +308,33 @@ const ModalProvider = ({
       return;
     }
     if (props.tags?.length) {
-      try {
-        const numberOfRecords = await Promise.all(
-          [ALL_TAG, ...props.tags].map((tag) =>
-            countChunks({
-              query: query,
-              trieve: trieve,
-              abortController,
-              ...(tag.tag !== "all" && { tag: tag.tag }),
-            })
-          )
+      if (props.pagefindOptions?.usePagefind) {
+        const filterCounts = await countChunksWithPagefind(
+          pagefind,
+          query,
+          props.tags
         );
-        setTagCounts(numberOfRecords);
-      } catch (e) {
-        if (
-          e != "AbortError" &&
-          e != "AbortError: signal is aborted without reason"
-        ) {
-          console.error(e);
+        setTagCounts(filterCounts);
+      } else {
+        try {
+          const numberOfRecords = await Promise.all(
+            [ALL_TAG, ...props.tags].map((tag) =>
+              countChunks({
+                query: query,
+                trieve: trieve,
+                abortController,
+                ...(tag.tag !== "all" && { tag: tag.tag }),
+              })
+            )
+          );
+          setTagCounts(numberOfRecords);
+        } catch (e) {
+          if (
+            e != "AbortError" &&
+            e != "AbortError: signal is aborted without reason"
+          ) {
+            console.error(e);
+          }
         }
       }
     }
@@ -296,6 +347,17 @@ const ModalProvider = ({
     }));
   }, [onLoadProps]);
 
+  useEffect(() => {
+    if (props.pagefindOptions?.usePagefind) {
+      const pagefind_base_url = `${props?.pagefindOptions.cdnBaseUrl}/${props.datasetId}`;
+      import(`${pagefind_base_url}/pagefind.js`).then((pagefind) => {
+        setPagefind(pagefind)
+        pagefind.filters().then(() => {
+        })
+      });
+    }
+  }, []);
+
   useEffect(() => {
     props.onOpenChange?.(open);
   }, [open]);

diff --git a/clients/search-component/src/utils/hooks/setClickTriggers.ts b/clients/search-component/src/utils/hooks/setClickTriggers.ts
@@ -24,7 +24,6 @@ export const setClickTriggers = (
     let element: Element | null = document.querySelector(trigger.selector);
     if (trigger.removeListeners ?? true) {
       element = removeAllClickListeners(trigger.selector);
-      console.log("Removed click listeners from", trigger.selector);
     }
 
     if (element) {