yt-embeddings-strapi-plugin 0.0.1

package/dist/server/src/tools/index.d.ts

@@ -0,0 +1,126 @@
+import { searchYtKnowledgeTool } from './search-yt-knowledge';
+import { getVideoTranscriptRangeTool } from './get-video-transcript-range';
+import { listYtVideosTool } from './list-yt-videos';
+import { getYtVideoSummaryTool } from './get-yt-video-summary';
+export declare const tools: ({
+    name: string;
+    description: string;
+    schema: import("zod").ZodObject<{
+        query: import("zod").ZodString;
+        limit: import("zod").ZodDefault<import("zod").ZodOptional<import("zod").ZodNumber>>;
+        videoId: import("zod").ZodOptional<import("zod").ZodString>;
+        topics: import("zod").ZodOptional<import("zod").ZodArray<import("zod").ZodString, "many">>;
+        contextWindowSeconds: import("zod").ZodDefault<import("zod").ZodOptional<import("zod").ZodNumber>>;
+        minSimilarity: import("zod").ZodDefault<import("zod").ZodOptional<import("zod").ZodNumber>>;
+    }, "strip", import("zod").ZodTypeAny, {
+        query?: string;
+        limit?: number;
+        videoId?: string;
+        topics?: string[];
+        contextWindowSeconds?: number;
+        minSimilarity?: number;
+    }, {
+        query?: string;
+        limit?: number;
+        videoId?: string;
+        topics?: string[];
+        contextWindowSeconds?: number;
+        minSimilarity?: number;
+    }>;
+    execute: (args: unknown, strapi: import("@strapi/types/dist/core").Strapi) => Promise<{
+        results: any[];
+        message: string;
+    } | {
+        results: any;
+        message?: undefined;
+    }>;
+    publicSafe: boolean;
+} | {
+    name: string;
+    description: string;
+    schema: import("zod").ZodObject<{
+        videoId: import("zod").ZodString;
+        startSeconds: import("zod").ZodNumber;
+        endSeconds: import("zod").ZodNumber;
+    }, "strip", import("zod").ZodTypeAny, {
+        videoId?: string;
+        startSeconds?: number;
+        endSeconds?: number;
+    }, {
+        videoId?: string;
+        startSeconds?: number;
+        endSeconds?: number;
+    }>;
+    execute: (args: unknown, strapi: import("@strapi/types/dist/core").Strapi) => Promise<{
+        transcript: any;
+        message: string;
+        videoId?: undefined;
+        range?: undefined;
+        deepLink?: undefined;
+    } | {
+        videoId: string;
+        range: string;
+        transcript: any;
+        deepLink: string;
+        message?: undefined;
+    }>;
+    publicSafe: boolean;
+} | {
+    name: string;
+    description: string;
+    schema: import("zod").ZodObject<{
+        page: import("zod").ZodDefault<import("zod").ZodOptional<import("zod").ZodNumber>>;
+        pageSize: import("zod").ZodDefault<import("zod").ZodOptional<import("zod").ZodNumber>>;
+    }, "strip", import("zod").ZodTypeAny, {
+        page?: number;
+        pageSize?: number;
+    }, {
+        page?: number;
+        pageSize?: number;
+    }>;
+    execute: (args: unknown, strapi: import("@strapi/types/dist/core").Strapi) => Promise<{
+        videos: any;
+        pagination: {
+            page: any;
+            pageSize: any;
+            total: any;
+            pageCount: number;
+        };
+    }>;
+    publicSafe: boolean;
+} | {
+    name: string;
+    description: string;
+    schema: import("zod").ZodObject<{
+        videoId: import("zod").ZodString;
+    }, "strip", import("zod").ZodTypeAny, {
+        videoId?: string;
+    }, {
+        videoId?: string;
+    }>;
+    execute: (args: unknown, strapi: import("@strapi/types/dist/core").Strapi) => Promise<{
+        error: boolean;
+        message: string;
+        videoId?: undefined;
+        title?: undefined;
+        topics?: undefined;
+        summary?: undefined;
+        chunkCount?: undefined;
+        embeddingStatus?: undefined;
+        watchLink?: undefined;
+        keyMoments?: undefined;
+    } | {
+        videoId: any;
+        title: any;
+        topics: any;
+        summary: any;
+        chunkCount: any;
+        embeddingStatus: any;
+        watchLink: string;
+        keyMoments: any;
+        error?: undefined;
+        message?: undefined;
+    }>;
+    publicSafe: boolean;
+})[];
+export { searchYtKnowledgeTool, getVideoTranscriptRangeTool, listYtVideosTool, getYtVideoSummaryTool, };

package/dist/server/src/tools/list-yt-videos.d.ts

@@ -0,0 +1,25 @@
+import type { Core } from '@strapi/strapi';
+export declare const listYtVideosTool: {
+    name: string;
+    description: string;
+    schema: import("zod").ZodObject<{
+        page: import("zod").ZodDefault<import("zod").ZodOptional<import("zod").ZodNumber>>;
+        pageSize: import("zod").ZodDefault<import("zod").ZodOptional<import("zod").ZodNumber>>;
+    }, "strip", import("zod").ZodTypeAny, {
+        page?: number;
+        pageSize?: number;
+    }, {
+        page?: number;
+        pageSize?: number;
+    }>;
+    execute: (args: unknown, strapi: Core.Strapi) => Promise<{
+        videos: any;
+        pagination: {
+            page: any;
+            pageSize: any;
+            total: any;
+            pageCount: number;
+        };
+    }>;
+    publicSafe: boolean;
+};

package/dist/server/src/tools/search-yt-knowledge.d.ts

@@ -0,0 +1,35 @@
+import type { Core } from '@strapi/strapi';
+export declare const searchYtKnowledgeTool: {
+    name: string;
+    description: string;
+    schema: import("zod").ZodObject<{
+        query: import("zod").ZodString;
+        limit: import("zod").ZodDefault<import("zod").ZodOptional<import("zod").ZodNumber>>;
+        videoId: import("zod").ZodOptional<import("zod").ZodString>;
+        topics: import("zod").ZodOptional<import("zod").ZodArray<import("zod").ZodString, "many">>;
+        contextWindowSeconds: import("zod").ZodDefault<import("zod").ZodOptional<import("zod").ZodNumber>>;
+        minSimilarity: import("zod").ZodDefault<import("zod").ZodOptional<import("zod").ZodNumber>>;
+    }, "strip", import("zod").ZodTypeAny, {
+        query?: string;
+        limit?: number;
+        videoId?: string;
+        topics?: string[];
+        contextWindowSeconds?: number;
+        minSimilarity?: number;
+    }, {
+        query?: string;
+        limit?: number;
+        videoId?: string;
+        topics?: string[];
+        contextWindowSeconds?: number;
+        minSimilarity?: number;
+    }>;
+    execute: (args: unknown, strapi: Core.Strapi) => Promise<{
+        results: any[];
+        message: string;
+    } | {
+        results: any;
+        message?: undefined;
+    }>;
+    publicSafe: boolean;
+};

package/dist/server/src/utils/chunking.d.ts

@@ -0,0 +1,44 @@
+/**
+ * Text chunking utilities for splitting large content into embeddable chunks
+ */
+export interface ChunkOptions {
+    /** Maximum characters per chunk (default: 4000, roughly ~1000 tokens) */
+    chunkSize?: number;
+    /** Number of characters to overlap between chunks (default: 200) */
+    chunkOverlap?: number;
+    /** Separator to use when splitting (default: splits on paragraphs, sentences, then words) */
+    separators?: string[];
+}
+export interface TextChunk {
+    /** The chunk text content */
+    text: string;
+    /** Zero-based chunk index */
+    chunkIndex: number;
+    /** Total number of chunks */
+    totalChunks: number;
+    /** Character offset in original text */
+    startOffset: number;
+    /** Character end offset in original text */
+    endOffset: number;
+}
+/**
+ * Estimate token count from character count
+ * OpenAI models average ~4 characters per token for English text
+ */
+export declare function estimateTokens(text: string): number;
+/**
+ * Check if content exceeds the recommended chunk size
+ */
+export declare function needsChunking(content: string, maxChars?: number): boolean;
+/**
+ * Split content into chunks suitable for embedding
+ *
+ * @param content - The text content to split
+ * @param options - Chunking options
+ * @returns Array of TextChunk objects
+ */
+export declare function chunkContent(content: string, options?: ChunkOptions): TextChunk[];
+/**
+ * Format chunk title with index information
+ */
+export declare function formatChunkTitle(baseTitle: string, chunkIndex: number, totalChunks: number): string;

package/dist/server/src/utils/preprocessing.d.ts

@@ -0,0 +1,26 @@
+/**
+ * Content preprocessing utilities for cleaning text before embedding
+ * Handles HTML stripping, markdown conversion, and text normalization
+ */
+export interface PreprocessOptions {
+    /** Strip HTML tags from content (default: true) */
+    stripHtml?: boolean;
+    /** Strip markdown syntax from content (default: true) */
+    stripMarkdown?: boolean;
+    /** Normalize whitespace (collapse multiple spaces/newlines) (default: true) */
+    normalizeWhitespace?: boolean;
+}
+/**
+ * Preprocess content for embedding
+ * Strips HTML, markdown, and normalizes whitespace
+ *
+ * @param content - The raw content to preprocess
+ * @param options - Preprocessing options
+ * @returns Cleaned plain text ready for embedding
+ */
+export declare function preprocessContent(content: string, options?: PreprocessOptions): string;
+/**
+ * Check if content needs preprocessing
+ * Returns true if content contains HTML or Markdown
+ */
+export declare function needsPreprocessing(content: string): boolean;

package/dist/server/src/utils/yt-chunker.d.ts

@@ -0,0 +1,16 @@
+export interface Segment {
+    text: string;
+    start: number;
+    end: number;
+    duration: number;
+}
+export interface YtChunk {
+    text: string;
+    startSeconds: number;
+    endSeconds: number;
+    durationSeconds: number;
+    chunkIndex: number;
+    segments: Segment[];
+    tokens: number;
+}
+export declare function chunkTranscript(segments: Segment[]): YtChunk[];

package/package.json

@@ -0,0 +1,106 @@
+{
+  "name": "yt-embeddings-strapi-plugin",
+  "version": "0.0.1",
+  "description": "Strapi v5 plugin for vector embeddings with OpenAI and Neon PostgreSQL. Enables semantic search, RAG chat, and MCP (Model Context Protocol) integration.",
+  "keywords": [
+    "strapi",
+    "strapi-plugin",
+    "embeddings",
+    "vector-search",
+    "rag",
+    "openai",
+    "neon",
+    "pgvector",
+    "mcp",
+    "semantic-search"
+  ],
+  "license": "MIT",
+  "author": "Paul Bratslavsky <codingafterthirty@gmail.com>",
+  "homepage": "https://github.com/PaulBrats662/strapi-content-embeddings#readme",
+  "repository": {
+    "type": "git",
+    "url": "git+https://github.com/PaulBratslavsky/strapi-content-embeddings.git"
+  },
+  "bugs": {
+    "url": "https://github.com/PaulBratslavsky/strapi-content-embeddings/issues"
+  },
+  "type": "commonjs",
+  "overrides": {
+    "@langchain/core": "^1.1.31"
+  },
+  "exports": {
+    "./package.json": "./package.json",
+    "./strapi-admin": {
+      "types": "./dist/admin/src/index.d.tsx",
+      "source": "./admin/src/index.tsx",
+      "import": "./dist/admin/index.mjs",
+      "require": "./dist/admin/index.js",
+      "default": "./dist/admin/index.js"
+    },
+    "./strapi-server": {
+      "types": "./dist/server/src/index.d.ts",
+      "source": "./server/src/index.ts",
+      "import": "./dist/server/index.mjs",
+      "require": "./dist/server/index.js",
+      "default": "./dist/server/index.js"
+    }
+  },
+  "files": [
+    "dist"
+  ],
+  "scripts": {
+    "build": "strapi-plugin build",
+    "setup-db": "node scripts/setup-db.js",
+    "test:ts:back": "run -T tsc -p server/tsconfig.json",
+    "test:ts:front": "run -T tsc -p admin/tsconfig.json",
+    "verify": "strapi-plugin verify",
+    "watch": "strapi-plugin watch",
+    "watch:link": "strapi-plugin watch:link"
+  },
+  "dependencies": {
+    "@langchain/community": "^1.1.2",
+    "@langchain/core": "^1.1.31",
+    "@langchain/openai": "^1.2.1",
+    "@mdxeditor/editor": "^3.52.3",
+    "@modelcontextprotocol/sdk": "^1.12.0",
+    "@strapi/design-system": "^2.0.0-rc.12",
+    "@strapi/icons": "^2.0.0-rc.12",
+    "html-to-text": "^9.0.5",
+    "langchain": "^1.2.4",
+    "pg": "^8.13.1",
+    "remove-markdown": "^0.5.5",
+    "qs": "^6.13.1",
+    "react-intl": "^6.8.4",
+    "react-markdown": "^10.1.0",
+    "zod": "^3.24.0"
+  },
+  "devDependencies": {
+    "@strapi/sdk-plugin": "^5.2.7",
+    "@strapi/strapi": "^5.2.0",
+    "@strapi/typescript-utils": "^5.2.0",
+    "@types/pg": "^8.11.10",
+    "@types/qs": "^6.9.17",
+    "@types/react": "^18.3.12",
+    "@types/react-dom": "^18.3.1",
+    "prettier": "^3.3.3",
+    "react": "^18.3.1",
+    "react-dom": "^18.3.1",
+    "react-router-dom": "^6.27.0",
+    "styled-components": "^6.1.13",
+    "typescript": "^5.6.3"
+  },
+  "peerDependencies": {
+    "@strapi/sdk-plugin": "^5.2.7",
+    "@strapi/strapi": "^5.2.0",
+    "react": "^18.3.1",
+    "react-dom": "^18.3.1",
+    "react-router-dom": "^6.27.0",
+    "styled-components": "^6.1.13"
+  },
+  "strapi": {
+    "kind": "plugin",
+    "name": "yt-embeddings-strapi-plugin",
+    "displayName": "YT Embeddings",
+    "description": "YouTube transcript vector embeddings with semantic search, RAG chat, and MCP integration."
+  }
+}