@freehour/supabase-langchain 1.0.0 → 1.0.1

package/dist/embedding-service.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"embedding-service.d.ts","sourceRoot":"","sources":["../lib/embedding-service.ts"],"names":[],"mappings":"AAKA,OAAO,KAAK,EAAE,mBAAmB,EAAc,oBAAoB,EAAE,OAAO,EAAgB,eAAe,EAAE,UAAU,EAAE,QAAQ,EAAE,UAAU,EAAE,eAAe,EAA+B,MAAM,yBAAyB,CAAC;AAU7N,OAAO,KAAK,EAAE,qBAAqB,EAAE,MAAM,4CAA4C,CAAC;AACxF,OAAO,EAAE,mBAAmB,EAAE,MAAM,4CAA4C,CAAC;AAEjF,OAAO,KAAK,EAAE,mBAAmB,EAAE,MAAM,4BAA4B,CAAC;AACtE,OAAO,KAAK,EAAE,sBAAsB,EAAE,MAAM,4BAA4B,CAAC;AACzE,OAAO,KAAK,EAAE,yBAAyB,EAAE,MAAM,8BAA8B,CAAC;AAG9E,OAAO,KAAK,EAAE,cAAc,EAAE,MAAM,uBAAuB,CAAC;AAE5D,OAAO,KAAK,EAAE,QAAQ,IAAI,iBAAiB,EAAE,MAAM,sBAAsB,CAAC;AAC1E,OAAO,KAAK,EAAE,YAAY,EAAE,mBAAmB,EAAE,eAAe,EAAE,MAAM,YAAY,CAAC;AAGrF,MAAM,MAAM,eAAe,CACvB,UAAU,SAAS,MAAM,EACzB,QAAQ,SAAS,UAAU,GAAG,UAAU,IACxC,QAAQ,CAAC,yBAAyB,CAAC,mBAAmB,CAAC,EAAE,aAAa,GAAG,QAAQ,CAAC,GAAG;IACrF,MAAM,CAAC,EAAE,qBAAqB,GAAG,OAAO,CAAC,eAAe,CAAC,UAAU,CAAC,GAAG,YAAY,GAAG,QAAQ,CAAC,CAAC;CACnG,CAAC;AAEF,MAAM,WAAW,sBAAsB,CACnC,QAAQ,SAAS,eAAe,CAAC,UAAU,CAAC,QAAQ,CAAC,CAAC,GAAG,iBAAiB,GAAG,iBAAiB,EAC9F,aAAa,SAAS,QAAQ,CAAC,mBAAmB,CAAC,GAAG,oBAAoB,CAAC,QAAQ,CAAC;IAEpF,UAAU,EAAE,mBAAmB,CAAC;IAChC,QAAQ,EAAE,cAAc,CAAC,QAAQ,EAAE,aAAa,CAAC,CAAC;IAClD,SAAS,CAAC,EAAE,MAAM,CAAC;IACnB,YAAY,CAAC,EAAE,MAAM,CAAC;CACzB;AAED,qBAAa,gBAAgB,CACzB,QAAQ,SAAS,eAAe,CAAC,UAAU,CAAC,QAAQ,CAAC,CAAC,GAAG,iBAAiB,GAAG,iBAAiB,EAC9F,aAAa,SAAS,QAAQ,CAAC,mBAAmB,CAAC,GAAG,oBAAoB,CAAC,QAAQ,CAAC,EACpF,UAAU,SAAS,MAAM,GAAG,MAAM,EAClC,QAAQ,SAAS,UAAU,GAAG,UAAU;IAGxC,OAAO,CAAC,QAAQ,CAAC,QAAQ,CAAqC;IAC9D,OAAO,CAAC,QAAQ,CAAC,OAAO,CAA6B;IACrD,OAAO,CAAC,QAAQ,CAAC,WAAW,CAAsB;IAClD,OAAO,CAAC,QAAQ,CAAC,eAAe,CAA8B;gBAElD,EACR,QAAQ,EACR,UAAU,EACV,SAAgB,EAChB,YAAkB,GACrB,EAAE,sBAAsB,CAAC,QAAQ,EAAE,aAAa,CAAC;IAWlD,OAAO,KAAK,UAAU,GAErB;IAED,OAAO,CAAC,SAAS;IA2CjB,OAAO,CAAC,WAAW;YA6BL,gBAAgB;YA8BhB,gBAAgB;YAYhB,gBAAgB;IAKxB,MAAM,CAAC,OAAO,EAAE,OAAO,CAAC,UAAU,CAAC,EAAE,QAAQ,EAAE,QAAQ,GAAG,mBAAmB,CAAC,UAAU,EAAE,QAAQ,CAAC,GAAG,OAAO,CAAC,eAAe,CAAC,UAAU,EAAE,QAAQ,CAAC,EAAE,CAAC;IAQtJ,MAAM,CAAC,MAAM,EAAE,UAAU,EAAE,QAAQ,EAAE,QAAQ,GAAG,mBAAmB,CAAC,UAAU,EAAE,QAAQ,CAAC,GAAG,OAAO,CACrG,oBAAoB,CAChB,eAAe,CAAC,UAAU,EAAE,QAAQ,CAAC,EAAE,CAC1C,EAAE,CACN;IAkBD,SAAS,CAAC,OAAO,EAAE,eAAe,CAAC,UAAU,EAAE,QAAQ,CAAC,GAAG,sBAAsB;CAIpF"}
+{"version":3,"file":"embedding-service.d.ts","sourceRoot":"","sources":["../lib/embedding-service.ts"],"names":[],"mappings":"AAKA,OAAO,KAAK,EAAE,mBAAmB,EAAc,oBAAoB,EAAE,OAAO,EAAgB,eAAe,EAAE,UAAU,EAAE,QAAQ,EAAE,UAAU,EAAE,eAAe,EAA+B,MAAM,yBAAyB,CAAC;AAU7N,OAAO,KAAK,EAAE,qBAAqB,EAAE,MAAM,4CAA4C,CAAC;AACxF,OAAO,EAAE,mBAAmB,EAAE,MAAM,4CAA4C,CAAC;AAEjF,OAAO,KAAK,EAAE,mBAAmB,EAAE,MAAM,4BAA4B,CAAC;AACtE,OAAO,KAAK,EAAE,sBAAsB,EAAE,MAAM,4BAA4B,CAAC;AACzE,OAAO,KAAK,EAAE,yBAAyB,EAAE,MAAM,8BAA8B,CAAC;AAG9E,OAAO,KAAK,EAAE,cAAc,EAAE,MAAM,uBAAuB,CAAC;AAE5D,OAAO,KAAK,EAAE,QAAQ,IAAI,iBAAiB,EAAE,MAAM,sBAAsB,CAAC;AAC1E,OAAO,KAAK,EAAE,YAAY,EAAE,mBAAmB,EAAE,eAAe,EAAE,MAAM,YAAY,CAAC;AAIrF,MAAM,MAAM,eAAe,CACvB,UAAU,SAAS,MAAM,EACzB,QAAQ,SAAS,UAAU,GAAG,UAAU,IACxC,QAAQ,CAAC,yBAAyB,CAAC,mBAAmB,CAAC,EAAE,aAAa,GAAG,QAAQ,CAAC,GAAG;IACrF,MAAM,CAAC,EAAE,qBAAqB,GAAG,OAAO,CAAC,eAAe,CAAC,UAAU,CAAC,GAAG,YAAY,GAAG,QAAQ,CAAC,CAAC;CACnG,CAAC;AAEF,MAAM,WAAW,sBAAsB,CACnC,QAAQ,SAAS,eAAe,CAAC,UAAU,CAAC,QAAQ,CAAC,CAAC,GAAG,iBAAiB,GAAG,iBAAiB,EAC9F,aAAa,SAAS,QAAQ,CAAC,mBAAmB,CAAC,GAAG,oBAAoB,CAAC,QAAQ,CAAC;IAEpF,UAAU,EAAE,mBAAmB,CAAC;IAChC,QAAQ,EAAE,cAAc,CAAC,QAAQ,EAAE,aAAa,CAAC,CAAC;IAClD,SAAS,CAAC,EAAE,MAAM,CAAC;IACnB,YAAY,CAAC,EAAE,MAAM,CAAC;CACzB;AAED,qBAAa,gBAAgB,CACzB,QAAQ,SAAS,eAAe,CAAC,UAAU,CAAC,QAAQ,CAAC,CAAC,GAAG,iBAAiB,GAAG,iBAAiB,EAC9F,aAAa,SAAS,QAAQ,CAAC,mBAAmB,CAAC,GAAG,oBAAoB,CAAC,QAAQ,CAAC,EACpF,UAAU,SAAS,MAAM,GAAG,MAAM,EAClC,QAAQ,SAAS,UAAU,GAAG,UAAU;IAGxC,OAAO,CAAC,QAAQ,CAAC,QAAQ,CAAqC;IAC9D,OAAO,CAAC,QAAQ,CAAC,OAAO,CAA6B;IACrD,OAAO,CAAC,QAAQ,CAAC,WAAW,CAAsB;IAClD,OAAO,CAAC,QAAQ,CAAC,eAAe,CAA8B;gBAElD,EACR,QAAQ,EACR,UAAU,EACV,SAAgB,EAChB,YAAkB,GACrB,EAAE,sBAAsB,CAAC,QAAQ,EAAE,aAAa,CAAC;IAWlD,OAAO,KAAK,UAAU,GAErB;IAED,OAAO,CAAC,SAAS;IA2CjB,OAAO,CAAC,WAAW;YA6BL,gBAAgB;YA8BhB,gBAAgB;YAYhB,gBAAgB;IAKxB,MAAM,CAAC,OAAO,EAAE,OAAO,CAAC,UAAU,CAAC,EAAE,QAAQ,EAAE,QAAQ,GAAG,mBAAmB,CAAC,UAAU,EAAE,QAAQ,CAAC,GAAG,OAAO,CAAC,eAAe,CAAC,UAAU,EAAE,QAAQ,CAAC,EAAE,CAAC;IAQtJ,MAAM,CAAC,MAAM,EAAE,UAAU,EAAE,QAAQ,EAAE,QAAQ,GAAG,mBAAmB,CAAC,UAAU,EAAE,QAAQ,CAAC,GAAG,OAAO,CACrG,oBAAoB,CAChB,eAAe,CAAC,UAAU,EAAE,QAAQ,CAAC,EAAE,CAC1C,EAAE,CACN;IAkBD,SAAS,CAAC,OAAO,EAAE,eAAe,CAAC,UAAU,EAAE,QAAQ,CAAC,GAAG,sBAAsB;CAGpF"}

package/dist/errors.d.ts

@@ -0,0 +1,26 @@
+import { TracedError } from '@freehour/supabase-core';
+export interface FileNotSupportedErrorOptions extends ErrorOptions {
+    /**
+     * The name of the file that caused the error.
+     */
+    fileName?: string;
+    /**
+     * The type of the file that is not supported.
+     */
+    fileType?: string;
+}
+/**
+ * An error that indicates that a file is not supported, e.g. due to its MIME type or extension.
+ */
+export declare class FileNotSupportedError extends TracedError {
+    /**
+     * The name of the file that caused the error.
+     */
+    readonly fileName?: string;
+    /**
+     * The type of the file that is not supported.
+     */
+    readonly fileType?: string;
+    constructor(message: string, { fileName, fileType, ...options }?: FileNotSupportedErrorOptions);
+}
+//# sourceMappingURL=errors.d.ts.map

package/dist/errors.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"errors.d.ts","sourceRoot":"","sources":["../lib/errors.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,WAAW,EAAE,MAAM,yBAAyB,CAAC;AAGtD,MAAM,WAAW,4BAA6B,SAAQ,YAAY;IAE9D;;OAEG;IACH,QAAQ,CAAC,EAAE,MAAM,CAAC;IAElB;;OAEG;IACH,QAAQ,CAAC,EAAE,MAAM,CAAC;CACrB;AAED;;GAEG;AACH,qBAAa,qBAAsB,SAAQ,WAAW;IAElD;;OAEG;IACH,QAAQ,CAAC,QAAQ,CAAC,EAAE,MAAM,CAAC;IAE3B;;OAEG;IACH,QAAQ,CAAC,QAAQ,CAAC,EAAE,MAAM,CAAC;gBAGvB,OAAO,EAAE,MAAM,EACf,EACI,QAAQ,EACR,QAAQ,EACR,GAAG,OAAO,EACb,GAAE,4BAAiC;CAM3C"}

package/dist/index.d.ts

@@ -1,3 +1,4 @@
 export type * from './document';
 export * from './embedding-service';
+export * from './errors';
 //# sourceMappingURL=index.d.ts.map

package/dist/index.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../lib/index.ts"],"names":[],"mappings":"AAAA,mBAAmB,YAAY,CAAC;AAChC,cAAc,qBAAqB,CAAC"}
+{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../lib/index.ts"],"names":[],"mappings":"AAAA,mBAAmB,YAAY,CAAC;AAChC,cAAc,qBAAqB,CAAC;AACpC,cAAc,UAAU,CAAC"}

package/dist/index.js

@@ -1,6 +1,6 @@
 import { Document as e } from "langchain";
 import { Mime as t } from "@freehour/mime";
-import { DatabaseService as n, FileNotSupportedError as r, StorageService as i } from "@freehour/supabase-core";
+import { DatabaseService as n, StorageService as r, TracedError as i } from "@freehour/supabase-core";
 import { JSONLinesLoader as a, JSONLoader as o } from "@langchain/classic/document_loaders/fs/json";
 import { TextLoader as s } from "@langchain/classic/document_loaders/fs/text";
 import { CSVLoader as c } from "@langchain/community/document_loaders/fs/csv";
@@ -11,14 +11,20 @@ import { PPTXLoader as f } from "@langchain/community/document_loaders/fs/pptx";
 import { SRTLoader as p } from "@langchain/community/document_loaders/fs/srt";
 import { SupabaseVectorStore as m } from "@langchain/community/vectorstores/supabase";
 import { RecursiveCharacterTextSplitter as h } from "@langchain/textsplitters";
-//#region lib/embedding-service.ts
-var g = class {
+//#region lib/errors.ts
+var g = class extends i {
+	fileName;
+	fileType;
+	constructor(e, { fileName: t, fileType: n, ...r } = {}) {
+		super(e, r), this.fileName = t, this.fileType = n;
+	}
+}, _ = class {
 	database;
 	storage;
 	vectorStore;
 	splitterOptions;
-	constructor({ supabase: e, embeddings: t, chunkSize: r = 1e3, chunkOverlap: a = 200 }) {
-		this.database = new n({ supabase: e }), this.storage = new i({
+	constructor({ supabase: e, embeddings: t, chunkSize: i = 1e3, chunkOverlap: a = 200 }) {
+		this.database = new n({ supabase: e }), this.storage = new r({
 			client: e.storage,
 			database: this.database
 		}), this.vectorStore = new m(t, {
@@ -26,7 +32,7 @@ var g = class {
 			tableName: "embeddings",
 			queryName: "match_documents"
 		}), this.splitterOptions = {
-			chunkSize: r,
+			chunkSize: i,
 			chunkOverlap: a
 		};
 	}
@@ -45,7 +51,7 @@ var g = class {
 		if (n.equals("application/vnd.openxmlformats-officedocument.presentationml.presentation")) return new f(e);
 		if (n.equals("application/x-subrip")) return new p(e);
 		if (n.type === "text") return new s(e);
-		throw new r(`File type ${e.type} is not supported`, {
+		throw new g(`File type ${e.type} is not supported`, {
 			fileName: e.name,
 			fileType: e.type
 		});
@@ -110,4 +116,4 @@ var g = class {
 	}
 };
 //#endregion
-export { g as EmbeddingService };
+export { _ as EmbeddingService, g as FileNotSupportedError };

package/package.json

@@ -3,7 +3,7 @@
   "private": false,
   "displayName": "Supabase-LangChain",
   "description": "Integration package for using LangChain with Supabase",
-  "version": "1.0.0",
+  "version": "1.0.1",
   "type": "module",
   "repository": {
     "type": "git",
@@ -17,7 +17,7 @@
   "files": [
     "dist",
     "supabase/migrations/0001_supabase_langchain.sql",
-    "supabase/schemas/supabase-langchain.sql",
+    "supabase/schemas/supabase-langchain",
     "scripts/copy-supabase-assets.js"
   ],
   "exports": {
@@ -28,26 +28,31 @@
   },
   "dependencies": {
     "@freehour/mime": "^1.1.7",
-    "@freehour/supabase-core": "^2.0.4",
-    "@langchain/classic": "^1.0.29",
+    "@freehour/supabase-core": "^2.1.0",
+    "@langchain/classic": "^1.0.31",
     "@langchain/community": "^1.1.27",
-    "@langchain/core": "^1.1.39",
+    "@langchain/core": "^1.1.41",
     "@langchain/textsplitters": "^1.0.1",
-    "@supabase/storage-js": "^2.103.0",
-    "@supabase/supabase-js": "^2.103.0",
-    "langchain": "^1.3.1",
-    "officeparser": "^6.0.7",
+    "langchain": "^1.3.4",
+    "officeparser": "^6.1.0",
     "pdf-parse": "^2.4.5",
     "srt-parser-2": "^1.2.3"
   },
+  "peerDependencies": {
+    "@supabase/supabase-js": "^2.104.0"
+  },
+  "overrides": {
+    "@browserbasehq/stagehand": "3.2.1"
+  },
   "devDependencies": {
     "@freehour/eslint-rules": "^1.0.19",
+    "@supabase/supabase-js": "^2.104.0",
     "ajv": "^8.18.0",
-    "bun-types": "^1.3.11",
+    "bun-types": "^1.3.13",
     "eslint": "9.39.4",
-    "supabase": "^2.88.1",
-    "typescript": "^6.0.2",
-    "vite": "^8.0.8",
+    "supabase": "^2.93.1",
+    "typescript": "^6.0.3",
+    "vite": "^8.0.10",
     "vite-plugin-dts": "^4.5.4"
   },
   "trustedDependencies": [

package/supabase/schemas/supabase-langchain/1_schemas.sql

@@ -0,0 +1,11 @@
+-- This file includes SQL commands for the langchain schema.
+-- This schema is used by services of the supabase-langchain package.
+
+-- Ensure the schema exists
+CREATE SCHEMA IF NOT EXISTS langchain;
+
+-- Grant access privileges
+GRANT USAGE ON SCHEMA langchain TO "anon";
+GRANT USAGE ON SCHEMA langchain TO "authenticated";
+GRANT USAGE ON SCHEMA langchain TO "service_role";
+

package/supabase/schemas/supabase-langchain/2_extensions.sql

@@ -0,0 +1,2 @@
+-- enable the "vector" extension for embeddings
+CREATE EXTENSION IF NOT EXISTS vector WITH SCHEMA extensions;

package/supabase/schemas/supabase-langchain/3_functions.sql

@@ -0,0 +1,71 @@
+-- Function to match documents based on a query embedding and optional metadata filter
+-- Returns the top 'match_count' documents ordered by similarity to the query embedding.
+-- The 'filter' parameter allows for additional filtering based on document metadata using json containment.
+-- Must be in the public schema to be compatible with langchain.
+CREATE OR REPLACE FUNCTION public.match_documents(
+    query_embedding vector(768),
+    match_count int,
+    filter jsonb DEFAULT '{}'::jsonb
+)
+RETURNS TABLE (
+    id uuid,
+    file_id uuid,
+    created_at timestamptz,
+    content text,
+    metadata jsonb,
+    embedding vector(768),
+    similarity float
+)
+SET search_path = ''
+AS $$
+#variable_conflict use_column
+BEGIN
+    RETURN QUERY
+    SELECT
+        id,
+        file_id,
+        created_at,
+        content,
+        metadata,
+        embedding,
+        1 - (embedding OPERATOR(extensions.<=>) query_embedding) AS similarity
+    FROM public.embeddings
+    WHERE metadata @> filter
+    ORDER BY embedding OPERATOR(extensions.<=>) query_embedding
+    LIMIT match_count;
+END;
+$$ LANGUAGE plpgsql;
+
+
+-- Function extracts the 'fileId' from metadata and assigns it to reference file_id.
+-- LangChain inserts embeddings automatically with a user defined metadata;
+-- this function is triggered before insert to populate the file_id column from the metadata.
+CREATE OR REPLACE FUNCTION langchain.embeddings_apply_metadata()
+RETURNS trigger
+SET search_path = ''
+AS $$
+BEGIN
+    IF NEW.metadata ? 'fileId' THEN
+        NEW.file_id := NEW.metadata ->> 'fileId';
+    END IF;
+    RETURN NEW;
+END;
+$$ LANGUAGE plpgsql;
+
+
+-- Function to return files in the specified bucket with no or outdated embeddings
+CREATE OR REPLACE FUNCTION langchain.get_outdated_embeddings(bucket text)
+RETURNS TABLE (
+    file_id uuid,
+    bucket_id uuid,
+    path_tokens text[]
+)
+SET search_path = ''
+AS $$
+BEGIN
+    RETURN QUERY
+    SELECT f.id, f.bucket_id, f.path_tokens
+    FROM langchain.outdated_embeddings AS f
+    WHERE f.bucket_id = bucket;
+END;
+$$ LANGUAGE plpgsql;

package/supabase/schemas/supabase-langchain/4_tables.sql

@@ -0,0 +1,48 @@
+-- ==========================================================
+-- Tables
+-- ==========================================================
+
+
+-- Create the embeddings table
+-- ---------------------------------------------------------
+-- The table holds embeddings for chunked text files.
+-- Must remain in the `public` schema for compatibility with LangChain.
+-- When a file is deleted from the storage.objects table,
+-- the corresponding embeddings are also deleted.
+CREATE TABLE IF NOT EXISTS public.embeddings (
+    id UUID PRIMARY KEY DEFAULT gen_random_uuid(),
+    file_id UUID NOT NULL REFERENCES storage.objects(id) ON DELETE CASCADE,  -- Not unique due to chunking
+    created_at timestamptz NOT NULL DEFAULT (now() AT TIME ZONE 'utc'),
+    content TEXT NOT NULL,
+    metadata JSONB,
+    embedding vector(768)  -- For nomic-embed-text (768 dimensions)
+);
+
+
+-- =========================================================
+-- Indexes
+-- =========================================================
+
+-- Vector similarity index (used by LangChain)
+CREATE INDEX IF NOT EXISTS public_embeddings_embedding_idx
+    ON public.embeddings
+    USING ivfflat (embedding vector_cosine_ops)
+    WITH (lists = 128);
+
+-- File lookup index
+CREATE INDEX IF NOT EXISTS public_embeddings_file_id_idx
+    ON public.embeddings (file_id);
+
+-- =========================================================
+-- Views
+-- =========================================================
+
+-- View to return files with no or outdated embeddings
+CREATE OR REPLACE VIEW langchain.outdated_embeddings AS
+SELECT f.*
+FROM storage.objects AS f
+LEFT JOIN public.embeddings AS e
+    ON e.file_id = f.id
+WHERE e.file_id IS NULL OR f.updated_at > e.created_at;
+
+GRANT SELECT ON langchain.outdated_embeddings TO authenticated, service_role, anon;

package/supabase/schemas/supabase-langchain/5_triggers.sql

@@ -0,0 +1,8 @@
+-- Trigger: Apply metadata before insert or update
+-- ---------------------------------------------------------
+-- LangChain inserts embeddings automatically; this trigger
+-- allows customizing the insert by populating columns based on the provided document metadata.
+CREATE TRIGGER langchain_embeddings_apply_metadata
+BEFORE INSERT OR UPDATE ON public.embeddings
+FOR EACH ROW
+EXECUTE FUNCTION langchain.embeddings_apply_metadata();