🔧 fix: Azure Blob Integration and File Source References (#6575)

* 🔧 fix: Update file source references to include 'azure_blob' for correct service initialization

* 🔧 fix: Add Azure Blob Storage Emulator entries to .gitignore

* fix: Update file source references to include 'azure_blob' for correct service initialization

* fix: Refactor Azure Blob Storage functions to use environment variables for access control and container name, fix deletion improper logging and improper params

* fix: Add basePath determination for agent file uploads based on MIME type

* fix: Implement file streaming to Azure Blob Storage to optimize memory usage during uploads (non-images)

* fix: Update SourceIcon to include 'azure_blob' class and adjust model setting in useSelectorEffects for assistants

* chore: import order

---------

Co-authored-by: Danny Avila <danny@librechat.ai>

.gitignore

@@ -37,6 +37,10 @@ client/public/main.js
 client/public/main.js.map
 client/public/main.js.LICENSE.txt
 
+# Azure Blob Storage Emulator (Azurite)
+__azurite**
+__blobstorage__/**/*
+
 # Dependency directorys
 # Deployed apps should consider commenting these lines out:
 # see https://npmjs.org/doc/faq.html#Should-I-check-my-node_modules-folder-into-git

api/server/services/AppService.js

@@ -52,7 +52,7 @@ const AppService = async (app) => {
 
   if (fileStrategy === FileSources.firebase) {
     initializeFirebase();
-  } else if (fileStrategy === FileSources.azure) {
+  } else if (fileStrategy === FileSources.azure_blob) {
     initializeAzureBlobService();
   } else if (fileStrategy === FileSources.s3) {
     initializeS3();

api/server/services/Files/Azure/crud.js

@@ -1,11 +1,13 @@
 const fs = require('fs');
 const path = require('path');
+const mime = require('mime');
 const axios = require('axios');
 const fetch = require('node-fetch');
 const { logger } = require('~/config');
 const { getAzureContainerClient } = require('./initialize');
 
 const defaultBasePath = 'images';
+const { AZURE_STORAGE_PUBLIC_ACCESS = 'true', AZURE_CONTAINER_NAME = 'files' } = process.env;
 
 /**
  * Uploads a buffer to Azure Blob Storage.
@@ -29,10 +31,9 @@ async function saveBufferToAzure({
 }) {
   try {
     const containerClient = getAzureContainerClient(containerName);
+    const access = AZURE_STORAGE_PUBLIC_ACCESS?.toLowerCase() === 'true' ? 'blob' : undefined;
     // Create the container if it doesn't exist. This is done per operation.
-    await containerClient.createIfNotExists({
-      access: process.env.AZURE_STORAGE_PUBLIC_ACCESS ? 'blob' : undefined,
-    });
+    await containerClient.createIfNotExists({ access });
     const blobPath = `${basePath}/${userId}/${fileName}`;
     const blockBlobClient = containerClient.getBlockBlobClient(blobPath);
     await blockBlobClient.uploadData(buffer);
@@ -97,25 +98,21 @@ async function getAzureURL({ fileName, basePath = defaultBasePath, userId, conta
  * Deletes a blob from Azure Blob Storage.
  *
  * @param {Object} params
- * @param {string} params.fileName - The name of the file.
- * @param {string} [params.basePath='images'] - The base folder where the file is stored.
- * @param {string} params.userId - The user's id.
- * @param {string} [params.containerName] - The Azure Blob container name.
+ * @param {ServerRequest} params.req - The Express request object.
+ * @param {MongoFile} params.file - The file object.
  */
-async function deleteFileFromAzure({
-  fileName,
-  basePath = defaultBasePath,
-  userId,
-  containerName,
-}) {
+async function deleteFileFromAzure(req, file) {
   try {
-    const containerClient = getAzureContainerClient(containerName);
-    const blobPath = `${basePath}/${userId}/${fileName}`;
+    const containerClient = getAzureContainerClient(AZURE_CONTAINER_NAME);
+    const blobPath = file.filepath.split(`${AZURE_CONTAINER_NAME}/`)[1];
+    if (!blobPath.includes(req.user.id)) {
+      throw new Error('User ID not found in blob path');
+    }
     const blockBlobClient = containerClient.getBlockBlobClient(blobPath);
     await blockBlobClient.delete();
     logger.debug('[deleteFileFromAzure] Blob deleted successfully from Azure Blob Storage');
   } catch (error) {
-    logger.error('[deleteFileFromAzure] Error deleting blob:', error.message);
+    logger.error('[deleteFileFromAzure] Error deleting blob:', error);
     if (error.statusCode === 404) {
       return;
     }
@@ -123,6 +120,65 @@ async function deleteFileFromAzure({
   }
 }
 
+/**
+ * Streams a file from disk directly to Azure Blob Storage without loading
+ * the entire file into memory.
+ *
+ * @param {Object} params
+ * @param {string} params.userId - The user's id.
+ * @param {string} params.filePath - The local file path to upload.
+ * @param {string} params.fileName - The name of the file in Azure.
+ * @param {string} [params.basePath='images'] - The base folder within the container.
+ * @param {string} [params.containerName] - The Azure Blob container name.
+ * @returns {Promise<string>} The URL of the uploaded blob.
+ */
+async function streamFileToAzure({
+  userId,
+  filePath,
+  fileName,
+  basePath = defaultBasePath,
+  containerName,
+}) {
+  try {
+    const containerClient = getAzureContainerClient(containerName);
+    const access = AZURE_STORAGE_PUBLIC_ACCESS?.toLowerCase() === 'true' ? 'blob' : undefined;
+
+    // Create the container if it doesn't exist
+    await containerClient.createIfNotExists({ access });
+
+    const blobPath = `${basePath}/${userId}/${fileName}`;
+    const blockBlobClient = containerClient.getBlockBlobClient(blobPath);
+
+    // Get file size for proper content length
+    const stats = await fs.promises.stat(filePath);
+
+    // Create read stream from the file
+    const fileStream = fs.createReadStream(filePath);
+
+    const blobContentType = mime.getType(fileName);
+    await blockBlobClient.uploadStream(
+      fileStream,
+      undefined, // Use default concurrency (5)
+      undefined, // Use default buffer size (8MB)
+      {
+        blobHTTPHeaders: {
+          blobContentType,
+        },
+        onProgress: (progress) => {
+          logger.debug(
+            `[streamFileToAzure] Upload progress: ${progress.loadedBytes} bytes of ${stats.size}`,
+          );
+        },
+      },
+    );
+
+    return blockBlobClient.url;
+  } catch (error) {
+    logger.error('[streamFileToAzure] Error streaming file:', error);
+    throw error;
+  }
+}
+
 /**
  * Uploads a file from the local file system to Azure Blob Storage.
  *
@@ -146,18 +202,19 @@ async function uploadFileToAzure({
 }) {
   try {
     const inputFilePath = file.path;
-    const inputBuffer = await fs.promises.readFile(inputFilePath);
-    const bytes = Buffer.byteLength(inputBuffer);
+    const stats = await fs.promises.stat(inputFilePath);
+    const bytes = stats.size;
     const userId = req.user.id;
     const fileName = `${file_id}__${path.basename(inputFilePath)}`;
-    const fileURL = await saveBufferToAzure({
+
+    const fileURL = await streamFileToAzure({
       userId,
-      buffer: inputBuffer,
+      filePath: inputFilePath,
       fileName,
       basePath,
       containerName,
     });
-    await fs.promises.unlink(inputFilePath);
+
     return { filepath: fileURL, bytes };
   } catch (error) {
     logger.error('[uploadFileToAzure] Error uploading file:', error);

api/server/services/Files/images/encode.js

@@ -37,7 +37,7 @@ const base64Only = new Set([
   EModelEndpoint.bedrock,
 ]);
 
-const blobStorageSources = new Set([FileSources.azure, FileSources.s3]);
+const blobStorageSources = new Set([FileSources.azure_blob, FileSources.s3]);
 
 /**
  * Encodes and formats the given files.

api/server/services/Files/process.js

@@ -492,7 +492,7 @@ const processAgentFileUpload = async ({ req, res, metadata }) => {
 
   let fileInfoMetadata;
   const entity_id = messageAttachment === true ? undefined : agent_id;
-
+  const basePath = mime.getType(file.originalname)?.startsWith('image') ? 'images' : 'uploads';
   if (tool_resource === EToolResources.execute_code) {
     const isCodeEnabled = await checkCapability(req, AgentCapabilities.execute_code);
     if (!isCodeEnabled) {
@@ -532,7 +532,7 @@ const processAgentFileUpload = async ({ req, res, metadata }) => {
       images,
       filename,
       filepath: ocrFileURL,
-    } = await handleFileUpload({ req, file, file_id, entity_id: agent_id });
+    } = await handleFileUpload({ req, file, file_id, entity_id: agent_id, basePath });
 
     const fileInfo = removeNullishValues({
       text,
@@ -582,6 +582,7 @@ const processAgentFileUpload = async ({ req, res, metadata }) => {
     file,
     file_id,
     entity_id,
+    basePath,
   });
 
   let filepath = _filepath;

api/server/services/Files/strategies.js

@@ -211,6 +211,8 @@ const getStrategyFunctions = (fileSource) => {
   } else if (fileSource === FileSources.openai) {
     return openAIStrategy();
   } else if (fileSource === FileSources.azure) {
+    return openAIStrategy();
+  } else if (fileSource === FileSources.azure_blob) {
     return azureStrategy();
   } else if (fileSource === FileSources.vectordb) {
     return vectorStrategy();

client/src/components/Chat/Input/Files/SourceIcon.tsx

@@ -10,7 +10,8 @@ const sourceToEndpoint = {
 
 const sourceToClassname = {
   [FileSources.openai]: 'bg-white/75 dark:bg-black/65',
-  [FileSources.azure]: 'azure-bg-color opacity-85',
+  [FileSources.azure]: 'azure-bg-color',
+  [FileSources.azure_blob]: 'azure-bg-color',
   [FileSources.execute_code]: 'bg-black text-white opacity-85',
   [FileSources.text]: 'bg-blue-500 dark:bg-blue-900 opacity-85 text-white',
   [FileSources.vectordb]: 'bg-yellow-700 dark:bg-yellow-900 opacity-85 text-white',

client/src/hooks/Endpoint/useSelectorEffects.ts

@@ -61,7 +61,7 @@ export default function useSelectorEffects({
       }
       const assistant = assistantsMap?.[endpoint ?? '']?.[assistant_id];
       if (assistant !== undefined) {
-        setOption('model')('');
+        setOption('model')(assistant.model);
         setOption('assistant_id')(assistant_id);
       }
     }

packages/data-provider/src/types/files.ts

@@ -4,6 +4,7 @@ export enum FileSources {
   local = 'local',
   firebase = 'firebase',
   azure = 'azure',
+  azure_blob = 'azure_blob',
   openai = 'openai',
   s3 = 's3',
   vectordb = 'vectordb',