Extend Cloud Storage with Cloud Functions

You can trigger a function in response to the uploading, updating, or deleting of files and folders in Cloud Storage.

Examples in this page are based on a sample function that triggers when image files are uploaded to Cloud Storage. This sample function demonstrates how to access event attributes, how to download a file to a Cloud Functions instance, and other fundamentals of handling Cloud Storage events.

Import the required modules

To get started, import the module required for handling Cloud Storage events:

Node.js

const{onObjectFinalized}=require("firebase-functions/storage");

Python

 fromfirebase_functionsimport storage_fn

To build out the full sample, also add the dependencies for the Firebase Admin SDK and image processing tools:

Node.js

const{initializeApp}=require("firebase-admin/app");
const{getStorage}=require("firebase-admin/storage");
constlogger=require("firebase-functions/logger");
constpath=require("path");
// library for image resizing
constsharp=require("sharp");
initializeApp();

Python

 importio
importpathlib
fromPILimport Image
fromfirebase_adminimport initialize_app
initialize_app()
fromfirebase_adminimport storage

Scope a Cloud Storage function

Use the following pattern to scope your function to a specific Cloud Storage bucket and set any desired options:

Node.js

// scope handler to a specific bucket, using storage options parameter
exports.archivedopts=onObjectArchived({bucket:"myBucket"},(event)=>{
//...
});

Python

# Scope handler to a specific bucket using storage options parameter
@storage_fn.on_object_archived(bucket="myBucket")
defarchived_bucket(event: storage_fn.CloudEvent[storage_fn.StorageObjectData]):
 # ...

By contrast, the example thumbnail generator function is scoped to the default bucket for the project:

Node.js

exports.generateThumbnail=onObjectFinalized({cpu:2},async(event)=>{
// ...
});

Python

@storage_fn.on_object_archived()
defgeneratethumbnail(event: storage_fn.CloudEvent[storage_fn.StorageObjectData]):
 # ...

Set the function location

A mismatch between locations can result in deployment failure. Also, distance between the location of a Cloud Storage bucket and the location of the function can create significant network latency. To avoid these situations, specify the function location so that it matches the bucket/trigger location in one of these ways:

  • Function location is the same as the trigger location
  • Function location is inside the trigger location (when the trigger region is dual/multi region)
  • Function may be in any location if the trigger region is set to us-central1

Handle Cloud Storage events

These handlers for responding to Cloud Storage events are available:

Node.js

  • onObjectArchived Only sent when a bucket has enabled object versioning. This event indicates that the live version of an object has become an archived version, either because it was archived or because it was overwritten by the upload of an object of the same name.
  • onObjectDeleted Sent when an object has been permanently deleted. This includes objects that are overwritten or are deleted as part of the bucket's lifecycle configuration. For buckets with object versioning enabled, this is not sent when an object is archived (see onArchive), even if archival occurs via the storage.objects.delete method.
  • onObjectFinalized Sent when a new object (or a new generation of an existing object) is successfully created in the bucket. This includes copying or rewriting an existing object. A failed upload does not trigger this event.
  • onMetadataUpdated Sent when the metadata of an existing object changes.

Python

  • on_object_archived Only sent when a bucket has enabled object versioning. This event indicates that the live version of an object has become an archived version, either because it was archived or because it was overwritten by the upload of an object of the same name.
  • on_object_deleted Sent when an object has been permanently deleted. This includes objects that are overwritten or are deleted as part of the bucket's lifecycle configuration. For buckets with object versioning enabled, this is not sent when an object is archived (see onArchive), even if archival occurs via the storage.objects.delete method.
  • on_object_finalized Sent when a new object (or a new generation of an existing object) is successfully created in the bucket. This includes copying or rewriting an existing object. A failed upload does not trigger this event.
  • on_metadata_updated Sent when the metadata of an existing object changes.

Access Cloud Storage object attributes

Cloud Functions exposes a number of Cloud Storage object attributes such as the object's size and content type for the file updated. The metageneration attribute is incremented whenever there's a change to the object's metadata. For new objects, the metageneration value is 1.

Node.js

constfileBucket=event.data.bucket;// Storage bucket containing the file.
constfilePath=event.data.name;// File path in the bucket.
constcontentType=event.data.contentType;// File content type.

Python

bucket_name = event.data.bucket
file_path = pathlib.PurePath(event.data.name)
content_type = event.data.content_type

The thumbnail generation sample uses some of these attributes to detect exit cases in which the function returns:

Node.js

// Exit if this is triggered on a file that is not an image.
if(!contentType.startsWith("image/")){
returnlogger.log("This is not an image.");
}
// Exit if the image is already a thumbnail.
constfileName=path.basename(filePath);
if(fileName.startsWith("thumb_")){
returnlogger.log("Already a Thumbnail.");
}

Python

# Exit if this is triggered on a file that is not an image.
if not content_type or not content_type.startswith("image/"):
 print(f"This is not an image. ({content_type})")
 return
# Exit if the image is already a thumbnail.
if file_path.name.startswith("thumb_"):
 print("Already a thumbnail.")
 return

Download, transform, and upload a file

For some cases, it may not be necessary to download files from Cloud Storage. However, to perform intensive tasks such as generating a thumbnail image from a file stored in Cloud Storage, you need to download files to the functions instance—that is, the virtual machine that runs your code.

Using Cloud Functions together with image-processing programs like sharp for Node.js and Pillow for Python, you can perform manipulations on graphical image files. The following is an example of how to create a thumbnail image for an uploaded image file:

Node.js

/**
 * When an image is uploaded in the Storage bucket,
 * generate a thumbnail automatically using sharp.
 */
exports.generateThumbnail=onObjectFinalized({cpu:2},async(event)=>{
constfileBucket=event.data.bucket;// Storage bucket containing the file.
constfilePath=event.data.name;// File path in the bucket.
constcontentType=event.data.contentType;// File content type.
// Exit if this is triggered on a file that is not an image.
if(!contentType.startsWith("image/")){
returnlogger.log("This is not an image.");
}
// Exit if the image is already a thumbnail.
constfileName=path.basename(filePath);
if(fileName.startsWith("thumb_")){
returnlogger.log("Already a Thumbnail.");
}
// Download file into memory from bucket.
constbucket=getStorage().bucket(fileBucket);
constdownloadResponse=awaitbucket.file(filePath).download();
constimageBuffer=downloadResponse[0];
logger.log("Image downloaded!");
// Generate a thumbnail using sharp.
constthumbnailBuffer=awaitsharp(imageBuffer).resize({
width:200,
height:200,
withoutEnlargement:true,
}).toBuffer();
logger.log("Thumbnail created");
// Prefix 'thumb_' to file name.
constthumbFileName=`thumb_${fileName}`;
constthumbFilePath=path.join(path.dirname(filePath),thumbFileName);
// Upload the thumbnail.
constmetadata={contentType:contentType};
awaitbucket.file(thumbFilePath).save(thumbnailBuffer,{
metadata:metadata,
});
returnlogger.log("Thumbnail uploaded!");
});

Download the file to a temporary directory on your Cloud Functions instance. In this location, you can process the file as needed and then upload to Cloud Storage. When performing asynchronous tasks, make sure you return a JavaScript promise in your callback.

Python

@storage_fn.on_object_finalized()
defgeneratethumbnail(event: storage_fn.CloudEvent[storage_fn.StorageObjectData]):
"""When an image is uploaded in the Storage bucket, generate a thumbnail
 automatically using Pillow."""
 bucket_name = event.data.bucket
 file_path = pathlib.PurePath(event.data.name)
 content_type = event.data.content_type
 # Exit if this is triggered on a file that is not an image.
 if not content_type or not content_type.startswith("image/"):
 print(f"This is not an image. ({content_type})")
 return
 # Exit if the image is already a thumbnail.
 if file_path.name.startswith("thumb_"):
 print("Already a thumbnail.")
 return
 bucket = storage.bucket(bucket_name)
 image_blob = bucket.blob(str(file_path))
 image_bytes = image_blob.download_as_bytes()
 image = Image.open(io.BytesIO(image_bytes))
 image.thumbnail((200, 200))
 thumbnail_io = io.BytesIO()
 image.save(thumbnail_io, format="png")
 thumbnail_path = file_path.parent / pathlib.PurePath(f"thumb_{file_path.stem}.png")
 thumbnail_blob = bucket.blob(str(thumbnail_path))
 thumbnail_blob.upload_from_string(thumbnail_io.getvalue(), content_type="image/png")

This code creates a 200x200 thumbnail for the image saved in a temporary directory, then uploads it back to Cloud Storage.

Except as otherwise noted, the content of this page is licensed under the Creative Commons Attribution 4.0 License, and code samples are licensed under the Apache 2.0 License. For details, see the Google Developers Site Policies. Java is a registered trademark of Oracle and/or its affiliates.

Last updated 2025年10月01日 UTC.