AWS S3 capabilities for iOS - ios

I decided to utilize Amazon S3 to upload files, however I find AWS Docs a bit confusing about S3 capabilities for iOS platform.
I would like to know how my app would act in the following scenarios:
Scenario 1: During the upload user has accidentally lost internet connection
Scenario 2: App crashes during the upload
I heard that iOS SDK takes care of such issues itself by resuming remaining upload when possible, I failed to find relevant information in docs, though.
Will AWSS3 framework cover both this scenarios? Does it need any additional lines od code to not be vulnerable for potential crashes and network errors?
I've found some relevant informations for Android platform
I'd love to know what can I expect from the following code
let image = UIImage(named: "12.jpeg")
let fileManager = FileManager.default
let imageData = UIImageJPEGRepresentation(image!, 0.99)
let path = (NSSearchPathForDirectoriesInDomains(.documentDirectory, .userDomainMask, true)[0] as NSString).appendingPathComponent("\(imageData!).jpeg")
fileManager.createFile(atPath: path as String, contents: imageData, attributes: nil)
let fileUrl = NSURL(fileURLWithPath: path)
let uploadRequest = AWSS3TransferManagerUploadRequest()
uploadRequest?.bucket = "bucketname"
uploadRequest?.key = "folder/12.jpeg"
uploadRequest?.contentType = "image/jpeg"
uploadRequest?.body = fileUrl as URL!
uploadRequest?.serverSideEncryption = AWSS3ServerSideEncryption.awsKms
uploadRequest?.uploadProgress = { (bytesSent, totalBytesSent, totalBytesExpectedToSend) -> Void in
DispatchQueue.main.async(execute: {
print("bytes sent \(bytesSent), total bytes sent \(totalBytesSent), of total \(totalBytesExpectedToSend)")
})
}
let transferManager = AWSS3TransferManager.default()
transferManager?.upload(uploadRequest).continue(with: AWSExecutor.mainThread(), withSuccessBlock: { (taskk: AWSTask) -> Any? in
if taskk.error != nil {
// Error.
} else {
// Do something with your result.
}
return nil
})
Is it already crash/network proof?
EDIT:
This is the part of docs that sounds ambiguous to me:
S3 provides a multipart upload feature that lets you upload a single
object as a set of parts. Each part is a contiguous portion of the
object's data, and the object parts are uploaded independently and in
any order. If transmission of any part fails, you can retransmit that
part without affecting other parts. After all parts of the object are
uploaded, S3 assembles these parts and creates the object.
Does that mean it has its own inherent mechanism to manage that? Let's say I kill app during uploading a file, when I relaunch it and start over the upload process , will it start with the last chunk where it left off before I killed the app?

Related

Saving a file without document directory ios

hope you are doing good in development. I have a question regarding Zipping currently i am using Zip framework which I am using for zip the all captured image in device but the problem here is I do not want to save the Zip file in document directory instead I wanna save this memory itself. I am struggling since last week please let me know how can we achieve the zipping without saving it in local directory filepath
Also when I captured the image successfully I saved it to local directory using
let documentsDirectory = FileManager.default.urls(for:
.documentDirectory, in: .userDomainMask).first!
let fileName = "image.jpg"
let fileURL = documentsDirectory.appendingPathComponent(fileName)
if let data = UIImageJPEGRepresentation(image, 1.0),
!FileManager.default.fileExists(atPath: fileURL.path) {
do {
// writes the image data to disk
try data.write(to: fileURL)
print("file saved")
} catch {
print("error saving file:", error)
}
}
After when I am trying to send the image before our server I want to make it all one file so I have implemented the Zip (framework)
Here filpathArray meant all captured image in local path
do {
var urls = [URL]()
for i in 0..<(self.filePathArray.count)
{
urls.append(self.filePathArray[i] as! URL)
}
if self.filePathArray.count > 0 {
let zipFiles = try Zip.quickZipFiles(urls, fileName: "AllFiles")
}
}
Here zipping is done successfully but it saved in local path so when using app container I can able to see the images, I do not want to see the images in device like apple sandbox or Xcode container itself
I want to make zipping on the flow like without saving it document directory, Thanks in advance.

SCNParticleSystem load from document Directory

I am trying load SCNParticleSystem from download bundle which i am not able to load.
Path for the resource.
file:///var/mobile/Containers/Data/Application/A91E9970-CDE1-43D8-B822-4B61EFC6149B/Documents/so/solarsystem.bundle/Contents/Resources/
let objScene = SCNParticleSystem(named: "stars", inDirectory: directory)
This object is nil.
This is a legitimate problem since SceneKit does not provide an out-of-the-box solution for initializing particle systems from files that are outside of the main bundle (the only init method SCNParticleSystem.init(named:inDirectory:) implies that SCNParticleSystem.scnp files are in the main bundle).
Luckily for us .scnp files are just encoded/archived SCNParticleSystem instances that we can easily decode/unarchive using NSKeyedUnarchiver:
extension SCNParticleSystem {
static func make(fromFileAt url: URL) -> SCNParticleSystem? {
guard let data = try? Data(contentsOf: url),
let object = try? NSKeyedUnarchiver.unarchiveTopLevelObjectWithData(data),
let system = object as? SCNParticleSystem else { return nil }
return system
}
}
If you do not need to support iOS 9 and iOS 10 you can use NSKeyedUnarchiver.unarchivedObject(ofClass: SCNParticleSystem.self, from: data) instead of NSKeyedUnarchiver.unarchiveTopLevelObjectWithData(_:) and type casting, which was introduced in iOS 11.0.
Another issue that you're most likely to encounter is missing particle images. That is because by default SceneKit will look for them in the main bundle. As of current versions of iOS (which is iOS 12) and Xcode (Xcode 10) particle images in .scnp files (particleImage property) are String values which are texture filenames in the main bundle (that might change, but probably won't, however there's not much else we could use).
So my suggestion is to take that filename and look for the texture file with the same name in the same directory where the .scnp file is:
extension SCNParticleSystem {
static func make(fromFileAt url: URL) -> SCNParticleSystem? {
guard let data = try? Data(contentsOf: url),
let object = try? NSKeyedUnarchiver.unarchiveTopLevelObjectWithData(data),
let system = object as? SCNParticleSystem else { return nil }
if let particleImageName = system.particleImage as? String {
let particleImageURL = url
.deletingLastPathComponent()
.appendingPathComponent(particleImageName)
if FileManager.default.fileExists(atPath: particleImageURL.path) {
system.particleImage = particleImageURL
}
}
return system
}
}
You can just set the URL of the image file and SceneKit will handle it from there.
As a little side-note, the recommended directory for downloadable content is Application Support directory, not Documents.
Application Support: Use this directory to store all app data files except those associated with the user’s documents. For example, you might use this directory to store app-created data files, configuration files, templates, or other fixed or modifiable resources that are managed by the app. An app might use this directory to store a modifiable copy of resources contained initially in the app’s bundle. A game might use this directory to store new levels purchased by the user and downloaded from a server.
(from File System Basics)
Don't have enough reps to add the comment so adding it as the answer.
The answer by Lësha Turkowski works for sure but was had issues with loading the particle images using only NSURL.
All particles were appearing square which meant,
If the value is nil (the default), SceneKit renders each particle as a
small white square (colorized by the particleColor property).
SCNParticleSystem particleImage
In the documentation it says You may specify an image using an
NSImage (in macOS) or UIImage (in iOS) instance, or an NSString or
NSURL instance containing the path or URL to an image file.
Instead of using the NSURL, ended up using the UIImage and it loaded up fine.
extension SCNParticleSystem {
static func make(fromFileAt url: URL) -> SCNParticleSystem? {
guard let data = try? Data(contentsOf: url),
let object = try? NSKeyedUnarchiver.unarchiveTopLevelObjectWithData(data),
let system = object as? SCNParticleSystem else { return nil }
if let particleImageName = system.particleImage as? String {
let particleImageURL = url
.deletingLastPathComponent()
.appendingPathComponent(particleImageName)
if FileManager.default.fileExists(atPath: particleImageURL.path) {
// load up the NSURL contents in UIImage
let particleUIImage = UIImage(contentsOfFile: particleImageURL.path)
system.particleImage = particleUIImage
}
}
return system
}
}
I found out, that sometimes when dragging a SCNParticleSystem file into your project (probably form a different project) a silent error can happen due to some bugs in Xcode. As a result you can't get a reference to an instance of your SCNParticleSystem.
Solution: Check your BuildSettings in your target. The SCNPaticleSystem AND the associated ImageFile should be listed there and then you should get it right. (see screenShot below)

Extensive memory usage when uploading assets (images, videos) to firebase in Swift?

Suppose I have an array of UIImage called photos, they are to be uploaded to Firebase storage. I wish to do the following things:
Upload them to Firebase storage
Get paths of the uploaded photos and store in an array called uploadedAssets (paths, not download url, it looks like this: "photos/folder_name/photo_id"), where "folder_name" is randomly generated and "photo_id" is an integer, representing the order of photos
Call Cloud Function and pass uploadedAssets to it. The server then uses the paths to find all pictures and generates a thumbnail for each one.
Finally, store the original photos' download urls and thumbnails' download urls in database.
I have something that's working, but uses too much memory (300+MB when uploading only 4 pictures):
// Swift
let dispatchGroup = DispatchGroup()
let dispatchQueue = DispatchQueue.init(label: "AssetQueue")
var uploadedAssets = [String]()
let folderName: String = UUID().uuidString
dispatchQueue.async {
for i in 0..<photos.count {
dispatchGroup.enter()
let photo: UIImage = photos[i]
let fileName: String = "\(folderName)/\(i)"
let assetRef = Storage.storage().reference().child("photos/\(fileName)")
let metaData = StorageMetaData()
metaData.contentType = "image/jpg"
if let dataToUpload = UIImageJPEGRepresentation(photo, 0.75) {
assetRef.putData(
dataToUpload,
metaData: metaData,
completion: { (_, error) in
uploadedAssets.append("photos/\(fileName)")
dispatchGroup.leave()
}
)
}
}
}
dispatchGroup.notify(queue: dispatchQueue) {
Alamofire.request(
"https://<some_url>",
method: .post,
parameters: [
"uploadedAssets": uploadedAssets
]
)
}
And the code that generates thumbnails runs on server side, therefore, in my opinion, is irrelevant, I won't post it here. So, the above code snippet consumes 300+MB of memory when there are 4 photos to upload. After successfully uploaded those photos, the memory usage stays at 300+MB and never drops. When I try to upload more, say another 4 photos, it could even go up to 450+MB. I know that's not normal, but can't seem to figure out why this would happen?

cloudkit error no authToken received for asset

Why do I get this error when I run the following code? :
"Internal Error" (1/1000); "No authToken received for asset"
I think it has something to do with the setObject code in the last line.
let documentsDirectoryPath:NSString = NSSearchPathForDirectoriesInDomains(.documentDirectory, .userDomainMask, true)[0] as NSString
var imageURL: URL!
let imageData = UIImageJPEGRepresentation(self.newImage, 1.0)
let path:String = documentsDirectoryPath.appendingPathComponent(self.newImage.description)
try? UIImageJPEGRepresentation(self.newImage, 1.0)!.write(to: URL(fileURLWithPath: path), options: [.atomicWrite])
imageURL = URL(fileURLWithPath: path)
try? imageData?.write(to: imageURL, options: [.atomicWrite])
let imageAsset:CKAsset? = CKAsset(fileURL: URL(fileURLWithPath: path))
curImages = record["Images"] as! [CKAsset]
curImages.append(imageAsset!)
print("saving image")
record.setObject(curImages as CKRecordValue?, forKey: "Images")
I've encountered this, too. It appears to be a bug in cloudkit, and--from what I can tell--it happens when you try to re-use any part of the "asset creation chain."
In other words, you have some initial data, you create an image from that data, you write it to a file, you load that file into a CKAsset, then you load the CKAsset into the CKRecrod. In my experiments, if you re-use any of those components... or if they just happen to be the same (that is, you create an image, then you happen to create a new-but-identical image later) you'll see this error.
For example, the following code reliably recreates the "no auth token" error when saving a record. All it does is create an array of assets and places it into the record:
for (int i = 0; i <= maxPlayers; i++)
{
int tempVal = 0xf;
NSData *tempData = [[NSData alloc] initWithBytes:&tempVal length:sizeof(tempVal)];
NSString *tempDataFilepath = [documentsDirectory stringByAppendingPathComponent:[NSString stringWithFormat:#"temp%d.dat",i]];
[tempData writeToFile:tempDataFilepath atomically:YES];
NSURL *tempDataURL = [NSURL fileURLWithPath:tempDataFilepath];
someArray[i] = [[CKAsset alloc] initWithFileURL:tempDataURL ];
}
someRecord[SOME_FIELD_NAME] = someArray;
Simply changing the third line to:
int tempVal = i; //force the temp value to be different every time
Completely solves the error.
Furthermore, this error occurs even when I tried to use a value in a different CKAsset **that was already used in a prior CKAsset For example, using int tempVal = 0xf in the first asset, then using int secondTempVal = 0xf in another CKAsset also produces the "no auth token" error.
In my case, I was able to force the asset value to always be a unique value, and completely solved the problem. In your case, I suggest the following possible work arounds:
Check if you're using identical images for your assets. If you are, try slightly modifying the images for each new CKAsset.
If you must re-use identical images, try saving the record after you set each asset. I do not know if that will solve the issue, and it certainly increases your network traffic. But it's worth an experiment to see if it helps.
In this question Saving CKAsset to CKRecord in CloudKit produces error: "No authToken received for asset" the OP was able to create separate copies of the image file that ultimately solved the problem.
Open a bug with Apple. I didn't bother doing this, as I've grown jaded watching similar bug reports sit open for years without attention. But who knows, you might have better luck.
This is not an answer to the specific problem (which as been solved by the accepted answer), but it solves another problem that creates the same error message, so it might be useful for somebody else:
I have an app that uses CoreData+Cloudkit with a .public database, i.e. my description for the NSPersistentCloudKitContainer uses
description.cloudKitContainerOptions!.databaseScope = .public
Whenever I change my CloudKit schema, I have to re-initialize it using
do {
try self.initializeCloudKitSchema()
} catch {
print("Could not initialize schema, error \(error)")
}
This creates the error
"Internal Error" (1/1000); "No authToken received for asset"
although I do not use any asset in my model.
I now realized that it has something to do with the .public database:
As soon as I out-comment the instruction that sets the database scope to .public, the re-initialization works without problems.
Now the CloudKit schema is independent of the database type (.private or .public). Thus, a re-initialization of the schema requires the following:
Set the database to .private (the default)
Execute the init code
Set the database to .public
Disable the init code
PS: I know I should write now a bug report, but I stopped doing so: Nearly none of my bug reports (about 15) have ever been answered or processed, so it is not worth the effort.

Get PHAsset from iOS Share Extension

I am developing a share extension for photos for my iOS app. Inside the extension, I am able to successfully retrieve the UIImage object from the NSItemProvider.
However, I would like to be able to share the image with my container app, without having to store the entire image data inside my shared user defaults. Is there a way to get the PHAsset of the image that the user has chosen in the share extension (if they have picked from their device)?
The documentation on the photos framework (https://developer.apple.com/library/ios/documentation/Photos/Reference/Photos_Framework/) has a line that says "This architecture makes it easy, safe, and efficient to work with the same assets from multiple threads or multiple apps and app extensions."
That line makes me think there is a way to share the same PHAsset between extension and container app, but I have yet to figure out any way to do that? Is there a way to do that?
This only works if the NSItemProvider gives you a URL with the format:
file:///var/mobile/Media/DCIM/100APPLE/IMG_0007.PNG
which is not always true for all your assets, but if it returns a URL as:
file:///var/mobile/Media/PhotoData/OutgoingTemp/2AB79E02-C977-4B4A-AFEE-60BC1641A67F.JPG
then PHAsset will never find your asset. Further more, the latter is a copy of your file, so if you happen to have a very large image/video, iOS will duplicate it in that OutgoingTemp directory. Nowhere in the documentation says when it's going to be deleted, hopefully soon enough.
I think this is a big gap Apple has left between Sharing Extensions and PHPhotoLibrary framework. Apple should've be creating an API to close it, and soon.
You can get PHAsset if image is shared from Photos app. The item provider will give you a URL that contains the image's filename, you use this to match PHAsset.
/// Assets that handle through handleImageItem:completionHandler:
private var handledAssets = [PHAsset]()
/// Key is the matched asset's original file name without suffix. E.g. IMG_193
private lazy var imageAssetDictionary: [String : PHAsset] = {
let options = PHFetchOptions()
options.includeHiddenAssets = true
let fetchResult = PHAsset.fetchAssetsWithOptions(options)
var assetDictionary = [String : PHAsset]()
for i in 0 ..< fetchResult.count {
let asset = fetchResult[i] as! PHAsset
let fileName = asset.valueForKey("filename") as! String
let fileNameWithoutSuffix = fileName.componentsSeparatedByString(".").first!
assetDictionary[fileNameWithoutSuffix] = asset
}
return assetDictionary
}()
...
provider.loadItemForTypeIdentifier(imageIdentifier, options: nil) { imageItem, _ in
if let image = imageItem as? UIImage {
// handle UIImage
} else if let data = imageItem as? NSData {
// handle NSData
} else if let url = imageItem as? NSURL {
// Prefix check: image is shared from Photos app
if let imageFilePath = imageURL.path where imageFilePath.hasPrefix("/var/mobile/Media/") {
for component in imageFilePath.componentsSeparatedByString("/") where component.containsString("IMG_") {
// photo: /var/mobile/Media/DCIM/101APPLE/IMG_1320.PNG
// edited photo: /var/mobile/Media/PhotoData/Mutations/DCIM/101APPLE/IMG_1309/Adjustments/FullSizeRender.jpg
// cut file's suffix if have, get file name like IMG_1309.
let fileName = component.componentsSeparatedByString(".").first!
if let asset = imageAssetDictionary[fileName] {
handledAssets.append(asset)
imageCreationDate = asset.creationDate
}
break
}
}
}

Resources