How can I clone FileStream type?
The problem is that I using MagickNet.Image(inputStream) to get the image distentions before saving it, but it seems to closing the inputStream.
So, I can send a clone of inputStream to the function, so the function cannot edit the real inputStream object?
This is my function:
public Picture GetPictureDimension(Stream inputStream, Picture picture)
{
var img = new MagickNet.Image(inputStream);
picture.Width = img.Columns;
picture.Height = img.Rows;
return picture;
}
You could just re-open the file? But to keep a single Stream without it getting closed, you could try NonClosingStreamWrapper in MiscUtil. Just be sure to reset the Position if appropriate.
You didn't post full code, but I imagine that MagickNet.Image(inputStream) gets the full image and you only use the Size from that, and later you load the Image a second time.
So a more practical solution would be to get and hold the Image in memory just once. That gives you access to the Size.
Edit:
You don't seem to realize it but you are asking how to load the Image twice (w/o reopening the stream). I do think it is more efficient to load it just once.
Picture is a XNA class, right? I don't know to much about that but you could try something like:
public Picture GetPictureDimension(Stream inputStream, ref Picture picture)
{
var img = new MagickNet.Image(inputStream);
picture = new Picture(img); // just guessing here
//picture.Width = img.Columns;
//picture.Height = img.Rows;
return picture;
}
Related
I have build a program which allows me to insert comment and the title of an Image through System.Image.Drawing so right now, I have trouble trying to overwrite the existing Jpeg file with the one that has comment and title added into it, but I am encountering error while deleting the file, so I'm not sure what to do as I have tried disposing the file but I cant saved it in that case, due to the fact that I disposed it too early, but I cant saved it because the existing file name is not deleted so I'm kinda stuck in the middle right now.
Here are my codes for it:
public string EditComment(string OriginalFilepath, string newFilename)
{
image = System.Drawing.Image.FromFile(OriginalFilepath);
PropertyItem propItem = image.PropertyItems[0];
using (var file = System.Drawing.Image.FromFile(OriginalFilepath))
{
propItem.Id = 0x9286; // this is the id for 'UserComment'
propItem.Type = 2;
propItem.Value = System.Text.Encoding.UTF8.GetBytes("HelloWorld\0");
propItem.Len = propItem.Value.Length;
file.SetPropertyItem(propItem);
PropertyItem propItem1 = file.PropertyItems[file.PropertyItems.Count() - 1];
file.Dispose();
image.Dispose();
string filepath = Filepath;
if (File.Exists(#"C:\Desktop\Metadata"))
{
System.IO.File.Delete(#"C:\Desktop\Metadata");
}
string newFilepath = filepath + newFilename;
file.Save(newFilepath, ImageFormat.Jpeg);//error appears here
return filepath;
}
}
The Error shown are:
An exception of type 'System.ArgumentException' occurred in System.Drawing.dll but was not handled in user code
Additional information: Parameter is not valid.
The problem is that opening an image from file locks the file. You can get around that by reading the file into a byte array, creating a memory stream from that, and then opening the image from that stream:
public string EditComment(string originalFilepath, string newFilename)
{
Byte[] bytes = File.ReadAllBytes(originalFilepath);
using (MemoryStream stream = new MemoryStream(bytes))
using (Bitmap image = new Bitmap(stream))
{
PropertyItem propItem = image.PropertyItems[0];
// Processing code
propItem.Id = 0x9286; // this is the id for 'UserComment'
propItem.Type = 2;
propItem.Value = System.Text.Encoding.UTF8.GetBytes("HelloWorld\0");
propItem.Len = propItem.Value.Length;
image.SetPropertyItem(propItem);
// Not sure where your FilePath comes from but I'm just
// putting it in the same folder with the new name.
String newFilepath;
if (newFilename == null)
newFilepath = originalFilePath;
else
newFilepath = Path.Combine(Path.GetDirectory(originalFilepath), newFilename);
image.Save(newFilepath, ImageFormat.Jpeg);
return newFilepath;
}
}
Make sure you do not dispose your image object inside the using block as you did in your test code. Not only does the using block exist exactly so you don't have to dispose manually, but it's also rather hard to save an image to disk that no longer exists in memory. Similarly, you seem to open the image from file twice. I'm just going to assume all of those were experiments to try to get around the problem, but do make sure to clean those up.
The basic rules to remember when opening images are these:
An Image object created from a file will lock the file during the life cycle of the image object, preventing the file from being overwritten or deleted until the image is disposed.
An Image object created from a stream will need the stream to remain open for the entire life cycle of the image object. Unlike with files, there is nothing actively enforcing this, but after the stream is closed, the image will give errors when saved, cloned or otherwise manipulated.
Contrary to what some people believe, a basic .Clone() call on the image object will not change this behaviour. The cloned object will still keep the reference to the original source.
Note, if you actually want a usable image object that is not contained in a using block, you can use LockBits and Marshal.Copy to copy the byte data of the image object into a new image with the same dimensions and the same PixelFormat, effectively making a full data clone of the original image. (Note: I don't think this works on animated GIF files) After doing that, you can safely dispose the original and just use the new cleanly-cloned version.
There are some other workarounds for actually getting the image out, but most of them I've seen aren't optimal. These are the two most common other valid workarounds for the locking problem:
Create a new Bitmap from an image loaded from file using the Bitmap(Image image) constructor. This new object will not have the link to that file, leaving you free to dispose the one that's locking the file. This works perfectly, but it changes the image's colour depth to 32-bit ARGB, which might not be desired. If you just need to show an image on the UI, this is an excellent solution, though.
Create a MemoryStream as shown in my code, but not in a using block, leaving the stream open as required. Leaving streams open doesn't really seem like a good idea to me. Though some people have said that, since a MemoryStream is just backed by a simple array, and not some external resource, the garbage collector apparently handles this case fine...
I've also seen some people use System.Drawing.ImageConverter to convert from bytes, but I looked into the internals of that process, and what it does is actually identical to the last method here, which leaves a memory stream open.
I want to know how many frames a TIFF picture contains. However, to minimize the execution time, I would like to know this information without actually loading the full file. How can I achieve this?
For your information, I am using the following code right now:
FileStream mystream = new FileStream("mypicture.tif", FileMode.Open);
_Image _currentImg = Image.FromStream(mystream );
mystream.Close();
FrameDimension myframeDimensions = new
FrameDimension(_currentImg.FrameDimensionsList[0]);
Int32 numberOfFrames = _currentImg.GetFrameCount(myframeDimensions );
I'm very new to this stuff of saving images to the DB, and even when I thought it was very straight forward, it wasn't. What I'm trying to do is read and image file from the same computer in any format, display it in a picture box, and then convert the image to bytes to save it in the DB. Until now, I can display the image in the picture box, but I can't convert the image to bytes. Here's my code:
private void DisplayImage()
{
if (openFileDialog.ShowDialog(this) == DialogResult.OK)
{
try
{
Stream file;
if ((archivo = openFileDialog.OpenFile()) != null)
{
using (file)
{
pictureBox.Image = Image.FromStream(file);
}
}
}
catch (Exception ex)
{
...
}
}
}
That's a simple method that just displays the image in the picture box. The real problem is with the following method:
public static byte[] ConvertImageToBytes(Image image)
{
if (image != null)
{
MemoryStream ms = new MemoryStream();
using (ms)
{
image.Save(ms, ImageFormat.Bmp);
byte[] bytes = ms.ToArray();
return bytes;
}
}
else
{
return null;
}
}
When it tries to save the image to the memory stream, I get the error:
System.Runtime.InteropServices.ExternalException: A generic error occurred in GDI+.
Any ideas on what's happening?
You should use the RawFormat property of the original image as a parameter to the Save method, not default to a Bitmap. This will avoid image format type errors. eg:
image.Save(ms, image.RawFormat);
ms.Position = 0;
byte [] bytes=ms.ToArray();
I'd advise actually saving images to the file-system and simply storing the file path (preferably relative) in the database.
BLOBs (ie images etc) in a database cannot be indexed, are often stored in a secondary, slower access database area and will quickly blow out the size of the database (slower backups etc).
Cant you simply Read the file and load it to a byte[] using the File class:
byte[] imgData = System.IO.File.ReadAllBytes(#"C:\My Pic\Myfile.jpg");
You can pick the image path from your Open Dialog box.
That particular exception generally means that you are trying to save the image as the wrong format. In your code you specify ImageFormat.Bmp - is it actually a bitmap image, or did you perhaps load it from a JPEG or PNG? Attempting to save as a different format from the one you loaded will fail with ExternalException, as specified in the documentation.
Incidentally, I don't recommend storing images in a database and I believe most people here will agree. Databases may be able to handle this task but they are not optimized for it, and you end up hurting the performance of both your database and your application. Unless you are using SQL Server 2008 FILESTREAM columns, it is more efficient to store images on the file system.
It may be stupid to answer my own question, but I just found out that if I want to convert the Image object to bytes, I have to leave the original stream open. I saw this issue in another page I can't quite remember, and I tested it by leaving the stream open and it's true. So the format wasn't the problem. But I will take the advice of all of you and store the images in a separate directory. Thanks for your help guys!
The problem with this is that stream must be open during the lifetime of of the image otherwise will fail.
One solution that worked for me is just to create a copy of the image like this:
using (var ms = new MemoryStream(bytes))
{
_image = new Bitmap(Image.FromStream(ms));
}
this is how my code look now:
System.Drawing.Image objImage = System.Drawing.Image.FromFile(Server.MapPath("aaa.jpg"));
int height = objImage.Height;
int width = objImage.Width;
System.Drawing.Bitmap bitmapimage = new System.Drawing.Bitmap(objImage, width, height);
System.Drawing.Graphics g = System.Drawing.Graphics.FromImage(bitmapimage);
System.Drawing.Image bitmap2 = (System.Drawing.Image)Bitmap.FromFile(Server.MapPath("sem.png"));
g.DrawImage(bitmap2, (objImage.Width - bitmap2.Width) / 2, (objImage.Height - bitmap2.Height) / 2);
MemoryStream stream = new MemoryStream();
bitmapimage.Save(stream, ImageFormat.Jpeg);
String saveImagePath = Server.MapPath("ImagesMerge/") + "aaa.jpg";
bitmapimage.Save(saveImagePath);
imgBig.ImageUrl = saveImagePath;
The problem I have now is that the image is not displayed in browser, I don't understand why .
like jmaglasang said, I would suggest to you to use an ashx file and if you don't need to keep the image, just send the image stream directly to the http without saving it on the disk
so you only need to do something like
<img src="Handler.ashx?action=merge&image1=blah.jpg&image2=bloh.jpg">
look at this code for an example of how to send an image made in memory that does not exist on the drive
Bitmap is a subclass of Image, so there no need to convert Bitmap to Image. It already is...
Probably because saveImagePath will be a local path (such as c:\somepath\aaa.jpg) that is not reachable from the browser. You probably want to set the ImageUrl = "ImagesMerge/aaa.jpg" instead.
You can also try:
imgBig.ImageUrl = ResolveUrl(saveImagePath);
EDIT:
If saveImagePath is under the WebApplication Directory, doing some modifications on the directory structure i.e. modifying files, deleting and creating can cause the application pool to recycle, and once it reaches the maximum recycle count the application pool will be stopped causing "Server unavailable" error.
I would suggests to add/save/modify images on a separate directory (not under the Apps Directory) then create a Handler(ASHX) that will read the images, just an advice though.
MapPath will give you a physycal address, not a virtual address which is what the browser needs to get to the image.
You might be forgetting to set the Response.Headers. Check out the following example that shows how to create bar chart images and then display it on the screen:
http://www.highoncoding.com/Articles/399_Creating_Bar_Chart_Using__NET_Graphics_API.aspx
I was loading a Bitmap Image from a File. When I tried to save the Image to another file I got the following error "A generic error occurred in GDI+". I believe this is because the file is locked by the image object.
Ok so tried calling the Image.Clone function. This still locks the file.
hmm. Next I try loading a Bitmap Image from a FileStream and load the image into memory so GDI+ doesn't lock the file. This works great except I need to generate thumbnails using Image.GetThumbnailImage method it throws an out of memory exception. Apparently I need to keep the stream open to stop this exception but if I keep the stream open then the file remains locked.
So no good with that method. In the end I created a copy of the file. So now I have 2 versions of the file. 1 I can lock and manipulate in my c# program. This other original file remains unlocked to which I can save modifications to. This has the bonus of allowing me to revert changes even after saving them because I'm manipulating the copy of the file which cant change.
Surely there is a better way of achieving this without having to have 2 versions of the image file. Any ideas?
Well if you're looking for other ways to do what you're asking, I reckon it should work to create a MemoryStream, and read out the FileStream to it, and load the Image from that stream...
var stream = new FileStream("original-image", FileMode.Open);
var bufr = new byte[stream.Length];
stream.Read(bufr, 0, (int)stream.Length);
stream.Dispose();
var memstream = new MemoryStream(bufr);
var image = Image.FromStream(memstream);
Or something prettier to that extent.
Whether or not that's the way you should go about solving that problem, I don't know. :)
I've had a similar problem and wound up fixing it like this.
I have since found an alternative method to clone the image without locking the file. Bob Powell has it all plus more GDI resources.
//open the file
Image i = Image.FromFile(path);
//create temporary
Image t=new Bitmap(i.Width,i.Height);
//get graphics
Graphics g=Graphics.FromImage(t);
//copy original
g.DrawImage(i,0,0);
//close original
i.Dispose();
//Can now save
t.Save(path)
I had a similar problem. But I knew, that I will save the image as a bitmap-file. So I did this:
public void SaveHeightmap(string path)
{
if (File.Exists(path))
{
Bitmap bitmap = new Bitmap(image); //create bitmap from image
image.Dispose(); //delete image, so the file
bitmap.Save(path); //save bitmap
image = (Image) bitmap; //recreate image from bitmap
}
else
//...
}
Sure, thats not the best way, but its working :-)