Once I can save this string as image, other time I can'

Question

Ok. Long story short.

My camera has a method which takes a photo and this is what it returns:

[160, 120, 3, 10, 1287848024, 96181, 'super long image string']

I am able to decode the string and save it as image right after I call the method like this:

for i in range(0, 10):
  image = camProxy.getImageRemote(nameId)
  imageWidth = image[0]
  imageHeight = image[1]
  imageByteArray = image[6]
  im = Image.fromstring("YCbCr",(imageWidth,imageHeight),imageByteArray)
  fileName = str(time.time())+".jpg"
  im.save(fileName, "JPEG")

This works nicely and I can open the saved images.

However, if I just save the string into a txt file and later I want to load it and save as image like this:

f = open("rawImage.txt", "r")
data = f.readline()
f.close()

# save as image
im = Image.frombuffer("YCbCr",(160,120),data)
im.save("test.jpg", "JPEG")

What I get is almost completely green image.

Here is an example string which I keep having problems with:

http://richardknop.com/rawImage.txt

Here is a complete output of the getImageRemote() method of the camera for that image:

http://richardknop.com/log.txt

Anybody got ideas what could be wrong? Is this some issue related to encoding? All files are saved as ASCII but I have tried saving them all as UTF-8 as well.

EDIT:

How I have written the image to file? I just redirected the output of the script:

python script.py > output.txt

And in the script I had:

print imageByteArray

Answer 1

I got it working by changing the file mode from "r" to "rb".

Here's the working code:

import time
import Image
image_data = [160, 120, 3, 10, 1287848024, 96181, 'really long string from http://richardknop.com/log.txt']
imageWidth = image_data[0]
imageHeight = image_data[1]
imageByteArray = image_data[6]
fout = open("image_data.txt", "wb")
fout.write(imageByteArray)
fout.close()
fin = open("image_data.txt", 'rb')
image_string = fin.read()
fin.close()
im = Image.fromstring("YCbCr",(imageWidth,imageHeight),image_string)
fileName = str(time.time())+".jpg"
im.save(fileName, "JPEG")

I verified that you are correct, in that read and readline make no difference here, but I still advise using read, because that says what you mean.

Here's my original answer:

Change data = f.readline() to data = f.read(). read grabs the whole file, readline grabs just one line.

Answer 2

Maybe you should read and write into your file using binary mode like this :

open('file_name', 'wb')
open('file_name', 'rb')

Answer 3

Read in the data:

import Image
import ast

with open('rawImage.txt','r') as f:
   raw_data=f.read()
with open('log.txt','r') as f:
   log_data=f.read()   
log_data=ast.literal_eval(log_data)
imageWidth=log_data[0]
imageHeight=log_data[1]
log_data=log_data[6]        

Let's try to see if raw_data (from rawImage.txt) is the same string as log_data (from log.txt). Oops: they're not the same length:

print(len(raw_data))
# 146843
print(len(log_data))
# 57600

Take a peek at the beginning of both strings. It appears raw_data has written 4 characters for '\x81' when a single character \x81 was intended.

print(list(raw_data[:10]))
# ['6', '}', '\\', 'x', '8', '1', '8', '}', '\\', 'x']
print(list(log_data[:10]))
# ['6', '}', '\x81', '8', '}', '\x81', '7', '\x90', '\x8a', '4']

This might have happened because rawImage.txt was opened in writing mode 'w' instead of 'wb'. The best solution is to write rawImage.txt using the right writing mode, as Steven Rumbalski does here.

But given this predicament, here is a way you can fix it:

raw_data_fixed=raw_data.decode('string_escape') 

Now this works:

im = Image.fromstring("YCbCr",(imageWidth,imageHeight),raw_data_fixed)
im.show()