Convert Python Bytes to String Without Encoding

Question

I am using Python 3.6 and I have an image as bytes:

img = b'\xff\xd8\xff\xe0\x00\x10JFIF\x00'

I need to convert the bytes into a string without encoding so it looks like:

raw_img = '\xff\xd8\xff\xe0\x00\x10JFIF\x00'

The goal is to incorporate this into an html image tag:

<img src="'data:image/png;base64," + base64.b64encode(raw_img) + "' />"

score 3 · Answer 1 · answered Apr 05 '18 at 17:06

3

Why not just call str and remove the b after?

In:

str(img)[2:-1]

Out:

'\xff\xd8\xff\xe0\x00\x10JFIF\x00'

answered Apr 05 '18 at 17:06

Primusa

13,136
3
33
53

1

I'm also seeing that backslashes are being converted to double backslashes in the process – slaw Apr 05 '18 at 17:11
2

@slaw No they aren't. – Stop harming Monica Apr 05 '18 at 17:12
That's the escape character... In reality there is only one backslash. Because of stuff like \n being newline, another backslash is added to escape it. The raw string representation is still just one backslash. – Primusa Apr 05 '18 at 17:12
1

repr(img) is the answer to this question. – eatmeimadanish Apr 05 '18 at 19:13

eatmeimadanish · Answer 2 · 2018-04-05T17:34:34.767

2

img.decode("utf-8")

You can decode the variable with the above. Then convert it to base64.

"<img src='data:image/png;base64,{}'/>".format( base64.b64encode(img.decode("utf-8")) )

UPDATED:

What you really want is this:

raw_img = repr(img)
"<img src='data:image/png;base64,{}'/>".format( base64.b64encode(raw_img) )

edited Apr 05 '18 at 17:34

answered Apr 05 '18 at 17:05

eatmeimadanish

3,809
1
14
20

2

`UnicodeDecodeError: 'utf-8' codec can't decode byte 0xff in position 0: invalid start byte` – user3483203 Apr 05 '18 at 17:06
why someone downvoted me when I provided the correct answer is unfortunate. – eatmeimadanish Apr 05 '18 at 19:14
Not the down-voter but I think it might because you used the wrong encoding. – Xantium Apr 05 '18 at 22:05
It was downvoted within seconds of my post. And I have since updated it with the resolution that address his expressed issue. – eatmeimadanish Apr 06 '18 at 14:30
Yours was too, but I upvoted it because I thought it was unfair. – eatmeimadanish Apr 06 '18 at 14:30
I get this: a bytes-like object is required, not 'str' – Melissa Guo May 02 '19 at 14:14
`img.encode("utf-8")` Or something similar to address that issue. – eatmeimadanish May 02 '19 at 15:16

Xantium · Answer 3 · 2018-04-05T18:02:08.893

Since you just need to convert the image to string why not just use str() function?

>>> img = b'\xff\xd8\xff\xe0\x00\x10JFIF\x00'
>>> type(img)
<class 'bytes'>
>>>
>>>raw_img = str(img)
>>> type(str(img)) 
<class 'str'>
>>>

img is in bytes, but when you use str() it is converted to type string.

An encoding can also be specified https://docs.python.org/3/library/stdtypes.html#str, which would be a more natural way to do things:

str(img, encoding='ansi')

As suggested in these answers

Pablo LION · Answer 4 · 2022-02-03T06:42:13.093

I didn't solve this but here's some research on it(3Feb2022): This encoding is latin (or latin-1) and it's hard to print because Python wants to print it in another format. But for your case they should be the same. And for a data:image/png;base64 base64 code should be used.

My test code:

import codecs

img = b"\xff\xd8\xff\xe0\x00\x10JFIF\x00"
desired = "\xff\xd8\xff\xe0\x00\x10JFIF\x00"
str_decode = img.decode("latin-1")
str_decode_2 = str(img, "latin-1")
codecs_decode = codecs.decode(img, "latin-1")
print(desired.encode("latin-1") == img)
print(str_decode == desired)
print(str_decode == str_decode_2)
print(str_decode == codecs_decode)
print("desired:", repr(desired))  ##devprint

This gives 4 True and a desired: ÿØÿà\x00\x10JFIF\x00 with Python 3.10.

score 1 · Answer 5 · answered Feb 11 '22 at 01:26

I've solved it (2022 - bit late to the party...) If you try img_raw.decode() you get the UnicodeDecodeError: 'utf-8' codec can't decode byte 0xff in position 0: invalid start byte error

But if you leave img_raw as a binary string and pass it into b64encode and then decode it, it doesn't have the UnicodeDecodeError, and you can pass it in as a data string to your image tag.

base64.b64encode(raw_image).decode()

chepner · Answer 6 · 2018-04-05T18:20:44.120

0

I'm pretty sure img is the byte string that you want to pass to base64.b64encode:

>>> import base64
>>> img = b'\xff\xd8\xff\xe0\x00\x10JFIF\x00'
>>> base64.b64encode(img)
b'/9j/4AAQSkZJRgA='

If you want to incorporate that into an HTML string, use

html = b'<img src="data:image/png;base64,' + base64.b64encode(img) + b' />'

edited Apr 05 '18 at 18:20

answered Apr 05 '18 at 17:12

chepner

497,756
71
530
681

Not quite. It still adds `b'` to the beginning of the string and `'` to the end but the image still doesn't show up – slaw Apr 05 '18 at 17:57
No, it doesn't; the `b` is just part of the *representation* of the value, not the value itself. Look at the result of `list(img)`. The value of `img` you posted doesn't seem to contain a complete image; it looks like just an image header. – chepner Apr 05 '18 at 18:01
This `"{}".format(base64.b64encode(img))` produces `"b'/9j/4AAQSkZJRgA='"`. Notice that the double quotes are around the `b'` which implies that it is being incorporated into the string. – slaw Apr 05 '18 at 18:09
OK, see pending update. I was apparently testing with a weird mix of Python 2 and Python 3. – chepner Apr 05 '18 at 18:20
(I forgot to keep `bytes` and `str` objects distinct, and `bytes.format` isn't defined in Python 3.) – chepner Apr 05 '18 at 18:26

Convert Python Bytes to String Without Encoding

6 Answers6

Linked