AttributeError: 'str' object has no attribute 'decode' about captionstransformer HOT 2 CLOSED

This is a python 2 vs 3 issue. You mention that you are using Python 3.4.1. This library is made for Python 2, but it is very easy to fix. (I hope toutpt will read this and post an update).

Python 3 always uses unicode for the str type, and the str type no longer has (or needs) a decode method. Reading a str from a file in python 3 will always give you unicode, which needs no further decoding. If you try to call str.decode() in python 3 you will get this error.

A quick-and-dirty solution to this is to change core.py in this library, so that decode only gets called if it exists (which is effectively the same as checking whether you are running python 2). You could also use a try block, to catch that specific exception.

The decode method is called in two places in core.py.

I'm providing some code compatible with both python2+3 that you can use:

...in the read() method of the Reader class.

    def read(self):
        self.rawcontent = self.fileobject.read()        
        if type(self.rawcontent) == str:
            if hasattr(self.rawcontent, "decode"):
                self.rawcontent = self.rawcontent.decode(self.encoding)            
        self.text_to_captions()
        return self.captions

...and in the set_text() method of the Caption class

    def set_text(self, value):
        if type(value) == str:
            if hasattr(str, "decode"):
                value = str.decode(self.encoding)
        elif type(value) != unicode:
            raise ValueError("text must be either encoded string or unicode")
        self._text = value

This isn't a completely bulletproof solution, but in the absence of malicious code, it should 'just work'. I am sure better solutions can be imagined, but in my case, I just want to get it working.

BTW: Switching the decode issue will make the captionstransformer library compatible with python 2 & 3.

from captionstransformer.

closed by PR #2

from captionstransformer.