如何从 python 中的 tarfile 将文件提取到不同的目标文件名？

Question

I have a tarfile.TarFile from which I would like to extract some files to a modified destination filename; there is an existing file with the same name as the archive member that I do not want to touch. Specifically, I want to append a suffix, eg a member in the archive called foo/bar.txt should be extracted as foo/bar.txt.mysuffix .

The two somewhat obvious but also somewhat unsatisfactory approaches are:

• extract each file using extractfile , create renamed file and copy content using shutil.copyfileobj ; however, this is either limited to regular files or all the special handling, eg for sparse files, symlinks, directories etc. implemented in tarfile would have to be replicated.
• extractall to a temporary directory and then rename and copy to destination; this just feels unnecessarily convoluted, requires more interaction with the host system and introduces new failure modes, and it seems easy to get this subtly wrong (eg see warnings on shutil.copy/copy2 ).

Is there no interface or hook on the TarFile that would allow to implement this concisely and correctly?

Answer 1

There is TarFile.getmembers() method which returns the members of the archive as a list. And there you can loop and choose what files you want to extract or not.Depending on the size of your tar,your second approach can be viable too but not the best.

object = tarfile.open('example.tar', 'r')
for member in object.getmembers():
    if "whatever" in member.name:
        object.extract(member, "example_dir")

Answer 2

Looking throughLib/tarfile.py , I came across this comment :

    #--------------------------------------------------------------------------
    # Below are the different file methods. They are called via
    # _extract_member() when extract() is called. They can be replaced in a
    # subclass to implement other functionality.

    def makedir(self, tarinfo, targetpath):
       #...
    
    def makefile(self, tarinfo, targetpath):
       # ...

These methods are not mentioned in the official reference documentation, but they appear to be fair game. To overwrite these on an existing open TarFile instance, we can create a subclass Facade/Wrapper:

class SuffixingTarFile(tarfile.TarFile):
    def __init__(self, suffix: str, wrapped: tarfile.TarFile):
        self.suffix = suffix
        self.wrapped = wrapped

    def __getattr__(self, attr):
        return getattr(self.wrapped, attr)

    def makefile(self, tarinfo, targetpath):
        super().makefile(tarinfo, targetpath + self.suffix)

    # overwrite makedir, makelink, makefifo, etc. as desired

Example:

tar = tarfile.open(...)
star = SuffixingTarFile(".foo", tar)
star.extractall()  # extracts all (regular) file members with .foo suffix appended