Attribute error using python scientific module on Windows

I get the following error when our program is run on Windows XP using ActivePerl 2.6 (although it runs fine on linux). Any ideas of what is going on? Is this a bug in the module or a problem with our code?

Error log:

      File "C:\Python26\lib\site-packages\Scientific\IO\PDB.py", line 1163, in __ini
    t__
        self.parseFile(PDBFile(file_or_filename))
      File "C:\Python26\lib\site-packages\Scientific\IO\PDB.py", line 1371, in parse
    File
        type, data = file.readLine()
     File "C:\Python26\lib\site-packages\Scientific\IO\PDB.py", line 200, in readLi
   ne
        line = self.file.readline()
    AttributeError: 'unicode' object has no attribute 'readline'
    Exception AttributeError: "'unicode' object has n开发者_StackOverflow社区o attribute 'close'" in <bound
    method PDBFile.__del__ of <Scientific.IO.PDB.PDBFile instance at 0x00FD0440>> ig
    nored

UPDATE: Following Joe's suggestion below, I get the following error instead:

Traceback (most recent call last):
  File "NewParseV1_00.py", line 47, in <module>
    PromptUser()
  File "NewParseV1_00.py", line 45, in PromptUser
    parseGroupFile(grpfilepath, outputPPYfilepath, sorcePDBfilepath)
  File "NewParseV1_00.py", line 39, in parseGroupFile
    return GeneratePPY(LL,GRPNAME,SRCPDB,OUTPPY)
  File "NewParseV1_00.py", line 10, in __init__
    self.PDBStruct = Scientific.IO.PDB.Structure(SRCPDB)
  File "C:\Python26\lib\site-packages\Scientific\IO\PDB.py", line 1163, in __ini
t__
    self.parseFile(PDBFile(file_or_filename))
  File "C:\Python26\lib\site-packages\Scientific\IO\PDB.py", line 161, in __init
__
    if isinstance(file_or_filename, basestr):
NameError: global name 'basestr' is not defined
Exception AttributeError: "PDBFile instance has no attribute 'open'" in <bound m
ethod PDBFile.__del__ of <Scientific.IO.PDB.PDBFile instance at 0x00FDB418>> ign
ored

I'm guessing, but it looks like a bug in Scientific.IO.PDB where they do something like:

if type(file_or_filename) == str:
    file = open(file_or_filename)
else:
    file = file_or_filename

rather than doing:

if isinstance(file_or_filename, basestr):
    ...

For whatever reason, you appear to be passing in a unicode filename on the windows side, and a raw string on the linux side.

Again, though, I'm just taking a guess. I'm assuming Scientific here is this module? http://dirac.cnrs-orleans.fr/plone/software/scientificpython/ (I'm not familiar with it...)

Edit: Yep!

Here's the relevant code from Scientific.IO.PDB:

class PDBFile:

    """
    X{PDB} file with access at the record level

    The low-level file access is handled by the module
    L{Scientific.IO.TextFile}, therefore compressed files and URLs
    (for reading) can be used as well.
    """

    def __init__(self, file_or_filename, mode = 'r', subformat = None):
        """
        @param file_or_filename: the name of the PDB file, or a file object
        @type file_or_filename: C{str} or C{file}
        @param mode: the file access mode, 'r' (read) or 'w' (write)
        @type mode: C{str}
        @param subformat: indicates a specific dialect of the PDB format.
                          Subformats are defined in
                          L{Scientific.IO.PDBExportFilters}; they are used
                          only when writing.
        @type subformat: C{str} or C{NoneType}
        """
        if isinstance(file_or_filename, str):
            self.file = TextFile(file_or_filename, mode)
        else:
            self.file = file_or_filename
        <snip>

It should be as simple as changing isinstance(file_or_filename, str) to isinstance(file_or_filename, basestr). You might want to file a bug report...

Changing the code in Scientific.IO.PDB as explained by Joe needs one final modification to work:

On line 161, change if isinstance(file_or_filename, str): to if isinstance(file_or_filename, basestring):

(Note that Joe wrote that the second argument should be basestr but that gives the error in the update in the original question.)