Python：字典到Web文件列表

Question

I have a dict representation of a filesystem, which looks like this.

{
    "en": {
        'file_name_1': [{meta representation of file_name_1}] or False,
        {'folder_name': {
            'index': [{meta repr. of the index}] or False
            'sub_file_1': [{meta repr. of file_name_1}] or False,
            }
        }
    }
    "fr": {
        <...>
    }
    'tree_root': path to the tree's parent node
}

Given the path to one leaf in this tree (file or folder), I would like to represent the tree leading to that leaf as HTML nested lists.

I do not believe my tree will ever become very deep, but for the sake of being fast, I am considering running only once through the path of the leaf.

Here is what I have now: ( self.tree is my dictionary, leaf_path my, hmm. leaf path. I tried and cleaned my code for simplicity)

rel_URI = [key for key in leaf_path.split(os.sep)]
stack = []

html = u'<ul>'
for key in rel_URI:
    stack.append(key)
    obj = reduce(dict.get, stack, self.tree)
    if isinstance(obj, dict): # a subdir
        html += u'<ul>'
        for _dir in obj.iterkeys() if isinstance(_dir, dict):
            if _dir == rel_URI[len(stack)]: # in the leaf_path
                html += u'<li class="folder">%s</li>' % self.getTitle(obj, _dir)
            else: # not in the leaf_path
                html += u'<li class="folder, far">%s</li>'% self.getTitle(obj,_dir)
        for _file in obj.iterkeys() if isinstance(_file, list):
            html += u'<li class="file, far">%s</li>'% obj[0]['title']
        html += u'</ul>'
    else: # the leaf file
        item = u'<li class="file">%s</li>' % obj[0]['title']
html += '</ul>'

I have trouble figuring out how to nest the next sub-tree into the iteration: the way it is I have entirely disconnected the next step from this one, and so I end up with a messed-up nested file list.

• Part of my trouble is that dicts are not ordered

How would you go about building the list?

Answer 1

So, following Hans's advice, I turned to OrderedDicts, from the collections module

from collections import OrderedDict

then build OrderedDicts explicitly instead of dicts. They are not prettyprintable, and so harder to debug, but they do the job right.

So, thanks Hans.

I also ended up using recursion after all to place the lists in the right place.