diff -r d09cc8b3709c -r e94ab812c0c8 lib/urltree.py --- /dev/null Thu Jan 01 00:00:00 1970 +0000 +++ b/lib/urltree.py Sun Feb 08 03:13:11 2009 +0200 @@ -0,0 +1,670 @@ +""" + Tree-based URL mapping +""" + +import re +import os.path + +# for Mapper +from lib import map + +class URLError (Exception) : + """ + Error with an URL definition + """ + + pass + +class LabelValue (object) : + """ + Represents the value of a ValueLabel... love these names + """ + + def __init__ (self, label, value) : + """ + Just store + """ + + self.label = label + self.value = value + + def __str__ (self) : + return "%s=%r" % (self.label.key, self.value) + + def __repr__ (self) : + return "<%s>" % self + +class Label (object) : + """ + Base class for URL labels (i.e. the segments of the URL between /s) + """ + + @staticmethod + def parse (mask, defaults, types) : + """ + Parse the given label-segment, and return a *Label instance + """ + + # empty? + if not mask : + return EmptyLabel() + + # simple value? + match = SimpleValueLabel.EXPR.match(mask) + + if match : + # key + key = match.group('key') + + # type + type = match.group("type") + + # lookup type, None for default + type = types[type] + + # defaults? + default = defaults.get(key) + + if not default : + default = match.group('default') + + if default : + # apply type to default + default = type(default) + + # build + return SimpleValueLabel(key, type, default) + + # static? + match = StaticLabel.EXPR.match(mask) + + if match : + return StaticLabel(match.group('name')) + + # invalid + raise URLError("Invalid label: %r" % (mask, )) + + def match (self, value=None) : + """ + Match this label against the given value, returning either True to match without a value, a LabelValue + object, or boolean false to not match. + + If value is None, this means that only a default value should be returned. + """ + + abstract + + def build (self, value_dict) : + """ + Return a string representing this label, using the values in the given value_dict if needed + """ + + abstract + +class EmptyLabel (Label) : + """ + An empty label, i.e. just a slash in the URL + """ + + def __eq__ (self, other) : + """ + Just compares type + """ + + return isinstance(other, EmptyLabel) + + def match (self, value=None) : + """ + Match empty string -> True + """ + + # no default + if value is None : + return False + + # only empty segments + if value == '' : + return True + + def build (self, values) : + return str(self) + + def __str__ (self) : + return '' + +class StaticLabel (Label) : + """ + A simple literal Label, used for fixed terms in the URL + """ + + EXPR = re.compile(r'^(?P[a-zA-Z_.-]+)$') + + def __init__ (self, name) : + """ + The given name is the literal name of this label + """ + + self.name = name + + def __eq__ (self, other) : + """ + Compares names + """ + + return isinstance(other, StaticLabel) and self.name == other.name + + def match (self, value=None) : + """ + Match exactly -> True + """ + + # no defaults + if value is None : + return False + + # match name + if value == self.name : + return True + + def build (self, values) : + return str(self) + + def __str__ (self) : + return self.name + +class ValueLabel (Label) : + """ + A label with a key and a value + + XXX: do we even need this? + """ + + def __init__ (self, key, default) : + """ + Set the key and default value. Default value may be None if there is no default value defined + """ + + self.key = key + self.default = default + + def __eq__ (self, other) : + """ + Compares keys + """ + + return isinstance(other, ValueLabel) and self.key == other.key + + def build (self, values) : + """ + Return either the assigned value from values, our default value, or raise an error + """ + + value = values.get(self.key) + + if not value and self.default : + value = self.default + + elif not value : + raise URLError("No value given for label %r" % (self.key, )) + + return value + +class SimpleValueLabel (ValueLabel) : + """ + A label that has a name and a simple string value + """ + + EXPR = re.compile(r'^\{(?P[a-zA-Z_][a-zA-Z0-9_]*)(:(?P[a-zA-Z_][a-zA-Z0-9_]*))?(=(?P[^}]+))?\}$') + + def __init__ (self, key, type=str, default=None) : + """ + The given key is the name of this label's value + """ + + # type + self.type = type + + # store + self.key = key + self.default = default + + def match (self, value=None) : + """ + Match -> LabelValue + """ + + # default? + if value is None and self.default : + return LabelValue(self, self.default) + + # only non-empty values! + elif value : + # convert with type + try : + value = self.type(value) + + except Exception, e : + raise URLError("Bad value %r for type %s: %s: %s" % (value, self.type.__name__, type(e).__name__, e)) + + return LabelValue(self, value) + + def __str__ (self) : + return '{%s%s%s}' % ( + self.key, + ':%s' % (self.type.__name__ ) if self.type != str else '', + '=%s' % (self.default, ) if self.default else '', + ) + +class URLConfig (object) : + """ + Global configuration relevant to all URLs + """ + + # built-in type codes + BUILTIN_TYPES = { + # default + None : str, + + # string + 'str' : str, + + # integer + 'int' : int, + } + + def __init__ (self, type_dict=None) : + """ + Create an URLConfig for use with URL + + If type_dict is given, it should be a mapping of type names -> callables, and they will be available for + type specifications in addition to the defaults. + """ + + # build our type_dict + self.type_dict = self.BUILTIN_TYPES.copy() + + # apply the given type_dict + if type_dict : + self.type_dict.update(type_dict) + +class URL (object) : + """ + Represents a specific URL + """ + + + def __init__ (self, config, url_mask, handler, type_dict=None, **defaults) : + """ + Create an URL using the given URLConfig, with the given url mask, handler, and default values. + """ + + # store + self.config = config + self.url_mask = url_mask + self.handler = handler + self.defaults = defaults + + # query string + self.query_args = dict() + + # parse any query string + # XXX: conflicts with regexp syntax + if '/?' in url_mask : + url_mask, query_mask = url_mask.split('/?') + + else : + query_mask = None + + # build our label path + self.label_path = [Label.parse(mask, defaults, config.type_dict) for mask in url_mask.split('/')] + + # build our query args list + if query_mask : + # split into items + for query_item in query_mask.split('&') : + # parse default + if '=' in query_item : + query_item, default = query_item.split('=') + + else : + default = None + + # parse type + if ':' in query_item : + query_item, type = query_item.split(':') + else : + type = None + + # parse key + key = query_item + + # type + type = self.config.type_dict[type] + + # add to query_args as (type, default) tuple + self.query_args[key] = (type, type(default) if default else default) + + def get_label_path (self) : + """ + Returns a list containing the labels in this url + """ + + # copy self.label_path + return list(self.label_path) + + def execute (self, request, label_values) : + """ + Invoke the handler, using the given label values + """ + + # start with the defaults + kwargs = self.defaults.copy() + + # then add all the values + for label_value in label_values : + kwargs[label_value.label.key] = label_value.value + + # then parse all query args + for key, value in request.get_args() : + # lookup spec + type, default = self.query_args[key] + + # normalize empty value to None + if not value : + value = None + + else : + # process value + value = type(value) + + # set default? + if not value : + if default : + value = default + + if default == '' : + # do not pass key at all + continue + + # otherwise, fail + raise URLError("No value given for required argument: %r" % (key, )) + + # set key + kwargs[key] = value + + # then check all query args + for key, (type, default) in self.query_args.iteritems() : + # skip those already present + if key in kwargs : + continue + + # apply default? + if default is None : + raise URLError("Missing required argument: %r" % (key, )) + + elif default == '' : + # skip empty default + continue + + else : + # set default + kwargs[key] = default + + # execute the handler + return self.handler(request, **kwargs) + + def build (self, request, **values) : + """ + Build an absolute URL pointing to this target, with the given values + """ + + # build URL from request page prefix and our labels + return request.page_prefix + '/'.join(label.build(values) for label in self.label_path) + + def __str__ (self) : + return '/'.join(str(label) for label in self.label_path) + + def __repr__ (self) : + return "URL(%r, %r)" % (str(self), self.handler) + +class URLNode (object) : + """ + Represents a node in the URLTree + """ + + def __init__ (self, parent, label) : + """ + Initialize with the given parent and label, empty children dict + """ + + # the parent URLNode + self.parent = parent + + # this node's Label + self.label = label + + # list of child URLNodes + self.children = [] + + # this node's URL, set by add_url for an empty label_path + self.url = None + + def _build_child (self, label) : + """ + Build, insert and return a new child Node + """ + + # build new child + child = URLNode(self, label) + + # add to children + self.children.append(child) + + # return + return child + + def add_url (self, url, label_path) : + """ + Add a URL object to this node under the given path. Uses recursion to process the path. + + The label_path argument is a (partial) label path as returned by URL.get_label_path. + + If label_path is empty (len zero, or begins with EmptyLabel), then the given url is assigned to this node, if no + url was assigned before. + """ + + # matches this node? + if not label_path or isinstance(label_path[0], EmptyLabel) : + if self.url : + raise URLError(url, "node already defined") + + else : + # set + self.url = url + + else : + # pop child label from label_path + child_label = label_path.pop(0) + + # look for the child to recurse into + child = None + + # look for an existing child with that label + for child in self.children : + if child.label == child_label : + # found, use this + break + + else : + # build a new child + child = self._build_child(child_label) + + # recurse to handle the rest of the label_path + child.add_url(url, label_path) + + def match (self, label_path) : + """ + Locate the URL object corresponding to the given label_path value under this node. + + Returns a (url, label_values) tuple + """ + + # determine value to use + value = None + + # empty label_path? + if not label_path or label_path[0] == '' : + # the search ends at this node + if self.url : + # this URL is the best match + return (self.url, []) + + elif not self.children : + # incomplete URL + raise URLError("no URL handler defined for this Node") + + else : + # use default value, i.e. Label.match(None) + label = None + + else : + # pop the next label from the label path + label = label_path.pop(0) + + # return one match... + match = value = None + + # recurse through our children, DFS + for child in self.children : + # match value + value = child.label.match(label) + + # skip those that don't match at all + if not value : + continue; + + # already found a match? :/ + if match : + raise URLError("Ambiguous URL") + + # ok, but continue looking to make sure there's no ambiguous URLs + match = child + + # found something? + if not match : + raise URLError("No child found for label: %s + %s + %s" % (self.get_url(), label, '/'.join(str(l) for l in label_path))) + + # ok, recurse into the match + url, label_value = match.match(label_path) + + # add our value? + if isinstance(value, LabelValue) : + label_value.append(value) + + # return the match + return url, label_value + + def get_url (self) : + """ + Returns the URL for this node, by iterating over our parents + """ + + # URL segments in reverse order + segments = [''] + + # start with ourself + node = self + + # iterate up to root + while node : + segments.append(str(node.label)) + + node = node.parent + + # reverse + segments.reverse() + + # return + return '/'.join(segments) + + def dump (self, indent=0) : + """ + Returns a multi-line string representation of this Node + """ + + return '\n'.join([ + "%-45s%s" % ( + ' '*indent + str(self.label) + ('/' if self.children else ''), + (' -> %r' % self.url) if self.url else '' + ) + ] + [ + child.dump(indent + 4) for child in self.children + ]) + + def __str__ (self) : + return "%s/[%s]" % (self.label, ','.join(str(child) for child in self.children)) + +class URLTree (map.Mapper) : + """ + Map requests to handlers, using a defined tree of URLs + """ + + def __init__ (self, url_list) : + """ + Initialize the tree using the given list of URLs + """ + + # root node + self.root = URLNode(None, EmptyLabel()) + + # just add each URL + for url in url_list : + self.add_url(url) + + def add_url (self, url) : + """ + Adds the given URL to the tree. The URL must begin with a root slash. + """ + # get url's label path + path = url.get_label_path() + + # should begin with root + root_label = path.pop(0) + assert root_label == self.root.label, "URL must begin with root" + + # add to root + self.root.add_url(url, path) + + def match (self, url) : + """ + Find the URL object best corresponding to the given url, matching any ValueLabels. + + Returns an (URL, [LabelValue]) tuple. + """ + + # split it into labels + path = url.split('/') + + # empty URL is empty + if url : + # ensure that it doesn't start with a / + assert not self.root.label.match(path[0]), "URL must not begin with root" + + # just match starting at root + return self.root.match(path) + + def handle_request (self, request) : + """ + Looks up the request's URL, and invokes its handler + """ + + # get the requested URL + request_url = request.get_page_name() + + # find the URL+values to use + url, label_values = self.match(request_url) + + # let the URL handle it + return url.execute(request, label_values) + +