X-Git-Url: https://git.mdrn.pl/librarian.git/blobdiff_plain/e0e7693096469f5952eef1c5847dabc9cf797f70..6a2a1ba8ab5cad50af8dc0b1ec0387e4e4f41f4b:/librarian/xmlutils.py diff --git a/librarian/xmlutils.py b/librarian/xmlutils.py index a3f9df9..ae3512a 100644 --- a/librarian/xmlutils.py +++ b/librarian/xmlutils.py @@ -8,14 +8,38 @@ from collections import defaultdict class Xmill(object): - """Transforms XML to some text. + """Transforms XML to some text. Used instead of XSLT which is difficult and cumbersome. - + """ - def __init__(self, options=None): + def __init__(self, options=None, state=None): self._options = [] + self.state = state or {} if options: self._options.append(options) + self.text_filters = [] + self.escaped_text_filters = [] + + def register_text_filter(self, fun): + self.text_filters.append(fun) + + def register_escaped_text_filter(self, fun): + self.escaped_text_filters.append(fun) + + def filter_text(self, text): + for flt in self.text_filters: + if text is None: + return None + else: + text = flt(text) + # TODO: just work on the tree and let lxml handle escaping. + e = etree.Element("x") + e.text = text + # This whole mixing text with ML is so wrong. + output = etree.tostring(e, encoding=unicode)[3:-4] + for flt in self.escaped_text_filters: + output = flt(output) + return output def generate(self, document): """Generate text from node using handlers defined in class.""" @@ -26,9 +50,9 @@ class Xmill(object): def options(self): """Returnes merged scoped options for current node. """ - # Here we can see how a decision not to return the modified map + # Here we can see how a decision not to return the modified map # leads to a need for a hack. - return reduce(lambda a, b: a.update(b) or a, self._options, defaultdict(lambda: False)) + return reduce(lambda a, b: a.update(b) or a, self._options, defaultdict(lambda: None)) @options.setter def options(self, opts): @@ -36,26 +60,23 @@ class Xmill(object): """ self._options.append(opts) - def _handle_for_element(self, element): ns = None tagname = None -# from nose.tools import set_trace - - if isinstance(element, etree._Comment): return None + # from nose.tools import set_trace if element.tag[0] == '{': for nshort, nhref in element.nsmap.items(): try: if element.tag.index('{%s}' % nhref) == 0: ns = nshort - tagname = element.tag[len('{%s}' % nhref):] + tagname = element.tag[len('{%s}' % nhref):] break except ValueError: pass if not ns: - raise ValueError("Strange ns for tag: %s, nsmap: %s" % - (element.tag, element.nsmap)) + raise ValueError("Strange ns for tag: %s, nsmap: %s" % + (element.tag, element.nsmap)) else: tagname = element.tag @@ -63,7 +84,7 @@ class Xmill(object): meth_name = "handle_%s__%s" % (ns, tagname) else: meth_name = "handle_%s" % (tagname,) - + handler = getattr(self, meth_name, None) return handler @@ -73,49 +94,72 @@ class Xmill(object): while True: sibling = element.getnext() - if sibling is not None: return sibling # found a new branch to dig into + if sibling is not None: + return sibling # found a new branch to dig into element = element.getparent() - if element is None: return None # end of tree + if element is None: + return None # end of tree def _handle_element(self, element): + if isinstance(element, etree._Comment): + return None + handler = self._handle_for_element(element) + if self.state.get('mute') and not getattr(handler, 'unmuter', False): + return None # How many scopes + options_scopes = len(self._options) try: - options_scopes = len(self._options) - if handler is None: - pre = [element.text] - post = [] + pre = [self.filter_text(element.text)] + post = [self.filter_text(element.tail)] else: vals = handler(element) # depending on number of returned values, vals can be None, a value, or a tuple. # how poorly designed is that? 9 lines below are needed just to unpack this. if vals is None: - return [] + return [self.filter_text(element.tail)] else: if not isinstance(vals, tuple): - return [vals, element.tail] + return [vals, self.filter_text(element.tail)] else: - pre = [vals[0], element.text] - post = [vals[1], element.tail] + pre = [vals[0], self.filter_text(element.text)] + post = [vals[1], self.filter_text(element.tail)] out = pre + [self._handle_element(child) for child in element] + post finally: # clean up option scopes if necessary self._options = self._options[0:options_scopes] + return out -def tag(name, classes=None, **attrs): +def tag_open_close(name_, classes_=None, **attrs): + u"""Creates tag beginning and end. + + >>> tag_open_close("a", "klass", x=u"ą<") + (u'', u'') + + """ + if classes_: + if isinstance(classes_, (tuple, list)): + classes_ = ' '.join(classes_) + attrs['class'] = classes_ + + e = etree.Element(name_) + e.text = " " + for k, v in attrs.items(): + e.attrib[k] = v + pre, post = etree.tostring(e, encoding=unicode).split(u"> <") + return pre + u">", u"<" + post + + +def tag(name_, classes_=None, **attrs): """Returns a handler which wraps node contents in tag `name', with class attribute set to `classes' and other attributes according to keyword paramters """ - if classes: - if isinstance(classes, (tuple, list)): classes = ' '.join(classes) - attrs['class'] = classes - a = ''.join([' %s="%s"' % (k,v) for (k,v) in attrs.items()]) def _hnd(self, element): - return "<%s%s>" % (name, a), "" % name + return tag_open_close(name_, classes_, **attrs) return _hnd @@ -124,13 +168,16 @@ def tagged(name, classes=None, **attrs): set to `classes' and other attributes according to keyword paramters """ if classes: - if isinstance(classes, (tuple,list)): classes = ' '.join(classes) + if isinstance(classes, (tuple, list)): + classes = ' '.join(classes) attrs['class'] = classes - a = ''.join([' %s="%s"' % (k,v) for (k,v) in attrs.items()]) + a = ''.join([' %s="%s"' % (k, v) for (k, v) in attrs.items()]) + def _decor(f): def _wrap(self, element): r = f(self, element) - if r is None: return + if r is None: + return prepend = "<%s%s>" % (name, a) append = "" % name @@ -155,6 +202,7 @@ def ifoption(**options): return _handler return _decor + def flatten(l, ltypes=(list, tuple)): """flatten function from BasicPropery/BasicTypes package """