-#!/usr/bin/python2.7
-# -*- coding: iso-8859-15 -*-
+#!/usr/bin/python3.4
# $Id$
# $HeadURL$
-"""This module contains general functions that help parsing the mediawiki markup.
-I looked for an already existing MediaWiki parser in Python but I didn't find anything
-that convinced me. However, here are the links:
+"""For parsing MediaWiki text, we rely on the package mwparserfromhell (https://github.com/earwig/mwparserfromhell).
+This module just contains a few additional useful functions.
+Other Python MediaWiki parsers:
* py-wikimarkup https://github.com/dcramer/py-wikimarkup
* mwlib http://code.pediapress.com/wiki/wiki
+* https://www.mediawiki.org/wiki/Alternative_parsers
"""
-import re
-import xml.etree.ElementTree
-import collections
-import formencode
-
class ParseError(RuntimeError):
"""Exception used by some of the functions"""
pass
-def find_template(wikitext, template_title):
- """Returns the tuple (start, end) of the first occurence of the template '{{template ...}} within wikitext'.
- (None, None) is returned if the template is not found.
- If you are sure that the wikitext contains the template, the template could be extracted like follows:
-
- >>> wikitext = u'This is a {{Color|red|red text}} template.'
- >>> start, end = find_template(wikitext, u'Color')
- >>> print wikitext[start:end]
- {{Color|red|red text}}
-
- or just:
-
- >>> print wikitext.__getslice__(*find_template(wikitext, u'Color'))
- {{Color|red|red text}}
-
- The search is done with regular expression. It gives wrong results when parsing a template
- containing the characters "}}"
-
- :param wikitext: The text (preferalbe unicode) that has the template in it.
- :param template_title: The page title of the template with or without namespace (but as in the wikitext).
- :return:
- (start, end) of the first occurence with start >= 0 and end > start.
- (None, None) if the template is not found.
- """
- match = re.search(u"\{\{" + template_title + "\s*(\|[^\}]*)?\}\}", wikitext, re.DOTALL)
- if match is None: return None, None
- return match.start(), match.end()
-
-
-class TemplateValidator(formencode.FancyValidator):
- def __init__(self, strip=True):
- """If strip is True, the title, and the parameter keys and values are stripped in to_python."""
- self.strip = (lambda s: s.strip()) if strip else (lambda s: s)
-
- def to_python(self, value, state=None):
- """Takes a template, like u'{{Color|red|text=Any text}}' and translates it to a Python tuple
- (title, anonym_params, named_params) where title is the template title,
- anonym_params is a list of anonymous parameters and named_params is a OrderedDict
- of named parameters. Whitespace of the parameters is stripped."""
- if not value.startswith(u'{{'):
- raise formencode.Invalid(u'Template does not start with "{{"', value, state)
- if not value.endswith(u'}}'):
- raise formencode.Invalid(u'Template does not end with "}}"', value, state)
- parts = value[2:-2].split(u'|')
-
- # template name
- title = self.strip(parts[0])
- if len(title) == 0:
- raise formencode.Invalid(u'Empty template tilte.', value, state)
- del parts[0]
-
- # anonymous parameters
- anonym_params = []
- while len(parts) > 0:
- equalsign_pos = parts[0].find(u'=')
- if equalsign_pos >= 0: break # named parameter
- anonym_params.append(self.strip(parts[0]))
- del parts[0]
-
- # named or numbered parameters
- named_params = collections.OrderedDict()
- while len(parts) > 0:
- equalsign_pos = parts[0].find(u'=')
- if equalsign_pos < 0:
- raise formencode.Invalid(u'Anonymous parameter after named parameter.', value, state)
- key, sep, value = parts[0].partition(u'=')
- key = self.strip(key)
- if len(key) == 0:
- raise formencode.Invalid(u'Empty key.', value, state)
- if named_params.has_key(key):
- raise formencode.Invalid(u'Duplicate key: "{0}"'.format(key), value, state)
- named_params[key] = self.strip(value)
- del parts[0]
-
- return title, anonym_params, named_params
-
-
-def split_template(template):
- """Deprecated legacy function.
-
- Takes a template, like u'{{Color|red|text=Any text}}' and translates it to a Python tuple
- (template_title, parameters) where parameters is a Python dictionary {u'1': u'red', u'text'=u'Any text'}.
- Anonymous parameters get integer keys (converted to unicode) starting with 1
- like in MediaWiki, named parameters are unicode strings.
- Whitespace is stripped.
- If an unexpected format is encountered, a ValueError is raised."""
- try:
- title, anonym_params, named_params = TemplateValidator().to_python(template)
- parameters = dict(named_params)
- for i in xrange(len(anonym_params)):
- parameters[unicode(i+1)] = anonym_params[i]
- except formencode.Invalid as e:
- raise ValueError(e[0])
- return title, parameters
-
-
-def create_template(template_title, anonym_params=[], named_param_keys=[], named_param_values=[], as_table=False, as_table_keylen=None):
- """Formats a MediaWiki template.
- :param template_title: Unicode string with the template name
- :param anonym_params: list with parameters without keys
- :param named_param_keys: list with keys of named parameters
- :param named_param_values: list with values of named parameters, corresponding to named_param_keys.
- :param as_table: formats the returned template in one row for each parameter
- :param as_table_keylen: length of the key field. None for "automatic".
- :return: unicode template"""
- pipe_char, equal_char, end_char = (u'\n| ', u' = ', u'\n}}') if as_table else (u'|', u'=', u'}}')
- parts = [u"{{" + template_title]
- parts += anonym_params
- if as_table and as_table_keylen is None:
- as_table_keylen = max([len(k) for k in named_param_keys])
- for i in xrange(len(named_param_keys)):
- key = named_param_keys[i]
- if as_table:
- key = key.ljust(as_table_keylen)
- parts.append((key + equal_char + named_param_values[i]).rstrip())
- else:
- parts.append(key + equal_char + named_param_values[i])
- return pipe_char.join(parts) + end_char
-
-
-def find_tag(wikitext, tagname, pos=0):
- """Returns position information of the first occurence of the tag '<tag ...>...</tag>'
- or '<tag ... />'.
- If you are sure that the wikitext contains the tag, the tag could be extracted like follows:
+def format_template_table(template, keylen=None):
+ """Reformat the given template to be tabular.
- >>> wikitext = u'This is a <tag>mytag</tag> tag.'
- >>> start, content, endtag, end = find_template(wikitext, u'tag')
- >>> print wikitext[start:end]
- <tag>mytag</tag>
+ >>> template
+ {{foo|bar|bazz=7}}
+ >>> format_template_table(template)
+ {{foo
+ | bar
+ | bazz = 7
+ }}
- :param wikitext: The text (preferalbe unicode) that has the template in it.
- :param tagname: Name of the tag, e.g. u'tag' for <tag>.
- :param pos: position within wikitext to start searching the tag.
- :return:
- (start, content, endtag, end). start is the position of '<' of the tag,
- content is the beginning of the content (after '>'), enttag is the
- beginning of the end tag ('</') and end is one position after the end tag.
- For single tags, (start, None, None, end) is returned.
- If the tag is not found (or only the start tag is present,
- (None, None, None, None) is returned.
+ :param keylen: length of the keys or None for automatic determination
"""
- # Find start tag
- regexp_starttag = re.compile(u"<{0}.*?(/?)>".format(tagname), re.DOTALL)
- match_starttag = regexp_starttag.search(wikitext, pos)
- if match_starttag is None:
- return None, None, None, None
-
- # does the tag have content?
- if len(match_starttag.group(1)) == 1: # group(1) is either '' or '/'.
- # single tag
- return match_starttag.start(), None, None, match_starttag.end()
-
- # tag with content
- regexp_endtag = re.compile(u'</{0}>'.format(tagname), re.DOTALL)
- match_endtag = regexp_endtag.search(wikitext, match_starttag.end())
- if match_endtag is None:
- # No closing tag - error in wikitext
- return None, None, None, None
- return match_starttag.start(), match_starttag.end(), match_endtag.start(), match_endtag.end()
-
-
-def parse_googlemap(wikitext):
- """Parses the (unicode) u'<googlemap ...>content</googlemap>' of the googlemap extension.
- If wikitext does not contain the <googlemap> tag or if the <googlemap> tag contains
- invalid formatted lines, a ParseError is raised.
- Use find_tag(wikitext, 'googlemap') to find the googlemap tag within an arbitrary
- wikitext before using this function.
-
- :param wikitext: wikitext containing the template. Example:
-
- wikitext = '''
- <googlemap version="0.9" lat="47.113291" lon="11.272337" zoom="15">
- (Parkplatz)47.114958,11.266026
- Parkplatz
-
- (Gasthaus) 47.114715, 11.266262, Alt Bärnbad (Gasthaus)
- 6#FF014E9A
- 47.114715,11.266262
- 47.114135,11.268381
- 47.113421,11.269322
- 47.11277,11.269979
- 47.112408,11.271119
- </googlemap>
- '''
- :returns: The tuple (attributes, coords, paths) is returned.
- attributes is a dict that contains the attribues that are present
- (e.g. lon, lat, zoom, width, height) converted to float (lon, lat) or int.
- coords is a list of (lon, lat, symbol, title) tuples.
- paths is a list of (style, coords) tuples.
- coords is again a list of (lon, lat, symbol, title) tuples."""
-
- def is_coord(line):
- """Returns True if the line contains a coordinate."""
- match = re.search('[0-9]{1,2}\.[0-9]+, ?[0-9]{1,2}\.[0-9]+', line)
- return not match is None
-
- def is_path(line):
- """Returns True if the line contains a path style definition."""
- match = re.match('[0-9]#[0-9a-fA-F]{8}', line)
- return not match is None
-
- def parse_coord(line):
- """Returns (lon, lat, symbol, title). If symbol or text is not present, None is returned."""
- match = re.match(u'\(([^)]+)\) ?([0-9]{1,2}\.[0-9]+), ?([0-9]{1,2}\.[0-9]+), ?(.*)', line)
- if not match is None: return (float(match.group(3)), float(match.group(2)), match.group(1), match.group(4))
- match = re.match(u'\(([^)]+)\) ?([0-9]{1,2}\.[0-9]+), ?([0-9]{1,2}\.[0-9]+)', line)
- if not match is None: return (float(match.group(3)), float(match.group(2)), match.group(1), None)
- match = re.match(u'([0-9]{1,2}\.[0-9]+), ?([0-9]{1,2}\.[0-9]+), ?(.*)', line)
- if not match is None: return (float(match.group(2)), float(match.group(1)), None, match.group(3))
- match = re.match(u'([0-9]{1,2}\.[0-9]+), ?([0-9]{1,2}\.[0-9]+)', line)
- if not match is None: return (float(match.group(2)), float(match.group(1)), None, None)
- return ParseError(u'Could not parse line ' + line)
-
- start, content, endtag, end = find_tag(wikitext, 'googlemap')
- if start is None:
- raise ParseError(u'<googlemap> tag not found.')
- if content is None:
- xml_only = wikitext[start:endtag]
- else:
- xml_only = wikitext[start:content]+wikitext[endtag:end]
-
- try:
- gm = xml.etree.ElementTree.XML(xml_only.encode('UTF8'))
- except xml.etree.ElementTree.ParseError as e:
- row, column = e.position
- raise ParseError(u"XML parse error in <googlemap ...>.")
-
- # parse attributes
- attributes = {}
- try:
- for key in ['lon', 'lat']:
- if gm.get(key) is not None:
- attributes[key] = float(gm.get(key))
- for key in ['zoom', 'width', 'height']:
- if gm.get(key) is not None:
- attributes[key] = int(gm.get(key))
- except ValueError as error:
- raise ParseError(u'Error at parsing attribute {0} of <googlemap>: {1}'.format(key, unicode(error)))
-
- # parse points and lines
- coords = []
- paths = []
- lines = wikitext[content:endtag].split("\n")
- i = 0
- while i < len(lines):
- line = lines[i].strip()
- i += 1
-
- # Skip whitespace
- if len(line) == 0: continue
-
- # Handle a path
- if is_path(line):
- match = re.match(u'([0-9]#[0-9a-fA-F]{8})', line)
- style = match.group(1)
- local_coords = []
- while i < len(lines):
- line = lines[i].strip()
- i += 1
- if is_path(line):
- i -= 1
- break
- if is_coord(line):
- lon, lat, symbol, title = parse_coord(line)
- local_coords.append((lon, lat, symbol, title))
- paths.append((style, local_coords))
- continue
-
- # Handle a coordinate
- if is_coord(line):
- lon, lat, symbol, title = parse_coord(line)
- while i < len(lines):
- line = lines[i].strip()
- i += 1
- if is_path(line) or is_coord(line):
- i -= 1
- break
- if len(line) > 0 and title is None: title = line
- coords.append((lon, lat, symbol, title))
- continue
-
- raise ParseError(u'Unknown line syntax: ' + line)
-
- return (attributes, coords, paths)
-
+ if keylen is None:
+ shown_keys = [len(param.name.strip()) for param in template.params if param.showkey]
+ keylen = max(shown_keys) if shown_keys else 0
+ template.name = '{}\n'.format(template.name.strip())
+ for param in template.params:
+ if param.showkey:
+ param.name = ' {{:{}}} '.format(keylen).format(param.name.strip())
+ value = param.value.strip()
+ if len(value) > 0:
+ param.value = ' {}\n'.format(value)
+ else:
+ param.value = '\n'
+
+
+def format_template_oneline(template):
+ """Formats a template like this: {{template_name|param| }}
+ (whitespace is stripped and empty parameters are replaced with one space)."""
+ template.name = template.name.strip()
+ for param in template.params:
+ if param.showkey:
+ param.name = param.name.strip()
+ value = param.value.strip()
+ if value == '':
+ value = ' '
+ param.value = value
+
+
+def dbkey_to_title(value):
+ """Converts a article database key to a article title. Private function secureAndSplit() of the Title class
+ on line 3316 of includes/Title.php says:
+ $this->mTextform = str_replace( '_', ' ', $this->mDbkeyform );
+ No check for None because a missing title is an error."""
+ return value.replace('_', ' ')
ToastFreeware / philipp / winterrodeln / wrpylib.git / blobdiff