def testTermCompression(self):
"""Test that unreported productions are compressed
Term compression is basically an inlining of terminal
expressions into the calling table. At the moment
the terminal expressions are all duplicated, which may
balloon the size of the grammar, not sure if this will
be an actual problem. As written, this optimization
should provide a significant speed up, but there may
the even more of a speed up if we allow for sharing
the terminal tuples as well.
This:
a:=b <b>:= -c* c:='this'
Should eventually compress to this:
a := -'this'*
"""
failures = []
for first, second in [
("""a:=b <b>:= -c* c:='this'""", """a := -'this'*"""),
("""a:=b >b<:= c c:= 'this'""", """a := c c:= 'this'"""),
("""a:=b >b<:= c <c>:= 'this'""", """a := 'this'"""),
("""a:=b >b<:= c+ <c>:= 'this'""", """a := 'this'+"""),
# The following will never work, so eventually may raise
# an error or at least give a warning!
("""a:=b,c >b<:= c+ <c>:= 'this'""", """a := 'this'+,'this'"""),
("""a:=b/c >b<:= c+ <c>:= 'this'""", """a := 'this'+/'this'"""),
# This is requiring group-compression, which isn't yet written
("""a:=-b/c >b<:= c+ <c>:= 'this'""", """a := -'this'+/'this'"""),
("""a := (table1 / table2 / any_line)*
<any_line> := ANY*, EOL
<ANY> := -EOL
<EOL> := '\n'
table1 := 'a'
table2 := 'b'
""", """a := (table1 / table2 / (-'\n'*, '\n'))*
table1 := 'a'
table2 := 'b'
"""),
("""a:= b,c <b>:= -c* <c>:= '\n'""", """a := -'\n'*,'\n'"""),
]:
pFirst = Parser(first, "a")
pSecond = Parser(second, "a")
tFirst = pFirst.buildTagger()
tSecond = pSecond.buildTagger()
if not rcmp(tFirst, tSecond):
tFirstRepr = pprint.pformat(tFirst)
tSecondRepr = pprint.pformat(tSecond)
failures.append(
"""%(first)r did not produce the same parser as %(second)r\n\t%(tFirstRepr)s\n\t%(tSecondRepr)s"""
% locals())
if failures:
raise ValueError("\n".join(failures))
def testTermCompression( self ):
"""Test that unreported productions are compressed
Term compression is basically an inlining of terminal
expressions into the calling table. At the moment
the terminal expressions are all duplicated, which may
balloon the size of the grammar, not sure if this will
be an actual problem. As written, this optimization
should provide a significant speed up, but there may
the even more of a speed up if we allow for sharing
the terminal tuples as well.
This:
a:=b <b>:= -c* c:='this'
Should eventually compress to this:
a := -'this'*
"""
failures = []
for first, second in [
("""a:=b <b>:= -c* c:='this'""", """a := -'this'*"""),
("""a:=b >b<:= c c:= 'this'""", """a := c c:= 'this'"""),
("""a:=b >b<:= c <c>:= 'this'""", """a := 'this'"""),
("""a:=b >b<:= c+ <c>:= 'this'""", """a := 'this'+"""),
# The following will never work, so eventually may raise
# an error or at least give a warning!
("""a:=b,c >b<:= c+ <c>:= 'this'""", """a := 'this'+,'this'"""),
("""a:=b/c >b<:= c+ <c>:= 'this'""", """a := 'this'+/'this'"""),
# This is requiring group-compression, which isn't yet written
("""a:=-b/c >b<:= c+ <c>:= 'this'""", """a := -'this'+/'this'"""),
("""a := (table1 / table2 / any_line)*
<any_line> := ANY*, EOL
<ANY> := -EOL
<EOL> := '\n'
table1 := 'a'
table2 := 'b'
""", """a := (table1 / table2 / (-'\n'*, '\n'))*
table1 := 'a'
table2 := 'b'
"""),
("""a:= b,c <b>:= -c* <c>:= '\n'""", """a := -'\n'*,'\n'"""),
]:
pFirst = Parser( first, "a")
pSecond = Parser( second, "a")
tFirst = pFirst.buildTagger()
tSecond = pSecond.buildTagger()
if not rcmp( tFirst , tSecond):
tFirstRepr = pprint.pformat(tFirst)
tSecondRepr = pprint.pformat(tSecond)
failures.append( """%(first)r did not produce the same parser as %(second)r\n\t%(tFirstRepr)s\n\t%(tSecondRepr)s"""%locals())
if failures:
raise ValueError( "\n".join(failures))
def testTermSharing( self ):
"""Test that shared terminal productions are using the same parser"""
first =""" a := b,b >b<:= d d:= 'this'"""
pFirst = Parser( first, "a")
tFirst = pFirst.buildTagger()
b,c = tFirst
assert b is c, """Not sharing the same tuple for b and c instances"""
def testTermSharing(self):
"""Test that shared terminal productions are using the same parser"""
first = """ a := b,b >b<:= d d:= 'this'"""
pFirst = Parser(first, "a")
tFirst = pFirst.buildTagger()
b, c = tFirst
assert b is c, """Not sharing the same tuple for b and c instances"""
class GeneratorAPI1: """Stand-in class supporting operation of SimpleParse 1.0 applications There was really only the one method of interest, parserbyname, everything else was internal (and is now part of simpleparsegrammar.py). """ def __init__( self, production, prebuilt=() ): from simpleparse.parser import Parser self.parser = Parser( production, prebuilts=prebuilt ) def parserbyname( self, name ): """Retrieve a tag-table by production name""" return self.parser.buildTagger( name )
class GeneratorAPI1:
"""Stand-in class supporting operation of SimpleParse 1.0 applications
There was really only the one method of interest, parserbyname,
everything else was internal (and is now part of
simpleparsegrammar.py).
"""
def __init__(self, production, prebuilt=()):
from simpleparse.parser import Parser
self.parser = Parser(production, prebuilts=prebuilt)
def parserbyname(self, name):
"""Retrieve a tag-table by production name"""
return self.parser.buildTagger(name)
class EBNFSpill(object):
DEFAULT_MAX_TIMES_CHAR = 35
DEFAULT_MAX_TIMES_FUNC = 10
DEFAULT_MAX_SELF_RECURSION = 25
DEFAULT_MAX_WALK_RECURSION = 100
def __init__(self,showTags=False,showTagsRecursive=False,recursionLevel=0):
self._reset()
self.showTags=showTags
self.showTagsRecursive=showTagsRecursive
self.recursionLevelObj=recursionLevel
if self.recursionLevelObj>self.DEFAULT_MAX_SELF_RECURSION: raise Exception("a")
#print "INIT",recursionLevel
pass
def __del__(self):
self.recursionLevelObj-=1
pass
def validate(self,data):
return self.parser.parse(data)
def setDeclaration(self,declaration,production):
self.parser = Parser(declaration, production)
self.table = self.parser.buildTagger(production=production)
def setTable(self,table,nodes=None):
self.table = table
self.nodes=nodes or self.nodes
def _reset(self):
self.nodes = {}
self.ctx = [] # context (infos like recurion for table2)
#self.recursionLevelObj=0
self.recursionLevelWalk=0
random.seed()
def setDefaults(self,**kwargs):
valid_defaults = [i for i in dir(self) if i.startswith("DEFAULT_")]
for k,v in kwargs.iteritems():
if k in valid_defaults:
setattr(self,k,v)
else:
raise Exception("Not allowed to change %s to %s (valid options: %s)"%(k,v,valid_defaults))
def getTable(self):
return self.table
def getTagName(self,node):
if self.showTags and node[0]:
return "<%s>"%node[0]
return ""
def checkTypeIterable(self,l):
return isinstance(l, collections.Iterable) and not isinstance(l, basestring)
def checkTypeIterableRecursive(self,l):
return isinstance(l, collections.Iterable) and not isinstance(l, basestring) and isinstance(l,tuple) and isinstance(l[0],list) and isinstance(l[1],int)
def checkTypeNodeBase(self,l):
#checks ( None|str, int, *)
return self.checkTypeIterable(l) and len(l)>=2 and (l[0]==None or isinstance(l[0],basestring)) and isinstance(l[1],int)
def checkTypeNodeWithChilds(self,l):
#print "check_",str(l)[:50]
try:
#print "check_metric",checkTypeNodeBase(l),len(l)>=3 , checkTypeIterable(l[2])
pass
except:
pass
return self.checkTypeNodeBase(l) and len(l)>=3 and self.checkTypeIterable(l[2])
def next(self):
return
def rndTimesFunc(self,sample_func,args,minlen=0,maxlen=None):
maxlen = maxlen or self.DEFAULT_MAX_TIMES_FUNC
maxlen+=1
out = ""
for i in range(random.randrange(minlen,maxlen)):
out+=sample_func(args)
return out
def rndTimes(self,sample,minlen=0,maxlen=None):
maxlen = maxlen or self.DEFAULT_MAX_TIMES_CHAR
maxlen+=1
out = ""
for i in range(random.randrange(minlen,maxlen)):
out+=sample
return out
def rndSelect(self,haystack,sample_len=1,minlen=0,maxlen=None):
maxlen = maxlen or self.DEFAULT_MAX_TIMES_CHAR
maxlen+=1
out = ""
for i in range(random.randrange(minlen,maxlen)):
out += "".join(random.sample(haystack,sample_len))
return out
def eval(self,node):
# different lenght commandos
#print node
#print id(node),node
#if self.recursionLevelObj>self.DEFAULT_MAX_SELF_RECURSION or self.recursionLevelWalk>self.DEFAULT_MAX_WALK_RECURSION:
# return "<recursion_exception>"
if not node:
return ""
if len(node)<3:
raise Exception( "<3 - %s"%repr(node) ) #this is an error!
elif node[1]==Tdef.MATCH_RECURSION_EXCEPTION:
return "<<"
elif node[1]==Tdef.MATCH_RECURSION:
# create a new EBNFSpill object, and resolv this one?
#print node[2],self.nodes[node[2]]
self.recursionLevelObj+=1
try:
x = EBNFSpill(showTags=self.showTagsRecursive,recursionLevel=self.recursionLevelObj)
x.setTable(self.table)
recr_node=self.nodes[node[2]]
except:
return ""
#
#print "REKR",node
#print "REKR2",self.nodes
#print "<DAMN_RECURSION %s wild=%s>"%(node[2],self.ctx)
#return "<RECURSION"
#print "EXCEPT:",node[2],self.nodes
#return self.rndTimes(x.generate(recr_node['obj']), 0, 3)
return self.getTagName(node)+x.generate(recr_node)
# single words/selections
elif len(node)==3:
if node[1]==Tdef.MATCH_WORD or node[1]==Tdef.MATCH_IS:
return self.getTagName(node)+node[2]
elif node[1]==Tdef.MATCH_ALLIN or node[1]==Tdef.MATCH_ISIN:
return self.getTagName(node)+self.rndSelect(node[2],minlen=1,maxlen=1)
elif node[1]==Tdef.MATCH_TABLE:
# (xyz,MATCH_TABLE, <table>, 1) == exact 1
# (xyz,MATCH_TABLE, <table>, 2,1) == *
return self.getTagName(node)+""
#return "<TABLE: %s>"%node[0]
# mostly recursive ones
elif len(node)>3:
# recursions and stuff
if node[1]==Tdef.MATCH_IS or node[1]==Tdef.MATCH_IS:
# like (none,"MATCH_IS",'c',1,0) - choose zero or xx times
return self.getTagName(node)+self.rndTimes(node[2])
elif node[1]==Tdef.MATCH_ALLIN or node[1]==Tdef.MATCH_ISIN:
return self.getTagName(node)+self.rndSelect(node[2])
elif node[1]>=Tdef.MATCH_CALL and node[1]<=Tdef.MATCH_SUBTABLEINLIST:
# (xyz,MATCH_TABLE, <table>, 1) == exact 1
# (xyz,MATCH_TABLE, <table>, 2,1) == *
self.recursionLevelObj+=1
try:
x = EBNFSpill(showTags=self.showTagsRecursive,recursionLevel=self.recursionLevelObj)
except:
return ""
x.setTable(self.table)
#print "<TABLE: %s | %s || %s || nodeid:%s>"%(node[0:1],node[3],self.ctx,id(node[3]))
#print node[2]
#return self.getTagName(node)+""
return self.getTagName(node)+self.rndTimesFunc(x.generate,(node[2]))
return self.getTagName(node)
def generate(self,node=None):
out = ""
for n in self.walk(node):
#print n
#print self.recursionLevelObj,self.recursionLevelWalk
out+= self.eval(n)
return out
def process(self,l):
if self.checkTypeNodeBase(l):
return (l[0],Tdef().toName()[l[1]])+l[2:]
return l
def _checkRecursion(self,node):
# return boolean if boolean=True
nID = id(node)
#print "-->",nID, " NODE ",node
if self.nodes.has_key(nID):
raise StopRecursionException(('[RECURSION of Node=%s]'%nID,Tdef.MATCH_RECURSION,nID))
#self.nodeIDs.append(nID)
#print nID,node
return node
def _trackNode(self,node,nodeID=None):
nID = nodeID or id(node)
#print node
if self.checkTypeNodeBase(node):
#print "ISIN1",Tdef.MATCH_CALL,Tdef.MATCH_SUBTABLEINLIST,node[1],node[1]>=Tdef.MATCH_CALL and node[1]<=Tdef.MATCH_SUBTABLEINLIST
if node[1]>=Tdef.MATCH_CALL and node[1]<=Tdef.MATCH_SUBTABLEINLIST:
#print "ISIN2"
#print "--- add BASE",id(node),node
self.nodes[nID]=node
elif self.checkTypeIterable(node):
#print "--- add LIST",id(node),node
self.nodes[nID]=node
return node
def _pushLevel(self,node):
# add one level. . to check recursion space
if node[1]>=Tdef.MATCH_CALL and node[1]<=Tdef.MATCH_SUBTABLEINLIST \
and len(node)>3 and node[3]==2:
#print "push__"
self.ctx.append(id(node))
return node
def _popLevel(self,node):
if node[1]>=Tdef.MATCH_CALL and node[1]<=Tdef.MATCH_SUBTABLEINLIST \
and len(node)>3 and node[3]==2:
#print "pop___"
return self.ctx.pop()
return node
def walk(self,table=None):
table=table or self.table
if not table: raise Exception("EBNF TagTable not set, please generate [.setDeclaration()] or set one [.setTable()]") #must not be !NONE!, please .setDeclaration() first!
retn = self._walk(table)
self._reset()
return retn
def _walk(self,l):
# check if (None|basestring, int, ... ) > 2
#import time
#time.sleep(0.8)
#print "BEGIN",str(l)[:50]
#recursion check
if self.recursionLevelObj>self.DEFAULT_MAX_SELF_RECURSION or self.recursionLevelWalk>self.DEFAULT_MAX_WALK_RECURSION:
#print self.recursionLevelWalk
#print self.recursionLevelObj
#nID=
#raise StopRecursionException(('[RECURSION of Node=%s]'%nID,Tdef.MATCH_RECURSION,nID))
#print self.nodes
#yield l
#print "StopIter",l
#print self.recursionLevelObj,self.recursionLevelWalk
raise StopIteration("HMM")
#yield (None,Tdef.MATCH_RECURSION_EXCEPTION,())
#raise StopRecursionException(("[RECURSION_EXCEPTION_LEVEL_REACHED]",Tdef.MATCH_RECURSION_EXCEPTION,None))
self.recursionLevelWalk+=1
#print id(l),len(l),l
try:
if self.checkTypeNodeWithChilds(l):
#print "Childs"
self._checkRecursion(l)
yield self._trackNode(l)
self._pushLevel(l)
for e in self._walk(l[2]):
yield e
self._popLevel(l)
elif self.checkTypeNodeBase(l):
#print "Base"
self._checkRecursion(l)
yield self._trackNode(l)
elif self.checkTypeIterableRecursive(l):
#print "xxx",l[0][0]
nID=id(l[0][0])
#print "IterReck"
#print '[RECURSION of Node=%s]'%nID
#TODO: does not work
#fixme: does not work - recurses too much
raise StopRecursionException(('[RECURSION of Node=%s]'%nID,Tdef.MATCH_RECURSION,nID))
elif self.checkTypeIterable(l):
#print "list"
self._checkRecursion(l)
self._trackNode(l) # checkTypeIterableRecursive refs one of these nodes :( // damn need to reparse if this doesnt work out
#self._pushLevel(l)
for e in l:
self._pushLevel(e)
for x in self._walk(e):
yield x #do not check recursion here.. this is not what we want
self._popLevel(e)
#self._popLevel(l)
else:
self._checkRecursion(l)
print "Elem? - ",l
#print self.checkTypeNodeWithChilds(l),self.checkTypeNodeBase(l),self.checkTypeIterable(l)
yield self._trackNode(l)
except StopRecursionException, e:
#print self.nodes[e.getObj()[2]]
#print "Except:",e.getObj()
yield e.getObj()
self.recursionLevelWalk-=1
class LegalRef:
# Kanske detta borde vara 1,2,4,8 osv, så att anroparen kan be om
# LAGRUM | FORESKRIFTER, och så vi kan definera samlingar av
# vanliga kombinationer (exv ALL_LAGSTIFTNING = LAGRUM |
# KORTLAGRUM | FORESKRIFTER | EGLAGSTIFTNING)
LAGRUM = 1 # hänvisningar till lagrum i SFS
KORTLAGRUM = 2 # SFS-hänvisningar på kortform
FORESKRIFTER = 3 # hänvisningar till myndigheters författningssamlingar
EGLAGSTIFTNING = 4 # EG-fördrag, förordningar och direktiv
INTLLAGSTIFTNING = 5 # Fördrag, traktat etc
FORARBETEN = 6 # proppar, betänkanden, etc
RATTSFALL = 7 # Rättsfall i svenska domstolar
MYNDIGHETSBESLUT = 8 # Myndighetsbeslut (JO, ARN, DI...)
EGRATTSFALL = 9 # Rättsfall i EG-domstolen/förstainstansrätten
INTLRATTSFALL = 10 # Europadomstolen
# re_urisegments = re.compile(r'([\w]+://[^/]+/[^\d]*)(\d+:(bih\.
# |N|)?\d+( s\.\d+|))#?(K(\d+)|)(P(\d+)|)(S(\d+)|)(N(\d+)|)')
re_urisegments = re.compile(
r'([\w]+://[^/]+/[^\d]*)(\d+:(bih\.[_ ]|N|)?\d+([_ ]s\.\d+|))#?(K([a-z0-9]+)|)(P([a-z0-9]+)|)(S(\d+)|)(N(\d+)|)')
re_escape_compound = re.compile(
r'\b(\w+-) (och) (\w+-?)(lagen|förordningen)\b', re.UNICODE)
re_escape_named = re.compile(
r'\B(lagens?|balkens?|förordningens?|formens?|ordningens?|kungörelsens?|stadgans?)\b', re.UNICODE)
re_descape_compound = re.compile(
r'\b(\w+-)_(och)_(\w+-?)(lagen|förordningen)\b', re.UNICODE)
re_descape_named = re.compile(
r'\|(lagens?|balkens?|förordningens?|formens?|ordningens?|kungörelsens?|stadgans?)')
re_xmlcharref = re.compile("&#\d+;")
def __init__(self, *args):
if not os.path.sep in __file__:
scriptdir = os.getcwd()
else:
scriptdir = os.path.dirname(__file__)
self.graph = Graph()
n3file = os.path.relpath(scriptdir + "/../../../res/etc/sfs-extra.n3")
# print "loading n3file %s" % n3file
self.graph.load(n3file, format="n3")
self.roots = []
self.uriformatter = {}
self.decl = "" # try to make it unicode clean all the way
self.namedlaws = {}
self.load_ebnf(scriptdir + "/../../../res/etc/base.ebnf")
self.args = args
if self.LAGRUM in args:
productions = self.load_ebnf(scriptdir + "/../../../res/etc/lagrum.ebnf")
for p in productions:
self.uriformatter[p] = self.sfs_format_uri
self.namedlaws.update(self.get_relations(RDFS.label))
self.roots.append("sfsrefs")
self.roots.append("sfsref")
if self.KORTLAGRUM in args:
# om vi inte redan laddat lagrum.ebnf måste vi göra det
# nu, eftersom kortlagrum.ebnf beror på produktioner som
# definerats där
if not self.LAGRUM in args:
self.load_ebnf(scriptdir + "/../../../res/etc/lagrum.ebnf")
productions = self.load_ebnf(
scriptdir + "/../../../res/etc/kortlagrum.ebnf")
for p in productions:
self.uriformatter[p] = self.sfs_format_uri
DCT = Namespace("http://purl.org/dc/terms/")
d = self.get_relations(DCT['alternate'])
self.namedlaws.update(d)
# lawlist = [x.encode(SP_CHARSET) for x in list(d.keys())]
lawlist = list(d.keys())
# Make sure longer law abbreviations come before shorter
# ones (so that we don't mistake "3 § MBL" for "3 § MB"+"L")
# lawlist.sort(cmp=lambda x, y: len(y) - len(x))
lawlist.sort(key=len, reverse=True)
lawdecl = "LawAbbreviation ::= ('%s')\n" % "'/'".join(lawlist)
self.decl += lawdecl
self.roots.insert(0, "kortlagrumref")
if self.EGLAGSTIFTNING in args:
productions = self.load_ebnf(scriptdir + "/../../../res/etc/eglag.ebnf")
for p in productions:
self.uriformatter[p] = self.eglag_format_uri
self.roots.append("eglagref")
if self.FORARBETEN in args:
productions = self.load_ebnf(
scriptdir + "/../../../res/etc/forarbeten.ebnf")
for p in productions:
self.uriformatter[p] = self.forarbete_format_uri
self.roots.append("forarbeteref")
if self.RATTSFALL in args:
productions = self.load_ebnf(scriptdir + "/../../../res/etc/rattsfall.ebnf")
for p in productions:
self.uriformatter[p] = self.rattsfall_format_uri
self.roots.append("rattsfallref")
if self.EGRATTSFALL in args:
productions = self.load_ebnf(scriptdir + "/../../../res/etc/egratt.ebnf")
for p in productions:
self.uriformatter[p] = self.egrattsfall_format_uri
self.roots.append("ecjcaseref")
rootprod = "root ::= (%s/plain)+\n" % "/".join(self.roots)
self.decl += rootprod
self.parser = Parser(self.decl.encode(SP_CHARSET), "root")
self.tagger = self.parser.buildTagger("root")
# util.writefile("tagger.tmp", repr(self.tagger), SP_CHARSET)
# print "tagger length: %d" % len(repr(self.tagger))
self.verbose = False
self.depth = 0
# SFS-specifik kod
self.currentlaw = None
self.currentchapter = None
self.currentsection = None
self.currentpiece = None
self.lastlaw = None
self.currentlynamedlaws = {}
def load_ebnf(self, file):
"""Laddar in produktionerna i den angivna filen i den
EBNF-deklaration som används, samt returnerar alla
*Ref och *RefId-produktioner"""
# base.ebnf contains 0x1A, ie the EOF character on windows,
# therefore we need to read it in binary mode
f = open(file, 'rb')
# assume our ebnf files use the same charset
content = f.read(os.stat(file).st_size).decode(SP_CHARSET)
self.decl += content
f.close()
return [x.group(1) for x in re.finditer(r'(\w+(Ref|RefID))\s*::=', content)]
def get_relations(self, predicate):
d = {}
for obj, subj in self.graph.subject_objects(predicate):
d[six.text_type(subj)] = six.text_type(obj)
return d
def parse(self, indata, baseuri="http://rinfo.lagrummet.se/publ/sfs/9999:999#K9P9S9P9", predicate=None):
assert isinstance(indata, six.text_type)
if indata == "":
return indata # this actually triggered a bug...
# h = hashlib.sha1()
# h.update(indata)
# print "Called with %r (%s) (%s)" % (indata, h.hexdigest(), self.verbose)
self.predicate = predicate
self.baseuri = baseuri
if baseuri:
m = self.re_urisegments.match(baseuri)
if m:
self.baseuri_attributes = {'baseuri': m.group(1),
'law': m.group(2),
'chapter': m.group(6),
'section': m.group(8),
'piece': m.group(10),
'item': m.group(12)}
else:
self.baseuri_attributes = {'baseuri': baseuri}
else:
self.baseuri_attributes = {}
# Det är svårt att få EBNF-grammatiken att känna igen
# godtyckliga ord som slutar på ett givet suffix (exv
# 'bokföringslagen' med suffixet 'lagen'). Därför förbehandlar
# vi indatasträngen och stoppar in ett '|'-tecken innan vissa
# suffix. Vi transformerar även 'Radio- och TV-lagen' till
# 'Radio-_och_TV-lagen'
#
# FIXME: Obviously, this shouldn't be done in a general class,
# but rather in a subclas or via proxy/adapter
fixedindata = indata # FIXME: Nonsensical
if self.LAGRUM in self.args:
fixedindata = self.re_escape_compound.sub(
r'\1_\2_\3\4', fixedindata)
fixedindata = self.re_escape_named.sub(r'|\1', fixedindata)
# print "After: %r" % type(fixedindata)
# SimpleParse har inget stöd för unicodesträngar, så vi
# konverterar intdatat till en bytesträng. Tyvärr får jag inte
# det hela att funka med UTF8, så vi kör xml character
# references istället
fixedindata = fixedindata.encode(SP_CHARSET, 'xmlcharrefreplace')
# Parsea texten med TextTools.tag - inte det enklaste sättet
# att göra det, men om man gör enligt
# Simpleparse-dokumentationen byggs taggertabellen om för
# varje anrop till parse()
if self.verbose:
print(("calling tag with '%s'" % (fixedindata.decode(SP_CHARSET))))
# print "tagger length: %d" % len(repr(self.tagger))
taglist = tag(fixedindata, self.tagger, 0, len(fixedindata))
result = []
root = NodeTree(taglist, fixedindata)
for part in root.nodes:
if part.tag != 'plain' and self.verbose:
sys.stdout.write(self.prettyprint(part))
if part.tag in self.roots:
self.clear_state()
# self.verbose = False
result.extend(self.formatter_dispatch(part))
else:
assert part.tag == 'plain', "Tag is %s" % part.tag
result.append(part.text)
# clear state
if self.currentlaw is not None:
self.lastlaw = self.currentlaw
self.currentlaw = None
if taglist[-1] != len(fixedindata):
log.error('Problem (%d:%d) with %r / %r' % (
taglist[-1] - 8, taglist[-1] + 8, fixedindata, indata))
raise ParseError(
"parsed %s chars of %s (...%s...)" % (taglist[-1], len(indata),
indata[(taglist[-1] - 2):taglist[-1] + 3]))
# Normalisera resultatet, dvs konkatenera intilliggande
# textnoder, och ta bort ev '|'-tecken som vi stoppat in
# tidigare.
normres = []
for i in range(len(result)):
if not self.re_descape_named.search(result[i]):
node = result[i]
else:
if self.LAGRUM in self.args:
text = self.re_descape_named.sub(r'\1', result[i])
text = self.re_descape_compound.sub(r'\1 \2 \3\4', text)
if isinstance(result[i], Link):
# Eftersom Link-objekt är immutable måste vi skapa
# ett nytt och kopiera dess attribut
if hasattr(result[i], 'predicate'):
node = LinkSubject(text, predicate=result[i].predicate,
uri=result[i].uri)
else:
node = Link(text, uri=result[i].uri)
else:
node = text
if (len(normres) > 0
and not isinstance(normres[-1], Link)
and not isinstance(node, Link)):
normres[-1] += node
else:
normres.append(node)
# and finally...
for i in range(len(normres)):
if isinstance(normres[i], Link):
# deal with these later
pass
else:
normres[i] = self.re_xmlcharref.sub(
self.unescape_xmlcharref, normres[i])
return normres
def unescape_xmlcharref(self, m):
# print "Changing %r to a %r" % (m.group(0)[2:-1], unichr(int(m.group(0)[2:-1])))
return chr(int(m.group(0)[2:-1]))
def find_attributes(self, parts, extra={}):
"""recurses through a parse tree and creates a dictionary of
attributes"""
d = {}
self.depth += 1
if self.verbose:
print(
(". " * self.depth + "find_attributes: starting with %s" % d))
if extra:
d.update(extra)
for part in parts:
current_part_tag = part.tag.lower()
if current_part_tag.endswith('refid'):
if ((current_part_tag == 'singlesectionrefid') or
(current_part_tag == 'lastsectionrefid')):
current_part_tag = 'sectionrefid'
d[current_part_tag[:-5]] = part.text.strip()
if self.verbose:
print((". " * self.depth +
"find_attributes: d is now %s" % d))
if part.nodes:
d.update(self.find_attributes(part.nodes, d))
if self.verbose:
print((". " * self.depth + "find_attributes: returning %s" % d))
self.depth -= 1
if self.currentlaw and 'law' not in d:
d['law'] = self.currentlaw
if self.currentchapter and 'chapter' not in d:
d['chapter'] = self.currentchapter
if self.currentsection and 'section' not in d:
d['section'] = self.currentsection
if self.currentpiece and 'piece' not in d:
d['piece'] = self.currentpiece
return d
def find_node(self, root, nodetag):
"""Returns the first node in the tree that has a tag matching nodetag. The search is depth-first"""
if root.tag == nodetag: # base case
return root
else:
for node in root.nodes:
x = self.find_node(node, nodetag)
if x is not None:
return x
return None
def find_nodes(self, root, nodetag):
if root.tag == nodetag:
return [root]
else:
res = []
for node in root.nodes:
res.extend(self.find_nodes(node, nodetag))
return res
def flatten_tokentree(self, part, suffix):
"""returns a 'flattened' tokentree ie for the following tree and the suffix 'RefID'
foo->bar->BlahongaRefID
->baz->quux->Blahonga2RefID
->Blahonga3RefID
->Blahonga4RefID
this should return [BlahongaRefID, Blahonga2RefID, Blahonga3RefID, Blahonga4RefID]"""
l = []
if part.tag.endswith(suffix):
l.append(part)
if not part.nodes:
return l
for subpart in part.nodes:
l.extend(self.flatten_tokentree(subpart, suffix))
return l
def formatter_dispatch(self, part):
# print "Verbositiy: %r" % self.verbose
self.depth += 1
# Finns det en skräddarsydd formatterare?
if "format_" + part.tag in dir(self):
formatter = getattr(self, "format_" + part.tag)
if self.verbose:
print(
((". " * self.depth) + "formatter_dispatch: format_%s defined, calling it" % part.tag))
res = formatter(part)
assert res is not None, "Custom formatter for %s didn't return anything" % part.tag
else:
if self.verbose:
print(
((". " * self.depth) + "formatter_dispatch: no format_%s, using format_tokentree" % part.tag))
res = self.format_tokentree(part)
if res is None:
print(((". " * self.depth) +
"something wrong with this:\n" + self.prettyprint(part)))
self.depth -= 1
return res
def format_tokentree(self, part):
# This is the default formatter. It converts every token that
# ends with a RefID into a Link object. For grammar
# productions like SectionPieceRefs, which contain
# subproductions that also end in RefID, this is not a good
# function to use - use a custom formatter instead.
res = []
if self.verbose:
print(((". " * self.depth) +
"format_tokentree: called for %s" % part.tag))
# this is like the bottom case, or something
if (not part.nodes) and (not part.tag.endswith("RefID")):
res.append(part.text)
else:
if part.tag.endswith("RefID"):
res.append(self.format_generic_link(part))
elif part.tag.endswith("Ref"):
res.append(self.format_generic_link(part))
else:
for subpart in part.nodes:
if self.verbose and part.tag == 'LawRef':
print(
((". " * self.depth) + "format_tokentree: part '%s' is a %s" % (subpart.text, subpart.tag)))
res.extend(self.formatter_dispatch(subpart))
if self.verbose:
print(
((". " * self.depth) + "format_tokentree: returning '%s' for %s" % (res, part.tag)))
return res
def prettyprint(self, root, indent=0):
res = "%s'%s': '%s'\n" % (
" " * indent, root.tag, re.sub(r'\s+', ' ', root.text))
if root.nodes is not None:
for subpart in root.nodes:
res += self.prettyprint(subpart, indent + 1)
return res
else:
return ""
def format_generic_link(self, part, uriformatter=None):
try:
uri = self.uriformatter[part.tag](self.find_attributes([part]))
except KeyError:
if uriformatter:
uri = uriformatter(self.find_attributes([part]))
else:
uri = self.sfs_format_uri(self.find_attributes([part]))
except AttributeError:
# Normal error from eglag_format_uri
return part.text
except:
exc = sys.exc_info()
# If something else went wrong, just return the plaintext
log.warning("(unknown): Unable to format link for text %s (production %s)" %
(part.text, part.tag))
return part.text
if self.verbose:
print((
(". " * self.depth) + "format_generic_link: uri is %s" % uri))
if not uri:
# the formatting function decided not to return a URI for
# some reason (maybe it was a partial/relative reference
# without a proper base uri context
return part.text
elif self.predicate:
return LinkSubject(part.text, uri=uri, predicate=self.predicate)
else:
return Link(part.text, uri=uri)
# FIXME: unify this with format_generic_link
def format_custom_link(self, attributes, text, production):
try:
uri = self.uriformatter[production](attributes)
except KeyError:
uri = self.sfs_format_uri(attributes)
if not uri:
# the formatting function decided not to return a URI for
# some reason (maybe it was a partial/relative reference
# without a proper base uri context
return part.text
elif self.predicate:
return LinkSubject(text, uri=uri, predicate=self.predicate)
else:
return Link(text, uri=uri)
#
# KOD FÖR LAGRUM
def clear_state(self):
self.currentlaw = None
self.currentchapter = None
self.currentsection = None
self.currentpiece = None
def normalize_sfsid(self, sfsid):
# sometimes '1736:0123 2' is given as '1736:0123 s. 2' or
# '1736:0123.2'. This fixes that.
sfsid = re.sub(r'(\d+:\d+)\.(\d)', r'\1 \2', sfsid)
# return sfsid.replace('s. ','').replace('s.','') # more advanced normalizations to come...
return sfsid
def normalize_lawname(self, lawname):
lawname = lawname.replace('|', '').replace('_', ' ').lower()
if lawname.endswith('s'):
lawname = lawname[:-1]
return lawname
def namedlaw_to_sfsid(self, text, normalize=True):
if normalize:
text = self.normalize_lawname(text)
nolaw = [
'aktieslagen',
'anordningen',
'anordningen',
'anslagen',
'arbetsordningen',
'associationsformen',
'avfallsslagen',
'avslagen',
'avvittringsutslagen',
'bergslagen',
'beskattningsunderlagen',
'bolagen',
'bolagsordningen',
'bolagsordningen',
'dagordningen',
'djurslagen',
'dotterbolagen',
'emballagen',
'energislagen',
'ersättningsformen',
'ersättningsslagen',
'examensordningen',
'finansbolagen',
'finansieringsformen',
'fissionsvederlagen',
'flygbolagen',
'fondbolagen',
'förbundsordningen',
'föreslagen',
'företrädesordningen',
'förhandlingsordningen',
'förlagen',
'förmånsrättsordningen',
'förmögenhetsordningen',
'förordningen',
'förslagen',
'försäkringsaktiebolagen',
'försäkringsbolagen',
'gravanordningen',
'grundlagen',
'handelsplattformen',
'handläggningsordningen',
'inkomstslagen',
'inköpssamordningen',
'kapitalunderlagen',
'klockslagen',
'kopplingsanordningen',
'låneformen',
'mervärdesskatteordningen',
'nummerordningen',
'omslagen',
'ordalagen',
'pensionsordningen',
'renhållningsordningen',
'representationsreformen',
'rättegångordningen',
'rättegångsordningen',
'rättsordningen',
'samordningen',
'samordningen',
'skatteordningen',
'skatteslagen',
'skatteunderlagen',
'skolformen',
'skyddsanordningen',
'slagen',
'solvärmeanordningen',
'storslagen',
'studieformen',
'stödformen',
'stödordningen',
'stödordningen',
'säkerhetsanordningen',
'talarordningen',
'tillslagen',
'tivolianordningen',
'trafikslagen',
'transportanordningen',
'transportslagen',
'trädslagen',
'turordningen',
'underlagen',
'uniformen',
'uppställningsformen',
'utvecklingsbolagen',
'varuslagen',
'verksamhetsformen',
'vevanordningen',
'vårdformen',
'ägoanordningen',
'ägoslagen',
'ärendeslagen',
'åtgärdsförslagen',
]
if text in nolaw:
return None
if text in self.currentlynamedlaws:
return self.currentlynamedlaws[text]
elif text in self.namedlaws:
return self.namedlaws[text]
else:
if self.verbose:
# print "(unknown): I don't know the ID of named law [%s]" % text
log.warning(
"(unknown): I don't know the ID of named law [%s]" % text)
return None
def sfs_format_uri(self, attributes):
piecemappings = {'första': '1',
'andra': '2',
'tredje': '3',
'fjärde': '4',
'femte': '5',
'sjätte': '6',
'sjunde': '7',
'åttonde': '8',
'nionde': '9'}
keymapping = {'lawref': 'L',
'chapter': 'K',
'section': 'P',
'piece': 'S',
'item': 'N',
'itemnumeric': 'N',
'element': 'O',
'sentence': 'M', # is this ever used?
}
attributeorder = ['law', 'lawref', 'chapter', 'section',
'element', 'piece', 'item', 'itemnumeric', 'sentence']
if 'law' in attributes:
if attributes['law'].startswith('http://'):
res = ''
else:
res = 'http://rinfo.lagrummet.se/publ/sfs/'
else:
if 'baseuri' in self.baseuri_attributes:
res = self.baseuri_attributes['baseuri']
else:
res = ''
resolvetobase = True
addfragment = False
justincase = None
for key in attributeorder:
if key in attributes:
resolvetobase = False
val = attributes[key]
elif (resolvetobase and key in self.baseuri_attributes):
val = self.baseuri_attributes[key]
else:
val = None
if val:
if not isinstance(val, six.text_type):
val = val.decode(SP_CHARSET)
if addfragment:
res += '#'
addfragment = False
if (key in ['piece', 'itemnumeric', 'sentence'] and val in piecemappings):
res += '%s%s' % (
keymapping[key], piecemappings[val.lower()])
else:
if key == 'law':
val = self.normalize_sfsid(val)
val = val.replace(" ", "_")
res += val
addfragment = True
else:
if justincase:
res += justincase
justincase = None
val = val.replace(" ", "")
val = val.replace("\n", "")
val = val.replace("\r", "")
res += '%s%s' % (keymapping[key], val)
else:
if key == 'piece':
justincase = "S1"
return res
def format_ChapterSectionRefs(self, root):
assert(root.tag == 'ChapterSectionRefs')
assert(len(root.nodes) == 3) # ChapterRef, wc, SectionRefs
part = root.nodes[0]
self.currentchapter = part.nodes[0].text.strip()
if self.currentlaw:
res = [self.format_custom_link({'law': self.currentlaw,
'chapter': self.currentchapter},
part.text,
part.tag)]
else:
res = [self.format_custom_link({'chapter': self.currentchapter},
part.text,
part.tag)]
res.extend(self.formatter_dispatch(root.nodes[1]))
res.extend(self.formatter_dispatch(root.nodes[2]))
self.currentchapter = None
return res
def format_ChapterSectionPieceRefs(self, root):
assert(root.nodes[0].nodes[0].tag == 'ChapterRefID')
self.currentchapter = root.nodes[0].nodes[0].text.strip()
res = []
for node in root.nodes:
res.extend(self.formatter_dispatch(node))
return res
def format_LastSectionRef(self, root):
# the last section ref is a bit different, since we want the
# ending double section mark to be part of the link text
assert(root.tag == 'LastSectionRef')
assert(len(root.nodes) == 3) # LastSectionRefID, wc, DoubleSectionMark
sectionrefid = root.nodes[0]
sectionid = sectionrefid.text
return [self.format_generic_link(root)]
def format_SectionPieceRefs(self, root):
assert(root.tag == 'SectionPieceRefs')
self.currentsection = root.nodes[0].nodes[0].text.strip()
res = [self.format_custom_link(self.find_attributes([root.nodes[2]]),
"%s %s" % (root.nodes[0]
.text, root.nodes[2].text),
root.tag)]
for node in root.nodes[3:]:
res.extend(self.formatter_dispatch(node))
self.currentsection = None
return res
def format_SectionPieceItemRefs(self, root):
assert(root.tag == 'SectionPieceItemRefs')
self.currentsection = root.nodes[0].nodes[0].text.strip()
self.currentpiece = root.nodes[2].nodes[0].text.strip()
res = [self.format_custom_link(self.find_attributes([root.nodes[2]]),
"%s %s" % (root.nodes[0]
.text, root.nodes[2].text),
root.tag)]
for node in root.nodes[3:]:
res.extend(self.formatter_dispatch(node))
self.currentsection = None
self.currentpiece = None
return res
# This is a special case for things like '17-29 och 32 §§ i lagen
# (2004:575)', which picks out the LawRefID first and stores it in
# .currentlaw, so that find_attributes finds it
# automagically. Although now it seems to be branching out and be
# all things to all people.
def format_ExternalRefs(self, root):
assert(root.tag == 'ExternalRefs')
# print "DEBUG: start of format_ExternalRefs; self.currentlaw is %s" % self.currentlaw
lawrefid_node = self.find_node(root, 'LawRefID')
if lawrefid_node is None:
# Ok, no explicit LawRefID found, lets see if this is a named law that we have the ID for
# namedlaw_node = self.find_node(root, 'NamedLawExternalLawRef')
namedlaw_node = self.find_node(root, 'NamedLaw')
if namedlaw_node is None:
# As a last chance, this might be a reference back to a previously
# mentioned law ("...enligt 4 § samma lag")
samelaw_node = self.find_node(root, 'SameLaw')
assert(samelaw_node is not None)
if self.lastlaw is None:
log.warning(
"(unknown): found reference to \"{samma,nämnda} {lag,förordning}\", but self.lastlaw is not set")
self.currentlaw = self.lastlaw
else:
# the NamedLaw case
self.currentlaw = self.namedlaw_to_sfsid(namedlaw_node.text)
if self.currentlaw is None:
# unknow law name - in this case it's better to
# bail out rather than resolving chapter/paragraph
# references relative to baseuri (which is almost
# certainly wrong)
return [root.text]
else:
self.currentlaw = lawrefid_node.text
if self.find_node(root, 'NamedLaw'):
namedlaw = self.normalize_lawname(
self.find_node(root, 'NamedLaw').text)
# print "remember that %s is %s!" % (namedlaw, self.currentlaw)
self.currentlynamedlaws[namedlaw] = self.currentlaw
# print "DEBUG: middle of format_ExternalRefs; self.currentlaw is %s" % self.currentlaw
if self.lastlaw is None:
# print "DEBUG: format_ExternalRefs: setting self.lastlaw to %s" % self.currentlaw
self.lastlaw = self.currentlaw
# if the node tree only contains a single reference, it looks
# better if the entire expression, not just the
# chapter/section part, is linked. But not if it's a
# "anonymous" law ('1 § i lagen (1234:234) om blahonga')
if (len(self.find_nodes(root, 'GenericRefs')) == 1 and
len(self.find_nodes(root, 'SectionRefID')) == 1 and
len(self.find_nodes(root, 'AnonymousExternalLaw')) == 0):
res = [self.format_generic_link(root)]
else:
res = self.format_tokentree(root)
return res
def format_SectionItemRefs(self, root):
assert(root.nodes[0].nodes[0].tag == 'SectionRefID')
self.currentsection = root.nodes[0].nodes[0].text.strip()
# res = self.formatter_dispatch(root.nodes[0]) # was formatter_dispatch(self.root)
res = self.format_tokentree(root)
self.currentsection = None
return res
def format_PieceItemRefs(self, root):
self.currentpiece = root.nodes[0].nodes[0].text.strip()
res = [self.format_custom_link(
self.find_attributes([root.nodes[2].nodes[0]]),
"%s %s" % (root.nodes[0].text, root.nodes[2].nodes[0].text),
root.tag)]
for node in root.nodes[2].nodes[1:]:
res.extend(self.formatter_dispatch(node))
self.currentpiece = None
return res
def format_ChapterSectionRef(self, root):
assert(root.nodes[0].nodes[0].tag == 'ChapterRefID')
self.currentchapter = root.nodes[0].nodes[0].text.strip()
return [self.format_generic_link(root)]
def format_AlternateChapterSectionRefs(self, root):
assert(root.nodes[0].nodes[0].tag == 'ChapterRefID')
self.currentchapter = root.nodes[0].nodes[0].text.strip()
# print "Self.currentchapter is now %s" % self.currentchapter
res = self.format_tokentree(root)
self.currentchapter = None
return res
def format_ExternalLaw(self, root):
self.currentchapter = None
return self.formatter_dispatch(root.nodes[0])
def format_ChangeRef(self, root):
id = self.find_node(root, 'LawRefID').data
return [self.format_custom_link({'lawref': id},
root.text,
root.tag)]
def format_SFSNr(self, root):
if self.baseuri is None:
sfsid = self.find_node(root, 'LawRefID').data
baseuri = 'http://rinfo.lagrummet.se/publ/sfs/%s#' % sfsid.decode(SP_CHARSET)
self.baseuri_attributes = {'baseuri': baseuri}
return self.format_tokentree(root)
def format_NamedExternalLawRef(self, root):
resetcurrentlaw = False
# print "format_NamedExternalLawRef: self.currentlaw is %r" % self.currentlaw
if self.currentlaw is None:
resetcurrentlaw = True
lawrefid_node = self.find_node(root, 'LawRefID')
if lawrefid_node is None:
self.currentlaw = self.namedlaw_to_sfsid(root.text)
else:
self.currentlaw = lawrefid_node.text
namedlaw = self.normalize_lawname(
self.find_node(root, 'NamedLaw').text)
# print "remember that %s is %s!" % (namedlaw, self.currentlaw)
self.currentlynamedlaws[namedlaw] = self.currentlaw
# print "format_NamedExternalLawRef: self.currentlaw is now %r" % self.currentlaw
# print "format_NamedExternalLawRef: self.baseuri is %r" % self.baseuri
if self.currentlaw is None: # if we can't find a ID for this law, better not <link> it
res = [root.text]
else:
res = [self.format_generic_link(root)]
# print "format_NamedExternalLawRef: self.baseuri is %r" % self.baseuri
if self.baseuri is None and self.currentlaw is not None:
# print "format_NamedExternalLawRef: setting baseuri_attributes"
# use this as the new baseuri_attributes
m = self.re_urisegments.match(self.currentlaw)
if m:
self.baseuri_attributes = {'baseuri': m.group(1),
'law': m.group(2),
'chapter': m.group(6),
'section': m.group(8),
'piece': m.group(10),
'item': m.group(12)}
else:
self.baseuri_attributes = {
'baseuri': 'http://rinfo.lagrummet.se/publ/sfs/' + self.currentlaw + '#'}
if resetcurrentlaw:
if self.currentlaw is not None:
self.lastlaw = self.currentlaw
self.currentlaw = None
return res
#
# KOD FÖR KORTLAGRUM
def format_AbbrevLawNormalRef(self, root):
lawabbr_node = self.find_node(root, 'LawAbbreviation')
self.currentlaw = self.namedlaw_to_sfsid(
lawabbr_node.text, normalize=False)
res = [self.format_generic_link(root)]
if self.currentlaw is not None:
self.lastlaw = self.currentlaw
self.currentlaw = None
return res
def format_AbbrevLawShortRef(self, root):
assert(root.nodes[0].tag == 'LawAbbreviation')
assert(root.nodes[2].tag == 'ShortChapterSectionRef')
self.currentlaw = self.namedlaw_to_sfsid(
root.nodes[0].text, normalize=False)
shortsection_node = root.nodes[2]
assert(shortsection_node.nodes[0].tag == 'ShortChapterRefID')
assert(shortsection_node.nodes[2].tag == 'ShortSectionRefID')
self.currentchapter = shortsection_node.nodes[0].text
self.currentsection = shortsection_node.nodes[2].text
res = [self.format_generic_link(root)]
self.currentchapter = None
self.currentsection = None
self.currentlaw = None
return res
#
# KOD FÖR FORARBETEN
def forarbete_format_uri(self, attributes):
# res = self.baseuri_attributes['baseuri']
res = 'http://rinfo.lagrummet.se/'
resolvetobase = True
addfragment = False
for key, val in list(attributes.items()):
if key == 'prop':
res += "publ/prop/%s" % val
elif key == 'bet':
res += "publ/bet/%s" % val
elif key == 'skrivelse':
res += "publ/rskr/%s" % val
elif key == 'celex':
if len(val) == 8: # incorrectly formatted, uses YY instead of YYYY
val = val[0] + '19' + val[1:]
res += "ext/eur-lex/%s" % val
if 'sidnr' in attributes:
res += "#s%s" % attributes['sidnr']
return res
def format_ChapterSectionRef(self, root):
assert(root.nodes[0].nodes[0].tag == 'ChapterRefID')
self.currentchapter = root.nodes[0].nodes[0].text.strip()
return [self.format_generic_link(root)]
#
# KOD FÖR EGLAGSTIFTNING
def eglag_format_uri(self, attributes):
res = 'http://rinfo.lagrummet.se/ext/celex/'
if not 'akttyp' in attributes:
if 'forordning' in attributes:
attributes['akttyp'] = 'förordning'
elif 'direktiv' in attributes:
attributes['akttyp'] = 'direktiv'
if 'akttyp' not in attributes:
raise AttributeError("Akttyp saknas")
# Om hur CELEX-nummer konstrueras
# https://www.infotorg.sema.se/infotorg/itweb/handbook/rb/hlp_celn.htm
# https://www.infotorg.sema.se/infotorg/itweb/handbook/rb/hlp_celf.htm
# Om hur länkning till EURLEX ska se ut:
# http://eur-lex.europa.eu/sv/tools/help_syntax.htm
# Absolut URI?
if 'ar' in attributes and 'lopnummer' in attributes:
sektor = '3'
rattslig_form = {'direktiv': 'L',
'förordning': 'R'}
if len(attributes['ar']) == 2:
attributes['ar'] = '19' + attributes['ar']
res += "%s%s%s%04d" % (sektor, attributes['ar'],
rattslig_form[attributes['akttyp']],
int(attributes['lopnummer']))
else:
if not self.baseuri_attributes['baseuri'].startswith(res):
# FIXME: should we warn about this?
# print "Relative reference, but base context %s is not a celex context" %
# self.baseuri_attributes['baseuri']
return None
if 'artikel' in attributes:
res += "#%s" % attributes['artikel']
if 'underartikel' in attributes:
res += ".%s" % attributes['underartikel']
return res
#
# KOD FÖR RATTSFALL
def rattsfall_format_uri(self, attributes):
# Listan härledd från containers.n3/rattsfallsforteckningar.n3 i
# rinfoprojektets källkod - en ambitiösare lösning vore att läsa
# in de faktiska N3-filerna i en rdflib-graf.
containerid = {'NJA': '/publ/rattsfall/nja/',
'RH': '/publ/rattsfall/rh/',
'MÖD': '/publ/rattsfall/mod/',
'RÅ': '/publ/rattsfall/ra/',
'RK': '/publ/rattsfall/rk/',
'MIG': '/publ/rattsfall/mig/',
'AD': '/publ/rattsfall/ad/',
'MD': '/publ/rattsfall/md/',
'FÖD': '/publ/rattsfall/fod/'}
# res = self.baseuri_attributes['baseuri']
if 'nja' in attributes:
attributes['domstol'] = attributes['nja']
assert 'domstol' in attributes, "No court provided"
assert attributes[
'domstol'] in containerid, "%s is an unknown court" % attributes['domstol']
res = "http://rinfo.lagrummet.se" + containerid[attributes['domstol']]
if 'lopnr' in attributes and ":" in attributes['lopnr']:
(attributes['ar'], attributes['lopnr']) = lopnr.split(":", 1)
if attributes['domstol'] == 'NJA':
# FIXME: URIs should be based on publikationsordinal, not
# pagenumber (which this in effect is) - but this requires
# a big lookup table/database/graph with
# pagenumber-to-ordinal-mappings
res += '%ss%s' % (attributes['ar'], attributes['sidnr'])
else:
res += '%s:%s' % (attributes['ar'], attributes['lopnr'])
return res
#
# KOD FÖR EGRÄTTSFALL
def egrattsfall_format_uri(self, attributes):
descriptormap = {'C': 'J', # Judgment of the Court
'T': 'A', # Judgment of the Court of First Instance
'F': 'W', # Judgement of the Civil Service Tribunal
}
# FIXME: Change this before the year 2054 (as ECJ will
# hopefully have fixed their case numbering by then)
if len(attributes['year']) == 2:
if int(attributes['year']) < 54:
year = "20" + attributes['year']
else:
year = "19" + attributes['year']
else:
year = attributes['year']
serial = '%04d' % int(attributes['serial'])
descriptor = descriptormap[attributes['decision']]
uri = "http://lagen.nu/ext/celex/6%s%s%s" % (year, descriptor, serial)
return uri
class Reference:
LAGRUM = 1
KORTALAGRUM = 2
FORESKRIFTER = 3
FORARBETEN = 6
reUriSegments = re.compile(r'([\w]+://[^/]+/[^\d]*)(\d+:(bih\.[_ ]|N|)?\d+([_ ]s\.\d+|))#?(K([a-z0-9]+)|)(P([a-z0-9]+)|)(S(\d+)|)(N(\d+)|)')
reEscapeCompound = re.compile(r'\b(\w+-) (och) (\w+-?)(lagen|förordningen)\b', re.UNICODE)
reEscapeNamed = re.compile(r'\B(lagens?|balkens?|förordningens?|formens?|ordningens?|kungörelsens?|stadgans?)\b', re.UNICODE)
reDescapeCompound = re.compile(r'\b(\w+-)_(och)_(\w+-?)(lagen|förordningen)\b', re.UNICODE)
reDescapeNamed = re.compile(r'\|(lagens?|balkens?|förordningens?|formens?|ordningens?|kungörelsens?|stadgans?)')
reXmlCharref = re.compile('&#\d+;')
def __init__(self, *args):
scriptDir = os.getcwd()
self.graph = Graph()
n3File = Util.relpath(scriptDir + '/etc/sfs-extra.n3')
self.graph.load(n3File, format='n3')
self.roots = []
self.uriFormatter = {}
self.decl = ''
self.namedLaws = {}
self.loadEbnf(scriptDir + '/etc/base.ebnf')
self.args = args
if self.LAGRUM in args:
prods = self.loadEbnf(scriptDir + '/etc/lagrum.ebnf')
for p in prods:
self.uriFormatter[p] = self.sfsFormatUri
self.namedLaws.update(self.getRelationship(RDFS.label))
self.roots.append('sfsrefs')
self.roots.append('sfsref')
if self.KORTALAGRUM in args:
# TODO: Fix korta lagrum also
pass
if self.FORARBETEN in args:
prods = self.loadEbnf(scriptDir + '/etc/forarbeten.ebnf')
for p in prods:
self.uriFormatter[p] = self.forarbeteFormatUri
self.roots.append('forarbeteref')
self.decl += 'root ::= (%s/plain)+\n' % '/'.join(self.roots)
self.parser = Parser(self.decl, 'root')
self.tagger = self.parser.buildTagger('root')
self.depth = 0
#SFS specific settings
self.currentLaw = None
self.currentChapter = None
self.currentSection = None
self.currentPiece = None
self.lastLaw = None
self.currentNamedLaws = {}
def loadEbnf(self, file):
"""Loads the syntax from a given EBNF file"""
f = open(file)
syntax = f.read()
self.decl += syntax
f.close()
return [x.group(1) for x in re.finditer(r'(\w+(Ref|RefID))\s*::=', syntax)]
def getRelationship(self, predicate):
d = {}
for obj, subj in self.graph.subject_objects(predicate):
d[unicode(subj)] = unicode(obj)
return d
def parse(self, indata, baseUri='http://rinfo.lagrummet.se/publ/sfs/9999:999#K9P9S9P9',predicate=None):
if indata == '':
return indata
self.predicate = predicate
self.baseUri = baseUri
if baseUri:
m = self.reUriSegments.match(baseUri)
if m:
self.baseUriAttrs = {'baseUri' : m.group(1),
'law' : m.group(2),
'chapter' : m.group(6),
'section' : m.group(8),
'piece' : m.group(10),
'item' : m.group(12)}
else:
self.baseUriAttrs = {'baseUri':baseUri}
else:
self.baseUriAttrs = {}
fixedIndata = unicode(indata)
if self.LAGRUM in self.args:
fixedIndata = self.reEscapeCompound.sub(r'\1_\2_\3\4', fixedIndata)
fixedIndata = self.reEscapeNamed.sub(r'|\1', fixedIndata)
if isinstance(fixedIndata, unicode):
fixedIndata = fixedIndata.encode(SP_CHARSET, 'xmlcharrefreplace')
tagList = tag(fixedIndata, self.tagger,0,len(fixedIndata))
res = []
root = NodeTree(tagList, fixedIndata)
for n in root.nodes:
if n.tag in self.roots:
self.clearState()
res.extend(self.formatterDispatch(n))
else:
assert n.tag == 'plain', 'Tag is %s' % n.tag
res.append(n.text)
if self.currentLaw != None:
self.lastLaw = self.currentLaw
self.currentLaw = None
if tagList[-1] != len(fixedIndata):
#TODO: Add Error
raise ParseError, 'Parsed %s chars of %s (...%s...)' % (tagList[-1], len(indata), indata[(tagList[-1]-2):tagList[-1]+3])
# Normalize the result, concat and remove '|'
result = []
for i in range(len(res)):
if not self.reDescapeNamed.search(res[i]):
node = res[i]
else:
if self.LAGRUM in self.args:
text = self.reDescapeNamed.sub(r'\1', res[i])
text = self.reDescapeCompound.sub(r'\1 \2 \3\4', text)
if isinstance(res[i], Link):
# A Link obj is immutable so we have to
# create a new and copy its attrs
if hasattr(res[i], 'predicate'):
node = LinkSubject(text, predicate=res[i].predicate, uri=res[i].uri)
else:
node = Link(text, uri=res[i].uri)
else:
node = text
if (len(result) > 0
and not isinstance(result[-1], Link)
and not isinstance(node, Link)):
result[-1] += node
else:
result.append(node)
for i in range(len(result)):
if isinstance(result[i], Link):
pass
else:
result[i] = self.reXmlCharref.sub(self.unescapeXmlCharref, result[i])
return result
def unescapeXmlCharref(self, m):
return unichr(int(m.group(0)[2:-1]))
def findAttrs(self, parts, extra={}):
"""Creates a dict of attributes through a tree"""
d = {}
self.depth += 1
if extra:
d.update(extra)
for part in parts:
currentPartTag = part.tag.lower()
if currentPartTag.endswith('refid'):
if ((currentPartTag == 'singelsectionrefid') or
(currentPartTag == 'lastsectionrefid')):
currentPartTag = 'sectionrefid'
d[currentPartTag[:-5]] = part.text.strip()
if part.nodes:
d.update(self.findAttrs(part.nodes, d))
self.depth -= 1
if self.currentLaw and 'law' not in d:
d['law'] = self.currentLaw
if self.currentChapter and 'chapter' not in d:
d['chapter'] = self.currentChapter
if self.currentSection and 'section' not in d:
d['section'] = self.currentSection
if self.currentPiece and 'piece' not in d:
d['piece'] = self.currentPiece
return d
def findNode(self, root, nodeTag):
"""Returns the first node in the tree that has a matching tag, dfs."""
if root.tag == nodeTag:
return root
else:
for node in root.nodes:
x = self.findNode(node, nodeTag)
if x != None:
return x
return None
def findNodes(self, root, nodeTag):
if root.tag == nodeTag:
return [root]
else:
res = []
for node in root.nodes:
res.extend(self.findNodes(node, nodeTag))
return res
def formatterDispatch(self, part):
self.depth += 1
if 'format_' + part.tag in dir(self):
formatter = getattr(self,'format_'+part.tag)
res = formatter(part)
assert res != None, 'Custom formatter for %s didnt return anythin' % part.tag
else:
res = self.formatTokentree(part)
self.depth -= 1
return res
def formatTokentree(self, part):
res = []
if (not part.nodes) and (not part.tag.endswith('RefID')):
res.append(part.text)
else:
if part.tag.endswith('RefID'):
res.append(self.formatGenericLink(part))
elif part.tag.endswith('Ref'):
res.append(self.formatGenericLink(part))
else:
for p in part.nodes:
res.extend(self.formatterDispatch(p))
return res
def formatGenericLink(self, part, uriFormatter=None):
try:
uri = self.uriFormatter[part.tag](self.findAttrs([part]))
except KeyError:
if uriFormatter:
uri = uriFormatter(self.findAttrs([part]))
else:
uri = self.sfsFormatUri(self.findAttrs([part]))
except AttributeError:
return part.text
except:
exc = sys.exc_info()
return part.text
if not uri:
return part.text
elif self.predicate:
return LinkSubject(part.text, uri=uri, predicate=self.predicate)
else:
return Link(part.text, uri=uri)
def formatCustomLink(self, attrs, text, production):
try:
uri = self.uriFormatter[production](attrs)
except KeyError:
uri = self.sfsFormatUri(attrs)
if not uri:
return part.text
elif self.predicate:
return LinkSubject(text, uri=uri, predicate=self.predicate)
else:
return Link(text, uri=uri)
def clearState(self):
self.currentLaw = None
self.currentChapter = None
self.currentSection = None
self.currentPiece = None
def normalizeSfsId(self, sfsId):
sfsId = re.sub(r'(\d+:\d+)\.(\d)', r'\1 \2', sfsId)
return sfsId
def normalizeLawName(self, lawName):
lawName = lawName.replace('|','').replace('_',' ').lower()
if lawName.endswith('s'):
lawName = lawName[:-1]
return lawName
def namedLawToSfsid(self, text, normalize=True):
if normalize:
text = self.normalizeLawName(text)
noLaw = [
u'aktieslagen',
u'anordningen',
u'anordningen',
u'anslagen',
u'arbetsordningen',
u'associationsformen',
u'avfallsslagen',
u'avslagen',
u'avvittringsutslagen',
u'bergslagen',
u'beskattningsunderlagen',
u'bolagen',
u'bolagsordningen',
u'bolagsordningen',
u'dagordningen',
u'djurslagen',
u'dotterbolagen',
u'emballagen',
u'energislagen',
u'ersättningsformen',
u'ersättningsslagen',
u'examensordningen',
u'finansbolagen',
u'finansieringsformen',
u'fissionsvederlagen',
u'flygbolagen',
u'fondbolagen',
u'förbundsordningen',
u'föreslagen',
u'företrädesordningen',
u'förhandlingsordningen',
u'förlagen',
u'förmånsrättsordningen',
u'förmögenhetsordningen',
u'förordningen',
u'förslagen',
u'försäkringsaktiebolagen',
u'försäkringsbolagen',
u'gravanordningen',
u'grundlagen',
u'handelsplattformen',
u'handläggningsordningen',
u'inkomstslagen',
u'inköpssamordningen',
u'kapitalunderlagen',
u'klockslagen',
u'kopplingsanordningen',
u'låneformen',
u'mervärdesskatteordningen',
u'nummerordningen',
u'omslagen',
u'ordalagen',
u'pensionsordningen',
u'renhållningsordningen',
u'representationsreformen',
u'rättegångordningen',
u'rättegångsordningen',
u'rättsordningen',
u'samordningen',
u'samordningen',
u'skatteordningen',
u'skatteslagen',
u'skatteunderlagen',
u'skolformen',
u'skyddsanordningen',
u'slagen',
u'solvärmeanordningen',
u'storslagen',
u'studieformen',
u'stödformen',
u'stödordningen',
u'stödordningen',
u'säkerhetsanordningen',
u'talarordningen',
u'tillslagen',
u'tivolianordningen',
u'trafikslagen',
u'transportanordningen',
u'transportslagen',
u'trädslagen',
u'turordningen',
u'underlagen',
u'uniformen',
u'uppställningsformen',
u'utvecklingsbolagen',
u'varuslagen',
u'verksamhetsformen',
u'vevanordningen',
u'vårdformen',
u'ägoanordningen',
u'ägoslagen',
u'ärendeslagen',
u'åtgärdsförslagen']
if text in noLaw:
return None
if self.currentNamedLaws.has_key(text):
return self.currentNamedLaws[text]
elif self.namedLaws.has_key(text):
return self.namedLaws[text]
else:
return None
def sfsFormatUri(self, attrs):
pieceMap = {u'första' :'1',
u'andra' :'2',
u'tredje' :'3',
u'fjärde' :'4',
u'femte' :'5',
u'sjätte' :'6',
u'sjunde' :'7',
u'åttonde':'8',
u'nionde' :'9'}
keyMap = {u'lawref' :'L',
u'chapter':'K',
u'section':'P',
u'piece' :'S',
u'item' :'N',
u'itemnumeric': 'N',
u'element':'O',
u'sentence': 'M'}
attrOrder = ['law', 'lawref', 'chapter', 'section', 'element', 'piece', 'item', 'itemnumeric', 'sentence']
if 'law' in attrs:
if attrs['law'].startswith('http://'):
res = ''
else:
res = 'http://rinfo.lagrummet.se/publ/sfs/'
else:
if 'baseUri' in self.baseUriAttrs:
res = self.baseUriAttrs['baseUri']
else:
res = ''
resolveBase = True
addFragment = False
justInCase = None
for key in attrOrder:
if attrs.has_key(key):
resolveBase = False
val = attrs[key]
elif (resolveBase and self.baseUriAttrs.has_key(key)):
val = self.baseUriAttrs[key]
else:
val = None
if val:
if addFragment:
res += '#'
addFragment = False
if (key in ['piece', 'itemnumeric', 'sentence'] and val in pieceMap):
res += '%s%s' % (keyMap[key], pieceMap[val.lower()])
else:
if key == 'law':
val = self.normalizeSfsId(val)
val = val.replace(' ', '_')
res += val
addFragment = True
else:
if justInCase:
res += justInCase
justInCase = None
val = val.replace(' ', '')
val = val.replace('\n', '')
val = val.replace('\r', '')
res += '%s%s' % (keyMap[key], val)
else:
if key == 'piece':
justInCase = 'S1'
return res
def format_SFSNr(self, root):
if self.baseUri == None:
sfsId = self.findNode(root, 'LawRefID').data
self.baseUriAttrs = {'baseUri':'http://rinfo.lagrummet.se/publ/sfs/'+sfsId+'#'}
return self.formatTokentree(root)
def format_ChangeRef(self, root):
id = self.findNode(root, 'LawRefID').data
return [self.formatCustomLink({'lawref':id},
root.text,
root.tag)]
def format_NamedExternalLawRef(self, root):
resetCurrentLaw = False
if self.currentLaw == None:
resetCurrentLaw = True
lawRefIdNode = self.findNode(root, 'LawRefID')
if lawRefIdNode == None:
self.currentLaw = self.namedLawToSfsid(root.text)
else:
self.currentLaw = lawRefIdNode.text
namedLaw = self.normalizeLawName(self.findNode(root, 'NamedLaw').text)
self.currentNamedLaws[namedLaw] = self.currentLaw
if self.currentLaw == None:
res = [root.text]
else:
res = [self.formatGenericLink(root)]
if self.baseUri == None and self.currentLaw != None:
m = self.reUriSegments.match(self.currentLaw)
if m:
self.baseUriAttrs = {'baseUri' : m.group(1),
'law': m.group(2),
'chapter': m.group(6),
'section': m.group(8),
'piece': m.group(10),
'item': m.group(12)}
else:
self.baseUriAttrs = {'baseUri': 'http://rinfo.lagrummet.se/publ/sfs/' + self.currentLaw + '#'}
if resetCurrentLaw:
if self.currentLaw != None:
self.lastLaw = self.currentLaw
self.currentLaw = None
return res
def format_ChapterSectionRef(self, root):
assert(root.nodes[0].nodes[0].tag == 'ChapterRefID')
self.currentChapter = root.nodes[0].nodes[0].text.strip()
return [self.formatGenericLink(root)]
def format_ChapterSectionPieceRefs(self, root):
assert(root.nodes[0].nodes[0].tag == 'ChapterRefID')
self.currentChapter = root.nodes[0].nodes[0].text.strip()
res = []
for node in root.nodes:
res.extend(self.formatterDispatch(node))
return res
def format_AlternativeChapterSectionRefs(self, root):
print "TODO: Implement me %s" % root.tag
def format_LastSectionRef(self, root):
# We want the ending double section mark to be
# a part of the link
assert(root.tag == 'LastSectionRef')
assert(len(root.nodes) == 3)
sectionRefId = root.nodes[0]
sectionId = sectionRefId.text
return [self.formatGenericLink(root)]
def format_SectionPieceRefs(self, root):
assert(root.tag == 'SectionPieceRefs')
self.currentSection = root.nodes[0].nodes[0].text.strip()
res = [self.formatCustomLink(self.findAttrs([root.nodes[2]]),
'%s %s' % (root.nodes[0].text, root.nodes[2].text),
root.tag)]
for node in root.nodes[3:]:
res.extend(self.formatterDispatch(node))
self.currentSection = None
return res
def format_SectionPieceItemRefs(self, root):
assert(root.tag == 'SectionPieceItemRefs')
self.currentSection = root.nodes[0].nodes[0].text.strip()
self.currentPiece = root.nodes[2].nodes[0].text.strip()
res = [self.formatCustomLink(self.findAttrs([root.nodes[2]]),
'%s %s' % (root.nodes[0].text, root.nodes[2].text),
root.tag)]
for node in root.nodes[3:]:
res.extend(self.formatterDispatch(node))
self.currentSection = None
self.currentPiece = None
return res
def format_SectionItemRefs(self, root):
assert(root.nodes[0].nodes[0].tag == 'SectionRefID')
self.currentSection = root.nodes[0].nodes[0].text.strip()
res = self.formatTokentree(root)
self.currentSection = None
return res
def format_PieceItemRefs(self, root):
self.currentPiece = root.nodes[0].nodes[0].text.strip()
res = [self.formatCustomLink(self.findAttrs([root.nodes[2].nodes[0]]),
'%s %s' % (root.nodes[0].text, root.nodes[2].nodes[0].text),
root.tag)]
for node in root.nodes[2].nodes[1:]:
res.extend(self.formatterDispatch(node))
self.currentPiece = None
return res
def format_ExternalLaw(self, root):
self.currentChapter = None
return self.formatterDispatch(root.nodes[0])
def format_ExternalRefs(self, root):
# Special case for things like '17-29 och 32 §§ i lagen
# (2004:575)' by picking the LawRefID and store it in
# currentLaw do findAttrs will find it.
assert(root.tag == 'ExternalRefs')
lawRefIdNode = self.findNode(root, 'LawRefID')
if lawRefIdNode == None:
namedLawNode = self.findNode(root, 'NamedLaw')
if namedLawNode == None:
sameLawNode = self.findNode(root, 'SameLaw')
assert(sameLawNode != None)
self.currentLaw = self.lastLaw
else:
self.currentLaw = self.namedLawToSfsid(namedLawNode.text)
if self.currentLaw == None:
# Unknown law name, return
return [root.text]
else:
self.currentLaw = lawRefIdNode.text
if self.findNode(root, 'NamedLaw'):
namedLaw = self.normalizeLawName(self.findNode(root, 'NamedLaw').text)
self.currentNamedLaws[namedLaw] = self.currentLaw
if self.lastLaw is None:
self.lastLaw = self.currentLaw
if (len(self.findNodes(root, 'GenericRefs')) == 1 and
len(self.findNodes(root, 'SectionRefID')) == 1 and
len(self.findNodes(root, 'AnonymousExternalLaw')) == 0):
res = [self.formatGenericLink(root)]
else:
res = self.formatTokentree(root)
return res
def forarbeteFormatUri(self, attrs):
res = 'http://rinfo.lagrummet.se/'
resolveBase = True
addFragment = False
for key, val in attrs.items():
if key == 'prop':
res += 'publ/prop/%s' % val
elif key == 'bet':
res += 'ext/bet/%s' % val
elif key == 'skrivelse':
res += 'ext/rskr/%s' % val
elif key == 'celex':
if len(val) == 8:
val = val[0] + '19' + val[1:]
res += 'ext/celex/%s' % val
if 'sidnr' in attrs:
res += '#s%s' % attrs['sidnr']
return res
class LegalRef:
# Kanske detta borde vara 1,2,4,8 osv, så att anroparen kan be om
# LAGRUM | FORESKRIFTER, och så vi kan definera samlingar av
# vanliga kombinationer (exv ALL_LAGSTIFTNING = LAGRUM |
# KORTLAGRUM | FORESKRIFTER | EGLAGSTIFTNING)
LAGRUM = 1 # hänvisningar till lagrum i SFS
KORTLAGRUM = 2 # SFS-hänvisningar på kortform
FORESKRIFTER = 3 # hänvisningar till myndigheters författningssamlingar
EGLAGSTIFTNING = 4 # EG-fördrag, förordningar och direktiv
INTLLAGSTIFTNING = 5 # Fördrag, traktat etc
FORARBETEN = 6 # proppar, betänkanden, etc
RATTSFALL = 7 # Rättsfall i svenska domstolar
MYNDIGHETSBESLUT = 8 # Myndighetsbeslut (JO, ARN, DI...)
EGRATTSFALL = 9 # Rättsfall i EG-domstolen/förstainstansrätten
INTLRATTSFALL = 10 # Europadomstolen
# re_urisegments = re.compile(r'([\w]+://[^/]+/[^\d]*)(\d+:(bih\. |N|)?\d+( s\.\d+|))#?(K(\d+)|)(P(\d+)|)(S(\d+)|)(N(\d+)|)')
re_urisegments = re.compile(r'([\w]+://[^/]+/[^\d]*)(\d+:(bih\.[_ ]|N|)?\d+([_ ]s\.\d+|))#?(K([a-z0-9]+)|)(P([a-z0-9]+)|)(S(\d+)|)(N(\d+)|)')
re_escape_compound = re.compile(r'\b(\w+-) (och) (\w+-?)(lagen|förordningen)\b', re.UNICODE)
re_escape_named = re.compile(r'\B(lagens?|balkens?|förordningens?|formens?|ordningens?|kungörelsens?|stadgans?)\b', re.UNICODE)
re_descape_compound = re.compile(r'\b(\w+-)_(och)_(\w+-?)(lagen|förordningen)\b', re.UNICODE)
re_descape_named = re.compile(r'\|(lagens?|balkens?|förordningens?|formens?|ordningens?|kungörelsens?|stadgans?)')
re_xmlcharref = re.compile("&#\d+;")
def __init__(self,*args):
if not os.path.sep in __file__:
scriptdir = os.getcwd()
else:
scriptdir = os.path.dirname(__file__)
#n3file = os.path.sep.join([scriptdir,"etc","sfs-extra.n3"])
#n3url = "file://" + n3file.replace("\\","/")
#print "scriptdir: %s" % scriptdir
#print "n3file: %s" % n3file
#print "n3url: %s" % n3url
self.graph = Graph()
n3file = Util.relpath(scriptdir + "/etc/sfs-extra.n3")
# print "loading n3file %s" % n3file
self.graph.load(n3file, format="n3")
self.roots = []
self.uriformatter = {}
self.decl = ""
self.namedlaws = {}
self.load_ebnf(scriptdir+"/etc/base.ebnf")
self.args = args
if self.LAGRUM in args:
productions = self.load_ebnf(scriptdir+"/etc/lagrum.ebnf")
for p in productions:
self.uriformatter[p] = self.sfs_format_uri
self.namedlaws.update(self.get_relations(RDFS.label))
self.roots.append("sfsrefs")
self.roots.append("sfsref")
if self.KORTLAGRUM in args:
# om vi inte redan laddat lagrum.ebnf måste vi göra det
# nu, eftersom kortlagrum.ebnf beror på produktioner som
# definerats där
if not self.LAGRUM in args:
self.load_ebnf(scriptdir+"/etc/lagrum.ebnf")
productions = self.load_ebnf(scriptdir+"/etc/kortlagrum.ebnf")
for p in productions:
self.uriformatter[p] = self.sfs_format_uri
DCT = Namespace("http://purl.org/dc/terms/")
d = self.get_relations(DCT['alternate'])
self.namedlaws.update(d)
lawlist = [x.encode(SP_CHARSET) for x in d.keys()]
# Make sure longer law abbreviations come before shorter
# ones (so that we don't mistake "3 § MBL" for "3 § MB"+"L")
lawlist.sort(cmp=lambda x,y:len(y)-len(x))
self.decl += "LawAbbreviation ::= ('%s')\n" % "'/'".join(lawlist)
self.roots.insert(0,"kortlagrumref")
if self.EGLAGSTIFTNING in args:
productions = self.load_ebnf(scriptdir+"/etc/eglag.ebnf")
for p in productions:
self.uriformatter[p] = self.eglag_format_uri
self.roots.append("eglagref")
if self.FORARBETEN in args:
productions = self.load_ebnf(scriptdir+"/etc/forarbeten.ebnf")
for p in productions:
self.uriformatter[p] = self.forarbete_format_uri
self.roots.append("forarbeteref")
if self.RATTSFALL in args:
productions = self.load_ebnf(scriptdir+"/etc/rattsfall.ebnf")
for p in productions:
self.uriformatter[p] = self.rattsfall_format_uri
self.roots.append("rattsfallref")
if self.EGRATTSFALL in args:
productions = self.load_ebnf(scriptdir+"/etc/egratt.ebnf")
for p in productions:
self.uriformatter[p] = self.egrattsfall_format_uri
self.roots.append("ecjcaseref")
self.decl += "root ::= (%s/plain)+\n" % "/".join(self.roots)
# pprint(productions)
# print self.decl.decode(SP_CHARSET,'ignore')
self.parser = Parser(self.decl, "root")
self.tagger = self.parser.buildTagger("root")
# print "tagger length: %d" % len(repr(self.tagger))
self.verbose = False
self.depth = 0
# SFS-specifik kod
self.currentlaw = None
self.currentchapter = None
self.currentsection = None
self.currentpiece = None
self.lastlaw = None
self.currentlynamedlaws = {}
def load_ebnf(self,file):
"""Laddar in produktionerna i den angivna filen i den
EBNF-deklaration som används, samt returnerar alla
*Ref och *RefId-produktioner"""
# print "%s: Loading %s" % (id(self), file)
f = open(file)
content = f.read()
self.decl += content
f.close()
return [x.group(1) for x in re.finditer(r'(\w+(Ref|RefID))\s*::=', content)]
def get_relations(self, predicate):
d = {}
for obj, subj in self.graph.subject_objects(predicate):
d[unicode(subj)] = unicode(obj)
return d
def parse(self, indata, baseuri="http://rinfo.lagrummet.se/publ/sfs/9999:999#K9P9S9P9",predicate=None):
if indata == "": return indata # this actually triggered a bug...
# h = hashlib.sha1()
# h.update(indata)
# print "Called with %r (%s) (%s)" % (indata, h.hexdigest(), self.verbose)
self.predicate = predicate
self.baseuri = baseuri
if baseuri:
m = self.re_urisegments.match(baseuri)
if m:
self.baseuri_attributes = {'baseuri':m.group(1),
'law':m.group(2),
'chapter':m.group(6),
'section':m.group(8),
'piece':m.group(10),
'item':m.group(12)}
else:
self.baseuri_attributes = {'baseuri':baseuri}
else:
self.baseuri_attributes = {}
# Det är svårt att få EBNF-grammatiken att känna igen
# godtyckliga ord som slutar på ett givet suffix (exv
# 'bokföringslagen' med suffixet 'lagen'). Därför förbehandlar
# vi indatasträngen och stoppar in ett '|'-tecken innan vissa
# suffix. Vi transformerar även 'Radio- och TV-lagen' till
# 'Radio-_och_TV-lagen'
#
# FIXME: Obviously, this shouldn't be done in a general class,
# but rather in a subclas or via proxy/adapter
# if we don't do the unicode conversion and pass
# BeautifulSoup.NavigableString, the later .encode call fails
# (since it's not a real unicode string)
fixedindata = unicode(indata)
# print "Before: %r" % type(fixedindata)
if self.LAGRUM in self.args:
fixedindata = self.re_escape_compound.sub(r'\1_\2_\3\4', fixedindata)
fixedindata = self.re_escape_named.sub(r'|\1', fixedindata)
# print "After: %r" % type(fixedindata)
# SimpleParse har inget stöd för unicodesträngar, så vi
# konverterar intdatat till en bytesträng. Tyvärr får jag inte
# det hela att funka med UTF8, så vi kör xml character
# references istället
if isinstance(fixedindata,unicode):
fixedindata = fixedindata.encode(SP_CHARSET,'xmlcharrefreplace')
# Parsea texten med TextTools.tag - inte det enklaste sättet
# att göra det, men om man gör enligt
# Simpleparse-dokumentationen byggs taggertabellen om för
# varje anrop till parse()
if self.verbose: print u"calling tag with '%s'" % (fixedindata.decode(SP_CHARSET))
# print "tagger length: %d" % len(repr(self.tagger))
taglist = tag(fixedindata, self.tagger,0,len(fixedindata))
result = []
root = NodeTree(taglist,fixedindata)
for part in root.nodes:
if part.tag != 'plain' and self.verbose:
sys.stdout.write(self.prettyprint(part))
if part.tag in self.roots:
self.clear_state()
# self.verbose = False
result.extend(self.formatter_dispatch(part))
else:
assert part.tag == 'plain',"Tag is %s" % part.tag
result.append(part.text)
# clear state
if self.currentlaw != None: self.lastlaw = self.currentlaw
self.currentlaw = None
if taglist[-1] != len(fixedindata):
log.error(u'Problem (%d:%d) with %r / %r' % (taglist[-1]-8,taglist[-1]+8,fixedindata,indata))
raise ParseError, "parsed %s chars of %s (...%s...)" % (taglist[-1], len(indata),
indata[(taglist[-1]-2):taglist[-1]+3])
# Normalisera resultatet, dvs konkatenera intilliggande
# textnoder, och ta bort ev '|'-tecken som vi stoppat in
# tidigare.
normres = []
for i in range(len(result)):
if not self.re_descape_named.search(result[i]):
node = result[i]
else:
if self.LAGRUM in self.args:
text = self.re_descape_named.sub(r'\1',result[i])
text = self.re_descape_compound.sub(r'\1 \2 \3\4', text)
if isinstance(result[i], Link):
# Eftersom Link-objekt är immutable måste vi skapa
# ett nytt och kopiera dess attribut
if hasattr(result[i],'predicate'):
node = LinkSubject(text, predicate=result[i].predicate,
uri=result[i].uri)
else:
node = Link(text,uri=result[i].uri)
else:
node = text
if (len(normres) > 0
and not isinstance(normres[-1],Link)
and not isinstance(node,Link)):
normres[-1] += node
else:
normres.append(node)
# and finally...
for i in range(len(normres)):
if isinstance(normres[i], Link):
# deal with these later
pass
else:
normres[i] = self.re_xmlcharref.sub(self.unescape_xmlcharref, normres[i])
return normres
def unescape_xmlcharref(self, m):
# print "Changing %r to a %r" % (m.group(0)[2:-1], unichr(int(m.group(0)[2:-1])))
return unichr(int(m.group(0)[2:-1]))
def find_attributes(self,parts,extra={}):
"""recurses through a parse tree and creates a dictionary of
attributes"""
d = {}
self.depth += 1
if self.verbose: print ". "*self.depth+"find_attributes: starting with %s"%d
if extra:
d.update(extra)
for part in parts:
current_part_tag = part.tag.lower()
if current_part_tag.endswith('refid'):
if ((current_part_tag == 'singlesectionrefid') or
(current_part_tag == 'lastsectionrefid')):
current_part_tag = 'sectionrefid'
d[current_part_tag[:-5]] = part.text.strip()
if self.verbose: print ". "*self.depth+"find_attributes: d is now %s" % d
if part.nodes:
d.update(self.find_attributes(part.nodes,d))
if self.verbose: print ". "*self.depth+"find_attributes: returning %s" % d
self.depth -= 1
if self.currentlaw and 'law' not in d : d['law'] = self.currentlaw
if self.currentchapter and 'chapter' not in d: d['chapter'] = self.currentchapter
if self.currentsection and 'section' not in d: d['section'] = self.currentsection
if self.currentpiece and 'piece' not in d : d['piece'] = self.currentpiece
return d
def find_node(self,root,nodetag):
"""Returns the first node in the tree that has a tag matching nodetag. The search is depth-first"""
if root.tag == nodetag: # base case
return root
else:
for node in root.nodes:
x = self.find_node(node,nodetag)
if x != None: return x
return None
def find_nodes(self,root,nodetag):
if root.tag == nodetag:
return [root]
else:
res = []
for node in root.nodes:
res.extend(self.find_nodes(node,nodetag))
return res
def flatten_tokentree(self,part,suffix):
"""returns a 'flattened' tokentree ie for the following tree and the suffix 'RefID'
foo->bar->BlahongaRefID
->baz->quux->Blahonga2RefID
->Blahonga3RefID
->Blahonga4RefID
this should return [BlahongaRefID, Blahonga2RefID, Blahonga3RefID, Blahonga4RefID]"""
l = []
if part.tag.endswith(suffix): l.append(part)
if not part.nodes: return l
for subpart in part.nodes:
l.extend(self.flatten_tokentree(subpart,suffix))
return l
def formatter_dispatch(self,part):
# print "Verbositiy: %r" % self.verbose
self.depth += 1
# Finns det en skräddarsydd formatterare?
if "format_"+part.tag in dir(self):
formatter = getattr(self,"format_"+part.tag)
if self.verbose: print (". "*self.depth)+ "formatter_dispatch: format_%s defined, calling it" % part.tag
res = formatter(part)
assert res != None, "Custom formatter for %s didn't return anything" % part.tag
else:
if self.verbose: print (". "*self.depth)+ "formatter_dispatch: no format_%s, using format_tokentree" % part.tag
res = self.format_tokentree(part)
if res == None: print (". "*self.depth)+ "something wrong with this:\n" + self.prettyprint(part)
self.depth -= 1
return res
def format_tokentree(self,part):
# This is the default formatter. It converts every token that
# ends with a RefID into a Link object. For grammar
# productions like SectionPieceRefs, which contain
# subproductions that also end in RefID, this is not a good
# function to use - use a custom formatter instead.
res = []
if self.verbose: print (". "*self.depth)+ "format_tokentree: called for %s" % part.tag
# this is like the bottom case, or something
if (not part.nodes) and (not part.tag.endswith("RefID")):
res.append(part.text)
else:
if part.tag.endswith("RefID"):
res.append(self.format_generic_link(part))
elif part.tag.endswith("Ref"):
res.append(self.format_generic_link(part))
else:
for subpart in part.nodes:
if self.verbose and part.tag == 'LawRef':
print (". "*self.depth) + "format_tokentree: part '%s' is a %s" % (subpart.text, subpart.tag)
res.extend(self.formatter_dispatch(subpart))
if self.verbose: print (". "*self.depth)+ "format_tokentree: returning '%s' for %s" % (res,part.tag)
return res
def prettyprint(self,root,indent=0):
res = u"%s'%s': '%s'\n" % (" "*indent,root.tag,re.sub(r'\s+', ' ',root.text))
if root.nodes != None:
for subpart in root.nodes:
res += self.prettyprint(subpart,indent+1)
return res
else: return u""
def format_generic_link(self,part,uriformatter=None):
try:
uri = self.uriformatter[part.tag](self.find_attributes([part]))
except KeyError:
if uriformatter:
uri = uriformatter(self.find_attributes([part]))
else:
uri = self.sfs_format_uri(self.find_attributes([part]))
except AttributeError:
# Normal error from eglag_format_uri
return part.text
except:
exc = sys.exc_info()
# If something else went wrong, just return the plaintext
log.warning("(unknown): Unable to format link for text %s (production %s)" % (part.text, part.tag))
return part.text
if self.verbose: print (". "*self.depth)+ "format_generic_link: uri is %s" % uri
if not uri:
# the formatting function decided not to return a URI for
# some reason (maybe it was a partial/relative reference
# without a proper base uri context
return part.text
elif self.predicate:
return LinkSubject(part.text, uri=uri, predicate=self.predicate)
else:
return Link(part.text, uri=uri)
# FIXME: unify this with format_generic_link
def format_custom_link(self, attributes, text, production):
try:
uri = self.uriformatter[production](attributes)
except KeyError:
uri = self.sfs_format_uri(attributes)
if not uri:
# the formatting function decided not to return a URI for
# some reason (maybe it was a partial/relative reference
# without a proper base uri context
return part.text
elif self.predicate:
return LinkSubject(text, uri=uri, predicate=self.predicate)
else:
return Link(text, uri=uri)
################################################################
# KOD FÖR LAGRUM
def clear_state(self):
self.currentlaw = None
self.currentchapter = None
self.currentsection = None
self.currentpiece = None
def normalize_sfsid(self,sfsid):
# sometimes '1736:0123 2' is given as '1736:0123 s. 2' or
# '1736:0123.2'. This fixes that.
sfsid = re.sub(r'(\d+:\d+)\.(\d)',r'\1 \2',sfsid)
#return sfsid.replace('s. ','').replace('s.','') # more advanced normalizations to come...
return sfsid
def normalize_lawname(self,lawname):
lawname=lawname.replace('|','').replace('_',' ').lower()
if lawname.endswith('s'):
lawname = lawname[:-1]
return lawname
def namedlaw_to_sfsid(self,text,normalize=True):
if normalize:
text = self.normalize_lawname(text)
nolaw = [
u'aktieslagen',
u'anordningen',
u'anordningen',
u'anslagen',
u'arbetsordningen',
u'associationsformen',
u'avfallsslagen',
u'avslagen',
u'avvittringsutslagen',
u'bergslagen',
u'beskattningsunderlagen',
u'bolagen',
u'bolagsordningen',
u'bolagsordningen',
u'dagordningen',
u'djurslagen',
u'dotterbolagen',
u'emballagen',
u'energislagen',
u'ersättningsformen',
u'ersättningsslagen',
u'examensordningen',
u'finansbolagen',
u'finansieringsformen',
u'fissionsvederlagen',
u'flygbolagen',
u'fondbolagen',
u'förbundsordningen',
u'föreslagen',
u'företrädesordningen',
u'förhandlingsordningen',
u'förlagen',
u'förmånsrättsordningen',
u'förmögenhetsordningen',
u'förordningen',
u'förslagen',
u'försäkringsaktiebolagen',
u'försäkringsbolagen',
u'gravanordningen',
u'grundlagen',
u'handelsplattformen',
u'handläggningsordningen',
u'inkomstslagen',
u'inköpssamordningen',
u'kapitalunderlagen',
u'klockslagen',
u'kopplingsanordningen',
u'låneformen',
u'mervärdesskatteordningen',
u'nummerordningen',
u'omslagen',
u'ordalagen',
u'pensionsordningen',
u'renhållningsordningen',
u'representationsreformen',
u'rättegångordningen',
u'rättegångsordningen',
u'rättsordningen',
u'samordningen',
u'samordningen',
u'skatteordningen',
u'skatteslagen',
u'skatteunderlagen',
u'skolformen',
u'skyddsanordningen',
u'slagen',
u'solvärmeanordningen',
u'storslagen',
u'studieformen',
u'stödformen',
u'stödordningen',
u'stödordningen',
u'säkerhetsanordningen',
u'talarordningen',
u'tillslagen',
u'tivolianordningen',
u'trafikslagen',
u'transportanordningen',
u'transportslagen',
u'trädslagen',
u'turordningen',
u'underlagen',
u'uniformen',
u'uppställningsformen',
u'utvecklingsbolagen',
u'varuslagen',
u'verksamhetsformen',
u'vevanordningen',
u'vårdformen',
u'ägoanordningen',
u'ägoslagen',
u'ärendeslagen',
u'åtgärdsförslagen',
]
if text in nolaw:
return None
if self.currentlynamedlaws.has_key(text):
return self.currentlynamedlaws[text]
elif self.namedlaws.has_key(text):
return self.namedlaws[text]
else:
if self.verbose:
# print "(unknown): I don't know the ID of named law [%s]" % text
log.warning("(unknown): I don't know the ID of named law [%s]" % text)
return None
def sfs_format_uri(self,attributes):
piecemappings = {u'första' :'1',
u'andra' :'2',
u'tredje' :'3',
u'fjärde' :'4',
u'femte' :'5',
u'sjätte' :'6',
u'sjunde' :'7',
u'åttonde':'8',
u'nionde' :'9'}
keymapping = {'lawref' :'L',
'chapter' :'K',
'section' :'P',
'piece' :'S',
'item' :'N',
'itemnumeric':'N',
'element' :'O',
'sentence':'M', # is this ever used?
}
attributeorder = ['law', 'lawref', 'chapter', 'section', 'element', 'piece', 'item', 'itemnumeric','sentence']
if 'law' in attributes:
if attributes['law'].startswith('http://'):
res = ''
else:
res = 'http://rinfo.lagrummet.se/publ/sfs/'
else:
if 'baseuri' in self.baseuri_attributes:
res = self.baseuri_attributes['baseuri']
else:
res = ''
resolvetobase = True
addfragment = False
justincase = None
for key in attributeorder:
if attributes.has_key(key):
resolvetobase = False
val = attributes[key]
elif (resolvetobase and self.baseuri_attributes.has_key(key)):
val = self.baseuri_attributes[key]
else:
val = None
if val:
if addfragment:
res += '#'
addfragment = False
if (key in ['piece', 'itemnumeric', 'sentence'] and val in piecemappings):
res += '%s%s' % (keymapping[key],piecemappings[val.lower()])
else:
if key == 'law':
val = self.normalize_sfsid(val)
val = val.replace(" ", "_")
res += val
addfragment = True
else:
if justincase:
res += justincase
justincase = None
val = val.replace(" ", "")
val = val.replace("\n", "")
val = val.replace("\r", "")
res += '%s%s' % (keymapping[key],val)
else:
if key == 'piece':
justincase = "S1"
return res
def format_ChapterSectionRefs(self,root):
assert(root.tag == 'ChapterSectionRefs')
assert(len(root.nodes) == 3) # ChapterRef, wc, SectionRefs
part = root.nodes[0]
self.currentchapter = part.nodes[0].text.strip()
if self.currentlaw:
res = [self.format_custom_link({'law':self.currentlaw,
'chapter':self.currentchapter},
part.text,
part.tag)]
else:
res = [self.format_custom_link({'chapter':self.currentchapter},
part.text,
part.tag)]
res.extend(self.formatter_dispatch(root.nodes[1]))
res.extend(self.formatter_dispatch(root.nodes[2]))
self.currentchapter = None
return res
def format_ChapterSectionPieceRefs(self,root):
assert(root.nodes[0].nodes[0].tag == 'ChapterRefID')
self.currentchapter = root.nodes[0].nodes[0].text.strip()
res = []
for node in root.nodes:
res.extend(self.formatter_dispatch(node))
return res
def format_LastSectionRef(self, root):
# the last section ref is a bit different, since we want the
# ending double section mark to be part of the link text
assert(root.tag == 'LastSectionRef')
assert(len(root.nodes) == 3) # LastSectionRefID, wc, DoubleSectionMark
sectionrefid = root.nodes[0]
sectionid = sectionrefid.text
return [self.format_generic_link(root)]
def format_SectionPieceRefs(self, root):
assert(root.tag == 'SectionPieceRefs')
self.currentsection = root.nodes[0].nodes[0].text.strip()
res = [self.format_custom_link(self.find_attributes([root.nodes[2]]),
"%s %s" % (root.nodes[0].text, root.nodes[2].text),
root.tag)]
for node in root.nodes[3:]:
res.extend(self.formatter_dispatch(node))
self.currentsection = None
return res
def format_SectionPieceItemRefs(self,root):
assert(root.tag == 'SectionPieceItemRefs')
self.currentsection = root.nodes[0].nodes[0].text.strip()
self.currentpiece = root.nodes[2].nodes[0].text.strip()
res = [self.format_custom_link(self.find_attributes([root.nodes[2]]),
"%s %s" % (root.nodes[0].text, root.nodes[2].text),
root.tag)]
for node in root.nodes[3:]:
res.extend(self.formatter_dispatch(node))
self.currentsection = None
self.currentpiece = None
return res
# This is a special case for things like '17-29 och 32 §§ i lagen
# (2004:575)', which picks out the LawRefID first and stores it in
# .currentlaw, so that find_attributes finds it
# automagically. Although now it seems to be branching out and be
# all things to all people.
def format_ExternalRefs(self,root):
assert(root.tag == 'ExternalRefs')
# print "DEBUG: start of format_ExternalRefs; self.currentlaw is %s" % self.currentlaw
lawrefid_node = self.find_node(root,'LawRefID')
if lawrefid_node == None:
# Ok, no explicit LawRefID found, lets see if this is a named law that we have the ID for
# namedlaw_node = self.find_node(root, 'NamedLawExternalLawRef')
namedlaw_node = self.find_node(root, 'NamedLaw')
if namedlaw_node == None:
# As a last chance, this might be a reference back to a previously mentioned law ("...enligt 4 § samma lag")
samelaw_node = self.find_node(root, 'SameLaw')
assert(samelaw_node != None)
if self.lastlaw == None:
log.warning(u"(unknown): found reference to \"{samma,nämnda} {lag,förordning}\", but self.lastlaw is not set")
self.currentlaw = self.lastlaw
else:
# the NamedLaw case
self.currentlaw = self.namedlaw_to_sfsid(namedlaw_node.text)
if self.currentlaw == None:
# unknow law name - in this case it's better to
# bail out rather than resolving chapter/paragraph
# references relative to baseuri (which is almost
# certainly wrong)
return [root.text]
else:
self.currentlaw = lawrefid_node.text
if self.find_node(root,'NamedLaw'):
namedlaw = self.normalize_lawname(self.find_node(root,'NamedLaw').text)
# print "remember that %s is %s!" % (namedlaw, self.currentlaw)
self.currentlynamedlaws[namedlaw] = self.currentlaw
#print "DEBUG: middle of format_ExternalRefs; self.currentlaw is %s" % self.currentlaw
if self.lastlaw is None:
#print "DEBUG: format_ExternalRefs: setting self.lastlaw to %s" % self.currentlaw
self.lastlaw = self.currentlaw
# if the node tree only contains a single reference, it looks
# better if the entire expression, not just the
# chapter/section part, is linked. But not if it's a
# "anonymous" law ('1 § i lagen (1234:234) om blahonga')
if (len(self.find_nodes(root,'GenericRefs')) == 1 and
len(self.find_nodes(root,'SectionRefID')) == 1 and
len(self.find_nodes(root,'AnonymousExternalLaw')) == 0):
res = [self.format_generic_link(root)]
else:
res = self.format_tokentree(root)
return res
def format_SectionItemRefs(self,root):
assert(root.nodes[0].nodes[0].tag == 'SectionRefID')
self.currentsection = root.nodes[0].nodes[0].text.strip()
#res = self.formatter_dispatch(root.nodes[0]) # was formatter_dispatch(self.root)
res = self.format_tokentree(root)
self.currentsection = None
return res
def format_PieceItemRefs(self,root):
self.currentpiece = root.nodes[0].nodes[0].text.strip()
res = [self.format_custom_link(self.find_attributes([root.nodes[2].nodes[0]]),
"%s %s" % (root.nodes[0].text, root.nodes[2].nodes[0].text),
root.tag)]
for node in root.nodes[2].nodes[1:]:
res.extend(self.formatter_dispatch(node))
self.currentpiece = None
return res
def format_ChapterSectionRef(self,root):
assert(root.nodes[0].nodes[0].tag == 'ChapterRefID')
self.currentchapter = root.nodes[0].nodes[0].text.strip()
return [self.format_generic_link(root)]
def format_AlternateChapterSectionRefs(self,root):
assert(root.nodes[0].nodes[0].tag == 'ChapterRefID')
self.currentchapter = root.nodes[0].nodes[0].text.strip()
# print "Self.currentchapter is now %s" % self.currentchapter
res = self.format_tokentree(root)
self.currentchapter = None
return res
def format_ExternalLaw(self,root):
self.currentchapter = None
return self.formatter_dispatch(root.nodes[0])
def format_ChangeRef(self,root):
id = self.find_node(root,'LawRefID').data
return [self.format_custom_link({'lawref':id},
root.text,
root.tag)]
def format_SFSNr(self,root):
if self.baseuri == None:
sfsid = self.find_node(root,'LawRefID').data
self.baseuri_attributes = {'baseuri':'http://rinfo.lagrummet.se/publ/sfs/'+sfsid+'#'}
return self.format_tokentree(root)
def format_NamedExternalLawRef(self,root):
resetcurrentlaw = False
#print "format_NamedExternalLawRef: self.currentlaw is %r" % self.currentlaw
if self.currentlaw == None:
resetcurrentlaw = True
lawrefid_node = self.find_node(root,'LawRefID')
if lawrefid_node == None:
self.currentlaw = self.namedlaw_to_sfsid(root.text)
else:
self.currentlaw = lawrefid_node.text
namedlaw = self.normalize_lawname(self.find_node(root,'NamedLaw').text)
# print "remember that %s is %s!" % (namedlaw, self.currentlaw)
self.currentlynamedlaws[namedlaw] = self.currentlaw
#print "format_NamedExternalLawRef: self.currentlaw is now %r" % self.currentlaw
#print "format_NamedExternalLawRef: self.baseuri is %r" % self.baseuri
if self.currentlaw == None: # if we can't find a ID for this law, better not <link> it
res = [root.text]
else:
res = [self.format_generic_link(root)]
#print "format_NamedExternalLawRef: self.baseuri is %r" % self.baseuri
if self.baseuri == None and self.currentlaw != None:
#print "format_NamedExternalLawRef: setting baseuri_attributes"
# use this as the new baseuri_attributes
m = self.re_urisegments.match(self.currentlaw)
if m:
self.baseuri_attributes = {'baseuri':m.group(1),
'law':m.group(2),
'chapter':m.group(6),
'section':m.group(8),
'piece':m.group(10),
'item':m.group(12)}
else:
self.baseuri_attributes = {'baseuri':'http://rinfo.lagrummet.se/publ/sfs/'+self.currentlaw+'#'}
if resetcurrentlaw:
if self.currentlaw != None: self.lastlaw = self.currentlaw
self.currentlaw = None
return res
################################################################
# KOD FÖR KORTLAGRUM
def format_AbbrevLawNormalRef(self,root):
lawabbr_node = self.find_node(root,'LawAbbreviation')
self.currentlaw = self.namedlaw_to_sfsid(lawabbr_node.text,normalize=False)
res = [self.format_generic_link(root)]
if self.currentlaw != None: self.lastlaw = self.currentlaw
self.currentlaw = None
return res
def format_AbbrevLawShortRef(self,root):
assert(root.nodes[0].tag == 'LawAbbreviation')
assert(root.nodes[2].tag == 'ShortChapterSectionRef')
self.currentlaw = self.namedlaw_to_sfsid(root.nodes[0].text,normalize=False)
shortsection_node = root.nodes[2]
assert(shortsection_node.nodes[0].tag == 'ShortChapterRefID')
assert(shortsection_node.nodes[2].tag == 'ShortSectionRefID')
self.currentchapter = shortsection_node.nodes[0].text
self.currentsection = shortsection_node.nodes[2].text
res = [self.format_generic_link(root)]
self.currentchapter = None
self.currentsection = None
self.currentlaw = None
return res
################################################################
# KOD FÖR FORARBETEN
def forarbete_format_uri(self,attributes):
# res = self.baseuri_attributes['baseuri']
res = 'http://rinfo.lagrummet.se/'
resolvetobase = True
addfragment = False
for key,val in attributes.items():
if key == 'prop':
res += "publ/prop/%s" % val
elif key == 'bet':
res += "ext/bet/%s" % val
elif key == 'skrivelse':
res += "ext/rskr/%s" % val
elif key == 'celex':
if len(val) == 8: # incorrectly formatted, uses YY instead of YYYY
val = val[0]+'19'+val[1:]
res += "ext/celex/%s" % val
if 'sidnr' in attributes:
res += "#s%s" % attributes['sidnr']
return res
def format_ChapterSectionRef(self,root):
assert(root.nodes[0].nodes[0].tag == 'ChapterRefID')
self.currentchapter = root.nodes[0].nodes[0].text.strip()
return [self.format_generic_link(root)]
################################################################
# KOD FÖR EGLAGSTIFTNING
def eglag_format_uri(self,attributes):
res = 'http://rinfo.lagrummet.se/ext/celex/'
if not 'akttyp' in attributes:
if 'forordning' in attributes:
attributes['akttyp'] = u'förordning';
elif 'direktiv' in attributes:
attributes['akttyp'] = u'direktiv';
if 'akttyp' not in attributes:
raise AttributeError("Akttyp saknas")
# Om hur CELEX-nummer konstrueras
# https://www.infotorg.sema.se/infotorg/itweb/handbook/rb/hlp_celn.htm
# https://www.infotorg.sema.se/infotorg/itweb/handbook/rb/hlp_celf.htm
# Om hur länkning till EURLEX ska se ut:
# http://eur-lex.europa.eu/sv/tools/help_syntax.htm
# Absolut URI?
if 'ar' in attributes and 'lopnummer' in attributes:
sektor = '3'
rattslig_form = {u'direktiv':'L',
u'förordning':'R'}
if len(attributes['ar']) == 2:
attributes['ar'] = '19'+attributes['ar']
res += "%s%s%s%04d" % (sektor,attributes['ar'],
rattslig_form[attributes['akttyp']],
int(attributes['lopnummer']))
else:
if not self.baseuri_attributes['baseuri'].startswith(res):
# FIXME: should we warn about this?
# print "Relative reference, but base context %s is not a celex context" % self.baseuri_attributes['baseuri']
return None
if 'artikel' in attributes:
res += "#%s" % attributes['artikel']
if 'underartikel' in attributes:
res += ".%s" % attributes['underartikel']
return res
################################################################
# KOD FÖR RATTSFALL
def rattsfall_format_uri(self,attributes):
# Listan härledd från containers.n3/rattsfallsforteckningar.n3 i
# rinfoprojektets källkod - en ambitiösare lösning vore att läsa
# in de faktiska N3-filerna i en rdflib-graf.
containerid = {u'NJA': '/publ/rattsfall/nja/',
u'RH': '/publ/rattsfall/rh/',
u'MÖD': '/publ/rattsfall/mod/',
u'RÅ': '/publ/rattsfall/ra/',
u'HFD': '/publ/rattsfall/hfd/',
u'RK': '/publ/rattsfall/rk/',
u'MIG': '/publ/rattsfall/mig/',
u'AD': '/publ/rattsfall/ad/',
u'MD': '/publ/rattsfall/md/',
u'FÖD': '/publ/rattsfall/fod/'}
# res = self.baseuri_attributes['baseuri']
if 'nja' in attributes:
attributes['domstol'] = attributes['nja']
assert 'domstol' in attributes, "No court provided"
assert attributes['domstol'] in containerid, "%s is an unknown court" % attributes['domstol']
res = "http://rinfo.lagrummet.se"+containerid[attributes['domstol']]
if 'lopnr' in attributes and ":" in attributes['lopnr']:
(attributes['ar'], attributes['lopnr']) = lopnr.split(":", 1)
if attributes['domstol'] == u'NJA':
# FIXME: URIs should be based on publikationsordinal, not
# pagenumber (which this in effect is) - but this requires
# a big lookup table/database/graph with
# pagenumber-to-ordinal-mappings
res += '%ss%s' % (attributes['ar'], attributes['sidnr'])
else:
res += '%s:%s' % (attributes['ar'], attributes['lopnr'])
return res
################################################################
# KOD FÖR EGRÄTTSFALL
def egrattsfall_format_uri(self,attributes):
descriptormap = {'C':'J', # Judgment of the Court
'T':'A', # Judgment of the Court of First Instance
'F':'W', # Judgement of the Civil Service Tribunal
}
# FIXME: Change this before the year 2054 (as ECJ will
# hopefully have fixed their case numbering by then)
if len(attributes['year']) == 2:
if int(attributes['year']) < 54:
year = "20"+attributes['year']
else:
year = "19"+attributes['year']
else:
year = attributes['year']
serial = '%04d' % int(attributes['serial'])
descriptor = descriptormap[attributes['decision']]
uri = "http://lagen.nu/ext/celex/6%s%s%s" % (year, descriptor, serial)
return uri