Merge branch 'master' into 5.0.0-dev

* master: (38 commits)
  hardening + warning for serializing invalid numericals (e.g., '"xy.z"^^xsd:float')
  make sure inf/Infinity and nan from numerical types are serialized as INF/NaN
  except inf/nan from plain serialization
  more tests for float and decimal inf/nan serialization (ensures datatypes)
  Only doubles can be infinity or NaN
  Serialisation of nan is "NaN"
  Remove test on serialisation (order is not fixed)
  Add tests requested by @joernhees
  change xxx to issue number (655)
  Test serialisation of infinite values
  Serialisation of infinity is "INF"
  Update requirements files for html5lib version
  Support newer transport_encoding parameter in html5lib parser call
  Set identifier of the parser sink to target graph. Fixes #432.
  Fix TurtleSerializer.isValidList failure on falsy values
  only run flake8 in latest py2 and py3 version
  pinning html5lib to 1.0b8 for now
  removed now unnecessary version restriction from SPARQLWrapper
  updated fuseki used in travis tests to 2.4.0
  travis updates setuptools and pip before installing
  ...

.travis.fuseki_install_optional.sh

@@ -2,11 +2,11 @@
 
 set -v
 
-uri="http://archive.eu.apache.org/dist/jena/binaries/jena-fuseki1-1.1.2-distribution.tar.gz"
+uri="http://archive.apache.org/dist/jena/binaries/apache-jena-fuseki-2.4.0.tar.gz"
 
 if wget "$uri" &&
-       tar -zxf jena-fuseki*-distribution.tar.gz &&
-       mv jena-fuseki*/ fuseki &&
+       tar -zxf *jena*fuseki*.tar.gz &&
+       mv *jena*fuseki*/ fuseki &&
        cd fuseki ; then
     # normal SPARQLStore & Dataset tests:
     bash fuseki-server --port 3030 --debug --update --mem /db &>fuseki.log &

.travis.yml

@@ -1,10 +1,13 @@
 # http://travis-ci.org/#!/RDFLib/rdflib
-language: python
 sudo: false
+language: python
 branches:
   only:
+    # only build master and release branches (merge request are built anyhow)
     - master
     - /^\d+\.\d+\.\d+(-.*)?$/
+git:
+  depth: 3
 
 python:
     # - 2.5
@@ -17,10 +20,11 @@ python:
     # - "pypy"
 
 before_install:
+    - pip install -U setuptools pip  # seems travis comes with a too old setuptools for html5lib
     - bash .travis.fuseki_install_optional.sh
 
 install:
-    - if [[ ${TRAVIS_PYTHON_VERSION%%.*} == '2' ]]; then pip install --default-timeout 60 -r requirements.py2.txt;  pip install --default-timeout 60 "html5lib";fi
+    - if [[ ${TRAVIS_PYTHON_VERSION%%.*} == '2' ]]; then pip install --default-timeout 60 -r requirements.py2.txt; fi
     - if [[ ${TRAVIS_PYTHON_VERSION%%.*} == '3' ]]; then pip install --default-timeout 60 -r requirements.py3.txt; fi
     - if [[ $TRAVIS_PYTHON_VERSION == '2.6' ]]; then pip install --default-timeout 60 ordereddict 'networkx<1.10' ; fi
     - if [[ $TRAVIS_PYTHON_VERSION != '2.6' ]]; then pip install --default-timeout 60 networkx ; fi
@@ -30,7 +34,7 @@ install:
     - python setup.py install
 
 before_script:
-    - if [[ ${TRAVIS_PYTHON_VERSION%%.*} == '2' ]]; then flake8 --exclude=pyRdfa,extras,host,transform,rdfs,sparql,results,pyMicrodata --exit-zero rdflib; fi
+    - if [[ $TRAVIS_PYTHON_VERSION == '2.7' ]] || [[ $TRAVIS_PYTHON_VERSION == '3.5' ]]; then flake8 --exclude=pyRdfa,extras,host,transform,rdfs,sparql,results,pyMicrodata --exit-zero rdflib; fi
 
 script:
     # Must run the tests in build/src so python3 doesn't get confused and run

README.md

@@ -1,5 +1,12 @@
 RDFLib
 ======
+[![Build Status](https://travis-ci.org/RDFLib/rdflib.png?branch=master)](https://travis-ci.org/RDFLib/rdflib)
+[![Coveralls branch](https://img.shields.io/coveralls/RDFLib/rdflib/master.svg)](https://coveralls.io/r/RDFLib/rdflib?branch=master)
+[![GitHub stars](https://img.shields.io/github/stars/RDFLib/rdflib.svg)](https://github.com/RDFLib/rdflib/stargazers)
+[![PyPI](https://img.shields.io/pypi/v/rdflib.svg)](https://pypi.python.org/pypi/rdflib)
+[![PyPI](https://img.shields.io/pypi/dm/rdflib.svg)](https://pypi.python.org/pypi/rdflib)
+[![PyPI](https://img.shields.io/pypi/pyversions/rdflib.svg)](https://pypi.python.org/pypi/rdflib)
+
 
 RDFLib is a Python library for working with RDF, a simple yet
 powerful language for representing information as graphs.
@@ -30,7 +37,7 @@ g=rdflib.Graph()
 g.load('http://dbpedia.org/resource/Semantic_Web')
 
 for s,p,o in g:
-  print s,p,o
+    print s,p,o
 ```
 
 The components of the triples are URIs (resources) or Literals
@@ -54,7 +61,7 @@ dbpedia=Namespace('http://dbpedia.org/ontology/')
 abstracts=list(x for x in g.objects(semweb, dbpedia['abstract']) if x.language=='en')
 ```
 
-See also *./examples*
+See also [./examples](./examples)
 
 
 Features
@@ -85,11 +92,6 @@ More information is available on the project webpage:
 
 https://github.com/RDFLib/rdflib/
 
-Continuous integration status details available from travis.ci, test coverage from coveralls:
-
-[![Build Status](https://travis-ci.org/RDFLib/rdflib.png?branch=master)](https://travis-ci.org/RDFLib/rdflib)
-[![Coverage Status](https://coveralls.io/repos/RDFLib/rdflib/badge.png?branch=master)](https://coveralls.io/r/RDFLib/rdflib?branch=master)
-
 The documentation can be built by doing::
 
     $ python setup.py build_sphinx

rdflib/__init__.py

@@ -73,10 +73,17 @@
 del sys
 
 import logging
-import __main__
-if not hasattr(__main__, '__file__'):
-    # show log messages in interactive mode
-    logging.basicConfig(level=logging.INFO)
+try:
+    import __main__
+    if not hasattr(__main__, '__file__'):
+        # show log messages in interactive mode
+        logging.basicConfig(level=logging.INFO)
+except ImportError:
+    #Main already imported from elsewhere
+    import warnings
+    warnings.warn('__main__ already imported', ImportWarning)
+    del warnings
+
 logger = logging.getLogger(__name__)
 logger.info("RDFLib Version: %s" % __version__)
 

rdflib/collection.py

@@ -129,7 +129,7 @@ def __setitem__(self, key, value):
         """TODO"""
         c = self._get_container(key)
         if c:
-            self.graph.add((c, RDF.first, value))
+            self.graph.set((c, RDF.first, value))
         else:
             raise IndexError(key)
 

rdflib/compare.py

@@ -143,11 +143,15 @@ def wrapped_f(*args, **kwargs):
 class IsomorphicGraph(ConjunctiveGraph):
     """An implementation of the RGDA1 graph digest algorithm.
 
-    An implementation of RGDA1 (publication forthcoming),
+    An implementation of RGDA1 (publication below),
     a combination of Sayers & Karp's graph digest algorithm using
     sum and SHA-256 <http://www.hpl.hp.com/techreports/2003/HPL-2003-235R1.pdf>
     and traces <http://pallini.di.uniroma1.it>, an average case
     polynomial time algorithm for graph canonicalization.
+    
+    McCusker, J. P. (2015). WebSig: A Digital Signature Framework for the Web.
+    Rensselaer Polytechnic Institute, Troy, NY.
+    http://gradworks.umi.com/3727015.pdf
     """
 
     def __init__(self, **kwargs):

rdflib/graph.py

@@ -925,7 +925,7 @@ def serialize(self, destination=None, format="xml",
         string. Format defaults to xml (AKA rdf/xml).
 
         Format support can be extended with plugins,
-        but 'xml', 'n3', 'turtle', 'nt', 'pretty-xml', trix' are built in.
+        but 'xml', 'n3', 'turtle', 'nt', 'pretty-xml', 'trix', 'trig' and 'nquads' are built in.
         """
         serializer = plugin.get(format, Serializer)(self)
         if destination is None:
@@ -1079,10 +1079,12 @@ def query(self, query_object, processor='sparql',
             query_object, initBindings, initNs, **kwargs))
 
     def update(self, update_object, processor='sparql',
-              initNs={}, initBindings={},
+              initNs=None, initBindings=None,
               use_store_provided=True, **kwargs):
-        """
-        """
+        """Update this graph with the given update query."""
+        initBindings = initBindings or {}
+        initNs = initNs or dict(self.namespaces())
+
         if hasattr(self.store, "update") and use_store_provided:
             try:
                 return self.store.update(

rdflib/plugins/parsers/ntriples.py

@@ -207,7 +207,7 @@ def eat(self, pattern):
         if not m:  # @@ Why can't we get the original pattern?
             # print(dir(pattern))
             # print repr(self.line), type(self.line)
-            raise ParseError("Failed to eat %s at %s" % (pattern, self.line))
+            raise ParseError("Failed to eat %s at %s" % (pattern.pattern, self.line))
         self.line = self.line[m.end():]
         return m
 

rdflib/plugins/parsers/pyRdfa/__init__.py

@@ -609,7 +609,10 @@ def copyErrors(tog, options) :
 					if self.charset :
 						# This means the HTTP header has provided a charset, or the
 						# file is a local file when we suppose it to be a utf-8
-						dom = parser.parse(input, encoding=self.charset)
+						try:
+							dom = parser.parse(input, encoding=self.charset)
+						except TypeError:
+							dom = parser.parse(input, transport_encoding=self.charset)
 					else :
 						# No charset set. The HTMLLib parser tries to sniff into the
 						# the file to find a meta header for the charset; if that

rdflib/plugins/parsers/trig.py

@@ -146,7 +146,7 @@ def parse(self, source, graph, encoding="utf-8"):
         # we're currently being handed a Graph, not a ConjunctiveGraph
         assert graph.store.context_aware, "TriG Parser needs a context-aware store!"
 
-        conj_graph = ConjunctiveGraph(store=graph.store)
+        conj_graph = ConjunctiveGraph(store=graph.store, identifier=graph.identifier)
         conj_graph.default_context = graph  # TODO: CG __init__ should have a
                                             # default_context arg
          # TODO: update N3Processor so that it can use conj_graph as the sink

rdflib/plugins/serializers/turtle.py

@@ -345,7 +345,7 @@ def isValidList(self, l):
         Checks if l is a valid RDF list, i.e. no nodes have other properties.
         """
         try:
-            if not self.store.value(l, RDF.first):
+            if self.store.value(l, RDF.first) is None:
                 return False
         except:
             return False

rdflib/plugins/sparql/aggregates.py

@@ -13,10 +13,14 @@
 """
 
 
-def _eval_rows(expr, group):
+def _eval_rows(expr, group, distinct):
+    seen = set()
     for row in group:
         try:
-            yield _eval(expr, row)
+            val = _eval(expr, row)
+            if not distinct or not val in seen:
+                yield val
+                seen.add(val)
         except:
             pass
 
@@ -25,9 +29,8 @@ def agg_Sum(a, group, bindings):
     c = 0
 
     dt = None
-    for x in group:
+    for e in _eval_rows(a.vars, group, a.distinct):
         try:
-            e = _eval(a.vars, x)
             n = numeric(e)
             if dt == None:
                 dt = e.datatype
@@ -51,15 +54,15 @@ def agg_Sum(a, group, bindings):
 def agg_Min(a, group, bindings):
     m = None
 
-    for x in group:
+    for v in _eval_rows(a.vars, group, None): # DISTINCT makes no difference for MIN
         try:
-            v = numeric(_eval(a.vars, x))
+            v = numeric(v)
             if m is None:
                 m = v
             else:
                 m = num_min(v, m)
         except:
-            return  # error in aggregate => no binding
+            continue # try other values
 
     if m is not None:
         bindings[a.res] = Literal(m)
@@ -68,9 +71,9 @@ def agg_Min(a, group, bindings):
 def agg_Max(a, group, bindings):
     m = None
 
-    for x in group:
+    for v in _eval_rows(a.vars, group, None): # DISTINCT makes no difference for MAX
         try:
-            v = numeric(_eval(a.vars, x))
+            v = numeric(v)
             if m is None:
                 m = v
             else:
@@ -83,46 +86,44 @@ def agg_Max(a, group, bindings):
 
 
 def agg_Count(a, group, bindings):
-
-    c = 0
-    for x in group:
-        try:
-            if a.vars != '*':
-                val = _eval(a.vars, x)
-                if isinstance(val, NotBoundError):
-                    continue
+    if a.vars == '*':
+        c = len(group)
+    else:
+        c = 0
+        for e in _eval_rows(a.vars, group, a.distinct):
             c += 1
-        except:
-            return  # error in aggregate => no binding
-            # pass  # simply dont count
 
     bindings[a.res] = Literal(c)
 
 
 def agg_Sample(a, group, bindings):
     for ctx in group:
-        val = _eval(a.vars, ctx)
-        if not isinstance(val, NotBoundError):
-            bindings[a.res] = val
+        try:
+            bindings[a.res] = _eval(a.vars, ctx)
             break
+        except NotBoundError:
+            pass
 
 
 def agg_GroupConcat(a, group, bindings):
 
     sep = a.separator or " "
+    if a.distinct:
+        agg = lambda x: x
+    else:
+        add = set
 
     bindings[a.res] = Literal(
-        sep.join(unicode(x) for x in _eval_rows(a.vars, group)))
+        sep.join(unicode(x) for x in _eval_rows(a.vars, group, a.distinct)))
 
 
 def agg_Avg(a, group, bindings):
 
     c = 0
     s = 0
     dt = None
-    for x in group:
+    for e in _eval_rows(a.vars, group, a.distinct):
         try:
-            e = _eval(a.vars, x)
             n = numeric(e)
             if dt == None:
                 dt = e.datatype

rdflib/plugins/sparql/evaluate.py

@@ -282,7 +282,7 @@ def evalGroup(ctx, group):
     else:
         res = collections.defaultdict(list)
         for c in p:
-            k = tuple(_eval(e, c) for e in group.expr)
+            k = tuple(_eval(e, c, False) for e in group.expr)
             res[k].append(c)
         return res
 

rdflib/plugins/sparql/evalutils.py

@@ -61,7 +61,7 @@ def _ebv(expr, ctx):
     return False
 
 
-def _eval(expr, ctx):
+def _eval(expr, ctx, raise_not_bound_error=True):
     if isinstance(expr, (Literal, URIRef)):
         return expr
     if isinstance(expr, Expr):
@@ -70,7 +70,10 @@ def _eval(expr, ctx):
         try:
             return ctx[expr]
         except KeyError:
-            return NotBoundError("Variable %s is not bound" % expr)
+            if raise_not_bound_error:
+                raise NotBoundError("Variable %s is not bound" % expr)
+            else:
+                return None
     elif isinstance(expr, CompValue):
         raise Exception(
             "Weird - _eval got a CompValue without evalfn! %r" % expr)