def test_convert_simple_value_boolean_query_to_and_boolean_queries():
parse_tree = \
parser.SimpleQuery(
parser.SpiresKeywordQuery(
parser.InspireKeyword('author'),
parser.Value(
parser.SimpleValueBooleanQuery(
parser.SimpleValue('foo'),
parser.And(),
parser.SimpleValueBooleanQuery(
parser.SimpleValue('bar'),
parser.Or(),
parser.SimpleValueNegation(parser.SimpleValue('foobar'))
)
)
)
)
)
expected_parse_tree = \
AndOp(
KeywordOp(Keyword('author'), Value('foo')),
OrOp(
KeywordOp(Keyword('author'), Value('bar')),
NotOp(KeywordOp(Keyword('author'), Value('foobar')))
)
)
restructuring_visitor = RestructuringVisitor()
parse_tree = parse_tree.accept(restructuring_visitor)
assert parse_tree == expected_parse_tree
def test_foo_bar():
query_str = 'find j Nucl.Phys. and not vol A531 and a ellis'
print("Parsing: " + query_str)
stateful_parser = StatefulParser()
restructuring_visitor = RestructuringVisitor()
_, parse_tree = stateful_parser.parse(query_str, parser.Query)
parse_tree = parse_tree.accept(restructuring_visitor)
expected_parse_tree = AndOp(
KeywordOp(Keyword('journal'), Value('Nucl.Phys.')),
KeywordOp(Keyword('author'), Value('ellis')))
assert parse_tree == expected_parse_tree
def visit_boolean_query(self, node):
"""Convert BooleanRule into AndOp or OrOp nodes."""
left = node.left.accept(self)
right = node.right.accept(self)
is_journal_keyword_op = isinstance(
left, KeywordOp) and left.left == Keyword('journal')
if is_journal_keyword_op:
journal_and_volume_conjunction = _restructure_if_volume_follows_journal(
left, right)
if journal_and_volume_conjunction:
return journal_and_volume_conjunction
return AndOp(left, right) if isinstance(node.bool_op, And) else OrOp(
left, right)
def _restructure_if_volume_follows_journal(left, right):
"""Remove volume node if it follows a journal logically in the tree hierarchy.
Args:
left (ast.ASTElement): The journal KeywordOp node.
right (ast.ASTElement): The rest of the tree to be restructured.
Return:
(ast.ASTElement): The restructured tree, with the volume node removed.
Notes:
This happens to support queries like "journal Phys.Rev. and vol d85". Appends the value of KeywordOp with
Keyword 'volume' and discards 'volume' KeywordOp node from the tree.
"""
def _get_volume_keyword_op_and_remaining_subtree(right_subtree):
if isinstance(right_subtree, NotOp) and isinstance(right_subtree.op, KeywordOp) \
and right_subtree.op.left == Keyword('volume'):
return None, None
elif isinstance(right_subtree, AndOp) and isinstance(right_subtree.left, NotOp) \
and isinstance(right_subtree.left.op, KeywordOp) and right_subtree.left.op.left == Keyword('volume'):
return None, right_subtree.right
elif isinstance(right_subtree,
KeywordOp) and right_subtree.left == Keyword('volume'):
return right_subtree, None
elif isinstance(
right_subtree,
AndOp) and right_subtree.left.left == Keyword('volume'):
return right_subtree.left, right_subtree.right
journal_value = left.right.value
volume_and_remaining_subtree = _get_volume_keyword_op_and_remaining_subtree(
right)
if not volume_and_remaining_subtree:
return
volume_node, remaining_subtree = volume_and_remaining_subtree
if volume_node:
left.right.value = ','.join([journal_value, volume_node.right.value])
return AndOp(left, remaining_subtree) if remaining_subtree else left
def visit_boolean_query(self, node):
"""Convert BooleanRule into AndOp or OrOp nodes."""
left = node.left.accept(self)
right = node.right.accept(self)
return AndOp(left, right) if isinstance(node.bool_op, And) else OrOp(
left, right)
RestructuringVisitor
@pytest.mark.parametrize(
['query_str', 'expected_parse_tree'],
[
# Find keyword combined with other production rules
('FIN author:\'ellis\'',
KeywordOp(Keyword('author'), PartialMatchValue('ellis'))),
('Find author "ellis"',
KeywordOp(Keyword('author'), ExactMatchValue('ellis'))),
('f author ellis', KeywordOp(Keyword('author'), Value('ellis'))),
# Invenio like search
('author:ellis and title:boson',
AndOp(KeywordOp(Keyword('author'), Value('ellis')),
KeywordOp(Keyword('title'), Value('boson')))),
('unknown_keyword:\'bar\'',
KeywordOp(Keyword('unknown_keyword'), PartialMatchValue('bar'))),
('dotted.keyword:\'bar\'',
KeywordOp(Keyword('dotted.keyword'), PartialMatchValue('bar'))),
# Boolean operator testing (And/Or)
('author ellis and title \'boson\'',
AndOp(KeywordOp(Keyword('author'), Value('ellis')),
KeywordOp(Keyword('title'), PartialMatchValue('boson')))),
('f a appelquist and date 1983',
AndOp(KeywordOp(Keyword('author'), Value('appelquist')),
KeywordOp(Keyword('date'), Value('1983')))),
('fin a henneaux and citedby a nicolai',
AndOp(
KeywordOp(Keyword('author'), Value('henneaux')),