robotframework-2.7.7.src.robot.parsing.datarow.py Maven / Gradle / Ivy
Go to download
Show more of this group Show more artifacts with this name
Show all versions of robotframework Show documentation
Show all versions of robotframework Show documentation
High level test automation framework
The newest version!
# Copyright 2008-2012 Nokia Siemens Networks Oyj
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
import re
class DataRow(object):
_row_continuation_marker = '...'
_whitespace_regexp = re.compile('\s+')
_ye_olde_metadata_prefix = 'meta:'
def __init__(self, cells):
self.cells, self.comments = self._parse(cells)
def _parse(self, row):
data = []
comments = []
for cell in row:
cell = self._collapse_whitespace(cell)
if cell.startswith('#') or comments:
comments.append(cell)
else:
data.append(cell)
return self._purge_empty_cells(data), self._purge_empty_cells(comments)
def _collapse_whitespace(self, cell):
return self._whitespace_regexp.sub(' ', cell).strip()
def _purge_empty_cells(self, row):
while row and not row[-1]:
row.pop()
# Cells with only a single backslash are considered empty
return [cell if cell != '\\' else '' for cell in row]
@property
def head(self):
return self.cells[0] if self.cells else ''
@property
def _tail(self):
return self.cells[1:]
@property
def all(self):
return self.cells
@property
def data(self):
if self.is_continuing():
index = self.cells.index(self._row_continuation_marker) + 1
return self.cells[index:]
return self.cells
def dedent(self):
datarow = DataRow([])
datarow.cells = self._tail
datarow.comments = self.comments
return datarow
def handle_old_style_metadata(self):
if self._is_metadata_with_olde_prefix(self.head):
self.cells = self._convert_to_new_style_metadata()
def _is_metadata_with_olde_prefix(self, value):
return value.lower().startswith(self._ye_olde_metadata_prefix)
def _convert_to_new_style_metadata(self):
return ['Metadata'] + [self.head.split(':', 1)[1].strip()] + self._tail
def starts_for_loop(self):
if self.head and self.head.startswith(':'):
return self.head.replace(':', '').replace(' ', '').upper() == 'FOR'
return False
def starts_test_or_user_keyword_setting(self):
head = self.head
return head and head[0] == '[' and head[-1] == ']'
def test_or_user_keyword_setting_name(self):
return self.head[1:-1].strip()
def is_indented(self):
return self.head == ''
def is_continuing(self):
for cell in self.cells:
if cell == self._row_continuation_marker:
return True
if cell:
return False
def is_commented(self):
return bool(not self.cells and self.comments)
def __nonzero__(self):
return bool(self.cells or self.comments)