project-config/tools/fix-neutron-stadium.py
Ian Wienand 4f6629021e Update hacking, fix errors/warnings
This version of hacking doesn't understand f-strings as usable in
Python 3.  Update to the latest and fix current issues, which are all
just formatting fixes.

Change-Id: I0a7d6f93f07477b6dd29ab143130dd9064c250be
2020-01-14 09:40:45 +11:00

191 lines
6.3 KiB
Python

#!/usr/bin/env python
# Copyright 2017 Red Hat, Inc.
#
# 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 collections
import subprocess
import ruamel.yaml
import yaml
# from :
# http://stackoverflow.com/questions/8640959/how-can-i-control-what-scalar-form-pyyaml-uses-for-my-data # noqa: E501
def should_use_block(value):
for c in u"\u000a\u000d\u001c\u001d\u001e\u0085\u2028\u2029":
if c in value:
return True
return False
def my_represent_scalar(self, tag, value, style=None):
if style is None:
if should_use_block(value):
style = '|'
else:
style = self.default_style
node = yaml.representer.ScalarNode(tag, value, style=style)
if self.alias_key is not None:
self.represented_objects[self.alias_key] = node
return node
def project_representer(dumper, data):
return dumper.represent_mapping('tag:yaml.org,2002:map',
data.items())
def construct_yaml_map(self, node):
data = collections.OrderedDict()
yield data
value = self.construct_mapping(node)
if isinstance(node, yaml.MappingNode):
self.flatten_mapping(node)
else:
raise yaml.constructor.ConstructorError(
None, None,
'expected a mapping node, but found %s' % node.id,
node.start_mark)
mapping = collections.OrderedDict()
for key_node, value_node in node.value:
key = self.construct_object(key_node, deep=False)
try:
hash(key)
except TypeError as exc:
raise yaml.constructor.ConstructorError(
'while constructing a mapping', node.start_mark,
'found unacceptable key (%s)' % exc, key_node.start_mark)
value = self.construct_object(value_node, deep=False)
mapping[key] = value
data.update(mapping)
class IndentedEmitter(yaml.emitter.Emitter):
def expect_block_sequence(self):
self.increase_indent(flow=False, indentless=False)
self.state = self.expect_first_block_sequence_item
class IndentedDumper(IndentedEmitter, yaml.serializer.Serializer,
yaml.representer.Representer, yaml.resolver.Resolver):
def __init__(self, stream,
default_style=None, default_flow_style=None,
canonical=None, indent=None, width=None,
allow_unicode=None, line_break=None,
encoding=None, explicit_start=None, explicit_end=None,
version=None, tags=None):
IndentedEmitter.__init__(
self, stream, canonical=canonical,
indent=indent, width=width,
allow_unicode=allow_unicode,
line_break=line_break)
yaml.serializer.Serializer.__init__(
self, encoding=encoding,
explicit_start=explicit_start,
explicit_end=explicit_end,
version=version, tags=tags)
yaml.representer.Representer.__init__(
self, default_style=default_style,
default_flow_style=default_flow_style)
yaml.resolver.Resolver.__init__(self)
def ordered_load(stream, *args, **kwargs):
yaml.add_constructor(yaml.resolver.BaseResolver.DEFAULT_MAPPING_TAG,
construct_yaml_map)
return yaml.safe_load(stream=stream, *args, **kwargs)
def ordered_dump(data, stream=None, *args, **kwargs):
dumper = IndentedDumper
# We need to do this because of how template expasion into a project
# works. Without it, we end up with YAML references to the expanded jobs.
dumper.ignore_aliases = lambda self, data: True
yaml.add_representer(collections.OrderedDict, project_representer,
Dumper=IndentedDumper)
output = yaml.dump(
data, default_flow_style=False,
Dumper=dumper, width=80, *args, **kwargs).replace(
'\n-', '\n\n-')
if stream:
stream.write(output)
else:
return output
def get_single_key(var):
if isinstance(var, str):
return var
elif isinstance(var, list):
return var[0]
return list(var.keys())[0]
def has_single_key(var):
if isinstance(var, list):
return len(var) == 1
if isinstance(var, str):
return True
dict_keys = list(var.keys())
if len(dict_keys) != 1:
return False
if var[get_single_key(var)]:
return False
return True
def main():
subprocess.run(['git', 'checkout', '--', 'zuul.d/projects.yaml'])
yaml = ruamel.yaml.YAML()
yaml.indent(mapping=2, sequence=4, offset=2)
projects = yaml.load(open('zuul.d/projects.yaml', 'r'))
for project in projects:
if project['project']['name'].split('/')[1].startswith('networking-'):
if 'templates' not in project['project']:
continue
templates = project['project']['templates']
for template in ('openstack-python-jobs',
'openstack-python35-jobs'):
if template in templates:
new_name = template + '-neutron'
templates[templates.index(template)] = new_name
yaml.dump(projects, open('zuul.d/projects.yaml', 'w'))
# Strip the extra 2 spaces that ruamel.yaml appends because we told it
# to indent an extra 2 spaces. Because the top level entry is a list it
# applies that indentation at the top. It doesn't indent the comment lines
# extra though, so don't do them.
with open('zuul.d/projects.yaml', 'r') as main_in:
main_content = main_in.readlines()
with open('zuul.d/projects.yaml', 'w') as main_out:
for line in main_content:
if '#' in line:
main_out.write(line)
else:
if line.startswith(' - project'):
main_out.write('\n')
main_out.write(line[2:])
if __name__ == '__main__':
main()