Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

feat(cfn): add plugin #584

Draft
wants to merge 4 commits into
base: dev
Choose a base branch
from
Draft
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
206 changes: 206 additions & 0 deletions jc/parsers/cfn.py
Original file line number Diff line number Diff line change
@@ -0,0 +1,206 @@
r"""
jc - JSON Convert `CloudFormation` file parser

> Note: `datetime` objects will be converted to strings.

Usage (cli):

$ cat template.yaml | jc --cfn

Usage (module):

import jc
result = jc.parse('cfn', yaml_file_output)

Schema:

YAML Document converted to a Dictionary.
See https://pypi.org/project/ruamel.yaml for details.

[
{
"key1": string/int/float/boolean/null/array/object,
"key2": string/int/float/boolean/null/array/object
}
]

Examples:

$ cat template.yaml
AWSTemplateFormatVersion: "2010-09-09"
Parameters:
EnvType:
Type: "String"
Default: "test"
Conditions:
CreateProdResources: !Equals [ !Ref EnvType, prod ]
Resources:
MyInstance:
Type: "AWS::EC2::Instance"
Properties:
ImageId: "ami-1234567890abcdefg"
InstanceType: "t2.micro"


$ cat template.yaml | jc --cfn -p
[
{
"AWSTemplateFormatVersion": "2010-09-09",
"Parameters": {
"EnvType": {
"Type": "String",
"Default": "test"
}
},
"Conditions": {
"CreateProdResources": [
"EnvType",
"prod"
]
},
"Resources": {
"MyInstance": {
"Type": "AWS::EC2::Instance",
"Properties": {
"ImageId": "ami-1234567890abcdefg",
"InstanceType": "t2.micro"
}
}
}
}
]
"""
from dataclasses import dataclass
from enum import Enum
from typing import List
import jc.utils
from jc.exceptions import LibraryNotInstalled


class info():
"""Provides parser metadata (version, author, etc.)"""
version = '1.7'
description = 'YAML file parser'
author = 'Ron Green'
author_email = '[email protected]'
details = 'Using the ruamel.yaml library at https://pypi.org/project/ruamel.yaml'
compatible = ['linux', 'darwin', 'cygwin', 'win32', 'aix', 'freebsd']
tags = ['standard', 'file', 'string']


__version__ = info.version


def _process(proc_data):
"""
Final processing to conform to the schema.

Parameters:

proc_data: (List of Dictionaries) raw structured data to process

Returns:

List of Dictionaries. Each dictionary represents a YAML document.
"""

# No further processing
return proc_data

class Option(Enum):
scalar = 1
sequence = 2
mapping = 3


@dataclass
class CustomTag:
name: str
options: List[Option]


custom_tags = [
CustomTag(name='!And', options=[Option.sequence]),
CustomTag(name='!Base64', options=[Option.scalar]),
CustomTag(name='!Cidr', options=[Option.scalar]),
CustomTag(name='!Condition', options=[Option.scalar]),
CustomTag(name='!Equals', options=[Option.scalar, Option.sequence]),
CustomTag(name='!FindInMap', options=[Option.sequence]),
CustomTag(name='!GetAZs', options=[Option.scalar]),
CustomTag(name='!GetAtt', options=[Option.scalar]), # , Option.sequence]),
CustomTag(name='!If', options=[Option.sequence]),
CustomTag(name='!ImportValue', options=[Option.scalar]),
CustomTag(name='!Join', options=[Option.sequence]),
CustomTag(name='!Not', options=[Option.sequence]),
CustomTag(name='!Or', options=[Option.sequence]),
CustomTag(name='!Ref', options=[Option.scalar]),
CustomTag(name='!Select', options=[Option.sequence]),
CustomTag(name='!Split', options=[Option.sequence]),
CustomTag(name='!Sub', options=[Option.scalar]),
CustomTag(name='!Transform', options=[Option.mapping]),
]


def parse(data, raw=False, quiet=False):
"""
Main text parsing function

Parameters:

data: (string) text data to parse
raw: (boolean) unprocessed output if True
quiet: (boolean) suppress warning messages if True

Returns:

List of Dictionaries representing the YAML documents.
"""
# check if yaml library is installed and fail gracefully if it is not
try:
from ruamel.yaml import YAML
from ruamel.yaml.constructor import SafeConstructor

except Exception:
raise LibraryNotInstalled('The ruamel.yaml library is not installed.')

jc.utils.compatibility(__name__, info.compatible, quiet)
jc.utils.input_type_check(data)

raw_output = []

if jc.utils.has_data(data):

# monkey patch to disable plugins since we don't use them and in
# ruamel.yaml versions prior to 0.17.0 the use of __file__ in the
# plugin code is incompatible with the pyoxidizer packager
YAML.official_plug_ins = lambda a: []

yaml = YAML(typ='safe')
for custom_tag in custom_tags:
for option in custom_tag.options:
if option == Option.mapping:
yaml.constructor.add_constructor(custom_tag.name, lambda loader, node: loader.construct_mapping(node))
continue
if option == Option.sequence:
yaml.constructor.add_constructor(custom_tag.name, lambda loader, node: loader.construct_sequence(node))
continue
if option == Option.scalar:
yaml.constructor.add_constructor(custom_tag.name, lambda loader, node: loader.construct_scalar(node))
continue

# yaml.Constructor = SafeConstructor
# modify the timestamp constructor to output datetime objects as
# strings since JSON does not support datetime objects
yaml.constructor.yaml_constructors['tag:yaml.org,2002:timestamp'] = \
yaml.constructor.yaml_constructors['tag:yaml.org,2002:str']




for document in yaml.load_all(data):
raw_output.append(document)

if raw:
return raw_output
else:
return _process(raw_output)
Loading