如何在boto3中模拟AWS CLI EC2过滤器

Question

I'm looking for a way to mimic the AWS CLI EC2 Filter using Boto3 let say i want to translate the filters portion of describe-instances command:

aws ec2 describe-instances --filters "Name=instance- 
type,Values=m1.small,t2.small"

into Boto3 describe_instances method:

response = client.describe_instances(
    Filters=[
        {
            'Name': 'instance- type',
            'Values': [
                'm1.small','t2.small',
            ]
        }
      ]
    )

So basically what i'm asking, what is good why in python to take the string:

"Name=instance-type,Values=m1.small,t2.small"

and convert it to:

[
  {
     'Name': 'instance- type',
     'Values': [
       'm1.small','t2.small',
     ]
  }
]

so that i can use it as a filter parameter in boto3's decribe_instances method.

Answer 1

The following will match the exact format given, but will run into problems if the format varies too much:

import re

filter='Name=instance-type,Values=m1.small,t2.small'

match = re.search('(.*)=(.*),(.*)=(.*)', filter)

f = {match.group(1) : match.group(2), match.group(3) : match.group(4).split(',')}

# f is a normal Python dictionary
print (f)

# Or, convert it to JSON
import json
print (json.dumps(f))

Output is:

{'Values': ['m1.small', 't2.small'], 'Name': 'instance-type'}
{"Values": ["m1.small", "t2.small"], "Name": "instance-type"}

Order doesn't matter for a dictionary. You can also wrap the output in '[]', but that makes it not strictly JSON.

A great site for testing Python regex expressions: Pythex

Answer 2

for cases where filter has multiparts seperated with a ; since the "Name" and "Values" are specific to this filter

def parse_filter_field(filter_str):
    filters = []
    regex = re.compile(r'name=([\w\d_:.-]+),values=([/\w\d_,.\*]+)', flags=re.I)
    for f in filter_str.split(';'):
        match = regex.match(f)
        if match is None:
            print 'could not parse filter: %s' % (f, )
            continue

        filters.append({
            'Name' : match.group(1),
            'Values' : match.group(2).split(',')
            })

return filters