YAML array extention in Python

example.yaml

CoreTeam: &CoreTeam
  Characters:
    - Mario
    - Luigi

ExtendedTeam: &ExtendedTeam
  <<: *CoreTeam
  Characters+:
    - Princess Peach
    - Yoshi
    - Toad

Games:
  SuperMario:
    <<: *CoreTeam
    Characters+:
      - Peach

  Mario Cart 6:
    <<: *ExtendedTeam
    Characters++:
      - Donkey Kong

  Mario Cart 7:
    <<: *ExtendedTeam
    Characters++:
      - Donkey Kong
      - Metal Mario

run_example.py

#!/usr/bin/env python3

import yaml
from yaml_utils import execute_merge_notation
import json

with open("example.yaml") as f:
    data = yaml.load(f)

print("--- Before --- ")
print(json.dumps(data, indent=4))
execute_merge_notation(data)
print("--- After --- ")
print(json.dumps(data, indent=4))

run_example_output.txt

--- Before --- 
{
    "CoreTeam": {
        "Characters": [
            "Mario",
            "Luigi"
        ]
    },
    "Games": {
        "SuperMario": {
            "Characters": [
                "Mario",
                "Luigi"
            ],
            "Characters+": [
                "Peach"
            ]
        },
        "Mario Cart 7": {
            "Characters": [
                "Mario",
                "Luigi"
            ],
            "Characters+": [
                "Princess Peach",
                "Yoshi",
                "Toad"
            ],
            "Characters++": [
                "Donkey Kong",
                "Metal Mario"
            ]
        },
        "Mario Cart 6": {
            "Characters": [
                "Mario",
                "Luigi"
            ],
            "Characters+": [
                "Princess Peach",
                "Yoshi",
                "Toad"
            ],
            "Characters++": [
                "Donkey Kong"
            ]
        }
    },
    "ExtendedTeam": {
        "Characters": [
            "Mario",
            "Luigi"
        ],
        "Characters+": [
            "Princess Peach",
            "Yoshi",
            "Toad"
        ]
    }
}
--- After --- 
{
    "CoreTeam": {
        "Characters": [
            "Mario",
            "Luigi"
        ]
    },
    "Games": {
        "SuperMario": {
            "Characters": [
                "Mario",
                "Luigi",
                "Peach"
            ]
        },
        "Mario Cart 7": {
            "Characters": [
                "Mario",
                "Luigi",
                "Princess Peach",
                "Yoshi",
                "Toad",
                "Donkey Kong",
                "Metal Mario"
            ]
        },
        "Mario Cart 6": {
            "Characters": [
                "Mario",
                "Luigi",
                "Princess Peach",
                "Yoshi",
                "Toad",
                "Donkey Kong"
            ]
        }
    },
    "ExtendedTeam": {
        "Characters": [
            "Mario",
            "Luigi",
            "Princess Peach",
            "Yoshi",
            "Toad"
        ]
    }
}

yaml_utils.py

''' 
YAML has a missing feature: ability to extend lists. This is a Python implementation that allows extending the lists,
as well as the maps of the elements referenced by an alias.
'''
def _trailing_plus_count(s):
    for i in range(0, len(s)):
        if s[-i-1] != '+':
            return i
    return len(s)


def _extend_list_or_dict(json, json2, key, merge_lists=True):
    if type(json) == dict and type(json2) == dict:
        for k2, v2 in json2.items():
           json[k2] = v2
        return json
    elif merge_lists and type(json) == list and type(json2) == list:
        json.extend(json2)
        return json
    return json2


def execute_merge_notation(json):
    if type(json) == dict:
        to_merge = None
        for k in json.keys():
            if type(k) == str and _trailing_plus_count(k) > 0:
                if not to_merge:
                    to_merge = [k]
                else:
                    to_merge.append(k)
        if to_merge:
            for k2 in sorted(to_merge, key=_trailing_plus_count):
                v2 = json[k2]
                k = k2[0: len(k2) - _trailing_plus_count(k2)]
                if len(k) > 0:
                    if k in json:
                        v = json[k]
                        json[k] = _extend_list_or_dict(v, v2, k)
                    else:
                        json[k] = v2
                    json.pop(k2)

        for v in json.values():
            if type(v) in (list, dict):
                execute_merge_notation(v)
    elif type(json) == list:
        for item in json:
            execute_merge_notation(item)

This code is written to aid my answer to this StackOverflow question: http://stackoverflow.com/questions/9254178/is-there-yaml-syntax-for-sharing-part-of-a-list-or-map/41728592#41728592

This does not work as intended, at least not with yaml.load(f, Loader=yaml.FullLoader)

Output has tons of duplicates -- seems references are overwritten, which means that additions to list elements pollute the original lists.

Changing the _extend_list_or_dict function works for me:

def _extend_list_or_dict(json, json2, key, merge_lists=True):
    if type(json) == dict and type(json2) == dict:
        for k2, v2 in json2.items():
           json[k2] = v2
        return json
    elif merge_lists and type(json) == list and type(json2) == list:
        res = []
        res.extend(json)
        res.extend(json2)
        #json.extend(json2)
        #return json
        return res
    return json2