Converts a list of tuples to a nested "family tree dict"
pip install list2tree
Tested against Windows 10 / Python 3.10 / Anaconda
from list2tree import treedict,edit_value_in_treedict,get_value_in_treedict
import pprint as pa
pprintx=pa.PrettyPrinter(indent=1,width=1,compact=True)
pprint=pprintx.pprint
nameconnections = [
("Maria", "Anna"),
("Anna", "Joao"),
("Kasimir", "Maria"),
("Hans", "Fritz"),
("Fritz", "Anna"),
("Günther", "Wolfgang"),
("Joao", "Wolfgang"),
]
print("0) --------------------------------------------------------------")
mapped, airvar = treedict(
pairs_list=nameconnections, main_mapping_keys=(), bi_rl_lr="lr"
)
pprint(mapped)
pprint(airvar)
print("1) --------------------------------------------------------------")
mapped, airvar = treedict(
pairs_list=nameconnections, main_mapping_keys=(), bi_rl_lr="rl"
)
pprint(mapped)
pprint(airvar)
print("2) --------------------------------------------------------------")
mapped, airvar = treedict(
pairs_list=nameconnections, main_mapping_keys=(), bi_rl_lr="bi"
)
pprint(mapped)
pprint(airvar)
print("3) --------------------------------------------------------------")
mapped, airvar = treedict(
pairs_list=nameconnections, main_mapping_keys=("Anna", "Günther"), bi_rl_lr="lr"
)
pprint(mapped)
pprint(airvar)
print("4) --------------------------------------------------------------")
mapped, airvar = treedict(
pairs_list=nameconnections, main_mapping_keys=("Anna", "Wolfgang"), bi_rl_lr="rl"
)
pprint(mapped)
pprint(airvar)
print("5) --------------------------------------------------------------")
mapped, airvar = treedict(
pairs_list=nameconnections, main_mapping_keys=("Anna",), bi_rl_lr="bi"
)
pprint(mapped)
pprint(airvar)
print("6) --------------------------------------------------------------")
v1 = get_value_in_treedict(airvar, ("Anna", "Joao", "Wolfgang", "Günther"))
print(v1)
edit_value_in_treedict(
airvar, ("Anna", "Joao", "Wolfgang", "Günther"), "new_value_edit"
)
pprint(airvar)
v2 = get_value_in_treedict(airvar, ("Anna", "Joao", "Wolfgang", "Günther"))
print(v2)
print("7) --------------------------------------------------------------")
0) --------------------------------------------------------------
[(1,
('Wolfgang',)),
(2,
('Günther',
'Wolfgang')),
(3,
('Joao',
'Wolfgang')),
(4,
('Anna',
'Joao',
'Wolfgang')),
(5,
('Fritz',
'Anna',
'Joao',
'Wolfgang')),
(6,
('Maria',
'Anna',
'Joao',
'Wolfgang')),
(7,
('Hans',
'Fritz',
'Anna',
'Joao',
'Wolfgang')),
(8,
('Kasimir',
'Maria',
'Anna',
'Joao',
'Wolfgang'))]
{'Anna': {'Joao': {'Wolfgang': 4}},
'Fritz': {'Anna': {'Joao': {'Wolfgang': 5}}},
'Günther': {'Wolfgang': 2},
'Hans': {'Fritz': {'Anna': {'Joao': {'Wolfgang': 7}}}},
'Joao': {'Wolfgang': 3},
'Kasimir': {'Maria': {'Anna': {'Joao': {'Wolfgang': 8}}}},
'Maria': {'Anna': {'Joao': {'Wolfgang': 6}}},
'Wolfgang': 1}
1) --------------------------------------------------------------
[(1,
('Günther',)),
(2,
('Hans',)),
(3,
('Kasimir',)),
(4,
('Fritz',
'Hans')),
(5,
('Maria',
'Kasimir')),
(6,
('Wolfgang',
'Günther')),
(7,
('Wolfgang',
'Joao',
'Anna',
'Maria',
'Kasimir')),
(8,
('Wolfgang',
'Joao',
'Anna',
'Fritz',
'Hans')),
(9,
('Anna',
'Maria',
'Kasimir')),
(10,
('Anna',
'Fritz',
'Hans')),
(11,
('Joao',
'Anna',
'Maria',
'Kasimir')),
(12,
('Joao',
'Anna',
'Fritz',
'Hans'))]
{'Anna': {'Fritz': {'Hans': 10},
'Maria': {'Kasimir': 9}},
'Fritz': {'Hans': 4},
'Günther': 1,
'Hans': 2,
'Joao': {'Anna': {'Fritz': {'Hans': 12},
'Maria': {'Kasimir': 11}}},
'Kasimir': 3,
'Maria': {'Kasimir': 5},
'Wolfgang': {'Günther': 6,
'Joao': {'Anna': {'Fritz': {'Hans': 8},
'Maria': {'Kasimir': 7}}}}}
2) --------------------------------------------------------------
[(1,
('Anna',
'Joao',
'Wolfgang',
'Günther')),
(2,
('Anna',
'Maria',
'Kasimir')),
(3,
('Anna',
'Fritz',
'Hans')),
(4,
('Fritz',
'Anna',
'Joao',
'Wolfgang',
'Günther')),
(5,
('Fritz',
'Anna',
'Maria',
'Kasimir')),
(6,
('Fritz',
'Hans')),
(7,
('Günther',
'Wolfgang',
'Joao',
'Anna',
'Maria',
'Kasimir')),
(8,
('Günther',
'Wolfgang',
'Joao',
'Anna',
'Fritz',
'Hans')),
(9,
('Hans',
'Fritz',
'Anna',
'Joao',
'Wolfgang',
'Günther')),
(10,
('Hans',
'Fritz',
'Anna',
'Maria',
'Kasimir')),
(11,
('Joao',
'Wolfgang',
'Günther')),
(12,
('Joao',
'Anna',
'Maria',
'Kasimir')),
(13,
('Joao',
'Anna',
'Fritz',
'Hans')),
(14,
('Kasimir',
'Maria',
'Anna',
'Joao',
'Wolfgang',
'Günther')),
(15,
('Kasimir',
'Maria',
'Anna',
'Fritz',
'Hans')),
(16,
('Maria',
'Anna',
'Joao',
'Wolfgang',
'Günther')),
(17,
('Maria',
'Anna',
'Fritz',
'Hans')),
(18,
('Maria',
'Kasimir')),
(19,
('Wolfgang',
'Günther')),
(20,
('Wolfgang',
'Joao',
'Anna',
'Maria',
'Kasimir')),
(21,
('Wolfgang',
'Joao',
'Anna',
'Fritz',
'Hans'))]
{'Anna': {'Fritz': {'Hans': 3},
'Joao': {'Wolfgang': {'Günther': 1}},
'Maria': {'Kasimir': 2}},
'Fritz': {'Anna': {'Joao': {'Wolfgang': {'Günther': 4}},
'Maria': {'Kasimir': 5}},
'Hans': 6},
'Günther': {'Wolfgang': {'Joao': {'Anna': {'Fritz': {'Hans': 8},
'Maria': {'Kasimir': 7}}}}},
'Hans': {'Fritz': {'Anna': {'Joao': {'Wolfgang': {'Günther': 9}},
'Maria': {'Kasimir': 10}}}},
'Joao': {'Anna': {'Fritz': {'Hans': 13},
'Maria': {'Kasimir': 12}},
'Wolfgang': {'Günther': 11}},
'Kasimir': {'Maria': {'Anna': {'Fritz': {'Hans': 15},
'Joao': {'Wolfgang': {'Günther': 14}}}}},
'Maria': {'Anna': {'Fritz': {'Hans': 17},
'Joao': {'Wolfgang': {'Günther': 16}}},
'Kasimir': 18},
'Wolfgang': {'Günther': 19,
'Joao': {'Anna': {'Fritz': {'Hans': 21},
'Maria': {'Kasimir': 20}}}}}
3) --------------------------------------------------------------
[(1,
('Günther',
'Wolfgang')),
(2,
('Anna',
'Joao',
'Wolfgang'))]
{'Anna': {'Joao': {'Wolfgang': 2}},
'Günther': {'Wolfgang': 1}}
4) --------------------------------------------------------------
[(1,
('Wolfgang',
'Günther')),
(2,
('Wolfgang',
'Joao',
'Anna',
'Maria',
'Kasimir')),
(3,
('Wolfgang',
'Joao',
'Anna',
'Fritz',
'Hans')),
(4,
('Anna',
'Maria',
'Kasimir')),
(5,
('Anna',
'Fritz',
'Hans'))]
{'Anna': {'Fritz': {'Hans': 5},
'Maria': {'Kasimir': 4}},
'Wolfgang': {'Günther': 1,
'Joao': {'Anna': {'Fritz': {'Hans': 3},
'Maria': {'Kasimir': 2}}}}}
5) --------------------------------------------------------------
[(1,
('Anna',
'Joao',
'Wolfgang',
'Günther')),
(2,
('Anna',
'Maria',
'Kasimir')),
(3,
('Anna',
'Fritz',
'Hans'))]
{'Anna': {'Fritz': {'Hans': 3},
'Joao': {'Wolfgang': {'Günther': 1}},
'Maria': {'Kasimir': 2}}}
6) --------------------------------------------------------------
1
{'Anna': {'Fritz': {'Hans': 3},
'Joao': {'Wolfgang': {'Günther': 'new_value_edit'}},
'Maria': {'Kasimir': 2}}}
new_value_edit
7) --------------------------------------------------------------
treedict(
pairs_list: list | tuple, main_mapping_keys: tuple | list = (), bi_rl_lr: str = "lr"
):
"""
This function takes a list or tuple of pairs, a tuple or list of main mapping keys, and a string indicating the type
of bidirectional mapping to be used. It returns a tuple of flattened dictionary and a nested dictionary.
:param pairs_list: A list or tuple of pairs.
:type pairs_list: list | tuple
:param main_mapping_keys: A tuple or list of main mapping keys.
:type main_mapping_keys: tuple | list
:param bi_rl_lr: A string indicating the type of bidirectional mapping to be used. Default is "lr", Valid options are
lr - left to right / rl - right to left / bi - bidirectional
:type bi_rl_lr: str
:return: A tuple of flattened dictionary and a nested dictionary.
:rtype: tuple
"""
edit_value_in_treedict(d: dict, keys: list | tuple, newvalue: Any) -> None:
"""
Edit the value of a nested dictionary given a list of keys.
Args:
d (dict): The nested dictionary to be edited.
keys (list,tuple): A list of keys representing the path to the value to be edited.
newvalue (Any): The new value to be assigned to the specified key path.
Returns:
None: This function does not return anything, but it changes the ORIGINAL dict! If you want to keep a copy of
the original dict, use copy.deepcopy(d) before you call the function
Raises:
KeyError: If any of the keys in the list do not exist in the dictionary.
Example:
d = {'a': {'b': {'c': 1}}}
keys = ['a', 'b', 'c']
newvalue = 2
edit_value_in_treedict(d, keys, newvalue)
# d is now {'a': {'b': {'c': 2}}}
"""
get_value_in_treedict(d: dict, keys: list | tuple) -> Any:
"""
Returns the value in a nested dictionary `d` at the specified `keys`.
Args:
d (dict): The nested dictionary to search for the value.
keys (list|tuple): The list or tuple of keys to traverse the nested dictionary.
Returns:
Any: The value at the specified `keys` in the nested dictionary `d`.
Raises:
TypeError: If `d` is not a dictionary or `keys` is not a list or tuple.
KeyError: If any of the keys in `keys` do not exist in the nested dictionary `d`.
"""
Metadata
Release files for list2tree 0.10
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | Actions |
|---|---|---|---|---|
| list2tree-0.10.tar.gz | 6.7 kB | Without provenance | Details for list2tree-0.10.tar.gz |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Actions Reset |
|---|---|---|---|---|
| list2tree-0.10-py3-none-any.whl | Python 3 | none | any | Details for list2tree-0.10-py3-none-any.whl |
Total release size: 16.2 kB
Release files / list2tree-0.10.tar.gz
| Download URL | list2tree-0.10.tar.gz |
|---|---|
| Size | 6.7 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
0eabfb47d77e96e24a8bfc444746db935aa550c08b85efd0996dc13606f11b6c
|
|
BLAKE2b-256 checksum How to use checksums |
58435bad03e51c36cd0a5140eff13c2c3595d3ad11bdff5ba7b4fe39b5e5c748
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/4.0.2 CPython/3.10.10
|
Release files / list2tree-0.10-py3-none-any.whl
| Download URL | list2tree-0.10-py3-none-any.whl |
|---|---|
| Size | 9.5 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
d820de0381e7cc18be168d3fa3d2cd2d0ee8aef1185c7c1f1e468d93685f90e7
|
|
BLAKE2b-256 checksum How to use checksums |
eea2c845e5b62b171bab1fbcc8ffd64f9acd6ae0fe8e26c1ecf67dbee7965d88
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/4.0.2 CPython/3.10.10
|
Security for list2tree 0.10
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
No provenance found
None of the files in this release were uploaded with provenance metadata.
A note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
To learn more about protecting yourself from malicious actors, visit our package security documentation.