Pytechecker is a utility library for type checking objects against sample objects.
- A simple example
- Embedded dicts
- What about lists?
- Dicts as list elements
- Tuples
- Dicts and lists in tuples
Pytechecker works excellently for checking API payloads. For how it can be used for this, here's an example of how it can be used together with a flask api.
Use pip to install pytechecker python3 -m pip install pytechecker.
Then you simply use the check function from the module.
from pytechecker import check
sample = {
"name": {
"required": True,
"allowed_types": [str]
},
"age": {
"required": True,
"allowed_types": [int]
}
}
obj = {
"name": "dcronqvist",
"age": 21
}
# succ will be True if it passed type checking
# errors will be an array of strings that tell you
# what went wrong during type checking, if succ is False.
succ, errors = check(sample, obj)Let's start by taking a look at a very simple sample object:
{
"name": {
"required": True,
"allowed_types": [str]
},
"age": {
"required": False,
"allowed_types": [int]
}
}Here we have defined a sample object which has one required key name and one optional key age. In the sample object we have also specified that the key name can only be of the type str, and age can only be an int. Let's see which objects that fit this sample.
{
"name": "Daniel",
"age": 21
}Above we have an object which fits the sample object. It's an object which has the required key name and the optional key age, and they both are their respective required types.
{
"name": "Daniel"
}Above is an object that still fits the sample object. Since the key age is optional, we can omit it from the object without it causing it to be unfit.
{
"age": 21
}The above object does, however, NOT fit the sample object. Upon attempting to match the object against the sample, you'll be met with the following error:
ERROR: Key 'name' is required, but was absent in supplied object.
We can also look at an example like this:
{
"name": "Daniel",
"age": 21.4
}The above object is unfit since one of its keys is of a type that is not allowed for that key. You'll be met with the following error:
ERROR: On key 'age', expected one of ['int'], got float.
So there we have it. That's a very simple example of how it works.
We don't want to be limited to dicts of only one level, so thankfully pytechecker can handle virtually infinite levels of dicts in dicts. Let's look at an example.
{
"tenant": {
"required": True,
"allowed_types": [dict],
"embedded_dict": {
"name": {
"required": True,
"allowed_types": [str]
},
"age": {
"required": True,
"allowed_types": [int]
}
}
},
"room": {
"required": True,
"allowed_types": [str]
}
}Above we have a sample object which has a top level key which is allowed to a dict. We can very easily also specify what this embedded dict must look like for it to fit. Let's look at some objects!
{
"tenant": {
"name": "Daniel",
"age": 21
},
"room": "RM-212512"
}This is an object which fits, like a glove! We have the key tenant which of course is a dict (or object) and then the key room which is a string, like specified! We can also see that all keys inside of tenant are required, so leaving one out would result in something like the following:
ERROR: Key 'tenant.age' is required, but was absent in supplied object.
Pytechecker is very nice and also tells you the entire key from top level and down, and won't just say that age is missing.
Like said, this works for a virtually infinite amount of dicts in dicts - it's done via recursion!
Lists are handled very nicely by pytechecker. It will go through all elements in the list and type check them for you. You can even have list elements that are dicts, and they'll also be handled! Let's look at a simple example with lists to begin with.
{
"nums": {
"required": True,
"allowed_types": [list],
"list_element": {
"allowed_types": [int, float]
}
}
}Here we have a sample object with a required key nums which must be a list of numbers.
{
"nums": [5, 1.2, 9, 12.4]
}Above here is an object which fits since it has they key nums and all elements are of the correct types.
{
"nums": [5, 1.2, 9, 12.4, "5"]
}Having something like the object above would give an error like the following:
ERROR: On key 'nums[4]', expected one of ['int', 'float'], got str.
See that [4]? Pytechecker also tells you which element(s) that are wrong in a list. If we would have multiple wrong elements:
{
"nums": [5, "1.2", 9, 12.4, "5"]
}We would get multiple errors like this:
[
"ERROR: On key 'nums[1]', expected one of ['int', 'float'], got str.",
"ERROR: On key 'nums[4]', expected one of ['int', 'float'], got str."
]Important to note is that all errors that have been shown before this example have also been part of an array, they have just been the only element in that array of errors.
{
"people": {
"required": True,
"allowed_types": [list],
"list_element": {
"allowed_types": [dict],
"embedded_dict": {
"name": {
"required": True,
"allowed_types": [str]
},
"age": {
"required": False,
"allowed_types": [int]
}
}
}
}
}Right, so here we have a sample object which contains a key people which should contain a list of objects with a required key name and optional key age. Let's look at some objects.
{
"people": [
{
"name": "Daniel",
"age": 21
},
{
"name": "John",
"age": 42
},
{
"name": "Betty"
}
]
}Here is our list of people with their names and optional age. This object fits, so let's look at one that doesn't fit.
{
"people": [
{
"name": "Daniel",
"age": 21
},
{
"name": "John",
"age": 42
},
{
"age": 35
}
]
}The person Betty now doesn't have a name, instead it only has an age. This will result in the following error:
ERROR: Key 'people[2].name' is required, but was absent in supplied object.
Once again, pytechecker tells you which of these objects that is unfit, and also what about that object that is unfit.
Tuples are much like lists, and handled a lot like them. However, tuples are specified by type order, so that the supplied object must match type order as well, let's look at an example.
{
"tuple": {
"required": True,
"allowed_types": [tuple],
"tuple_order": [str, int, int, float]
}
}Here we have a sample object which specifies a tuple with the specific order of (str, int, int, float). Let's look at objects!
{
"tuple": ("Hello there!", 5, 2, 1.1)
}There we are, an object which fits the tuple description perfectly. If we, however, were to flip around some values to not be in the same type order:
{
"tuple": ("Hello there!", 5, 1.1, 2)
}Then we'd be met with the following error:
ERROR: On key 'tuple', expected tuple with order (str,int,int,float), got tuple with order (str,int,float,int).
{
"tuple": {
"required": True,
"allowed_types": [tuple],
"tuple_order": [list, int, dict],
"list_element": {
"allowed_types": [dict],
"embedded_dict": {
"name": {
"required": True,
"allowed_types": [str]
}
}
},
"embedded_dict": {
"num": {
"required": True,
"allowed_types": [int, float]
}
}
}
}A sample object which should contain a key tuple, which also should be a tuple. This tuple should, respectively, contain a list of objects, an int and a dict.
{
"tuple": (
[
{
"name": "Daniel"
},
{
"name": "John"
}
],
5,
{
"num": 25
}
)
}The above object fits like a glove!