基于层次匹配的规则引擎

Question

Say, I have a dataset (cities) with this structure:

• ID
• City
• State
• Country
• Continent

And I have a configuration table (key, value) where the key can be a combination of above parameters

For eg:

{
    "continent": "asia"
}

or

{
    "continent": "asia",
    "country" : "india"
}

or

{
    "continent": "asia",
    "country" : "india",
    "state" : "maharashtra",
    "city" : "mumbai"
}

Now, I want to do a closest match of cities from the 1st dataset to the entries in configuration.

For example, if I have

city: mumbai, state: maharashtra, country: india, continent: asia

it should match, 3rd entry in the config above.

If I have

city: tokyo, state: Kantō, country: japan, continent: asia

it should match 1st entry in the config above.

I am looking for suggestions if something is readily available for this kind of a scenario.

I am open to storing the configurations in some different way if there is any.

Ideally I would like a solution in which I can pass multiple entries (inputs) and it should return the closest matching configuration for each of the inputs.

Answer 1

With the risk that I misunderstand the question, I would:

• collect configuration entries in a tree-like data structure, for faster retrieval
• find a match for a single data entry at a time, by traversing through that tree

Here is how that would look:

class Config:
    def __init__(self):
        self.continents = {}

    def add(self, config):
        collection = self.continents
        for prop in ["continent", "country", "state", "city"]:
            if prop not in config:
                break
            key = config[prop]
            if key not in collection:
                collection[key] = {}
            collection = collection[key]
        collection["config"] = config

    def get(self, data):
        collection = self.continents
        config = None
        for prop in ["continent", "country", "state", "city"]:
            if prop not in data or data[prop] not in collection:
                break
            collection = collection[data[prop]]
            if "config" in collection:
                config = collection["config"]
        return config

The above class can be used as follows.

First create an instance, and populate it with configuration entries:

config = Config()

config.add({ "continent": "asia"})
config.add({ "continent": "asia", "country": "india" })
config.add({ "continent": "asia", "country": "india", 
             "state": "maharashtra", "city": "mumbai" })

Then iterate your data entries, and for each call config.get to get the best matching configuration. For example:

print(config.get({ "city": "mumbai", "state": "maharashtra", 
                   "country": "india", "continent": "asia"}))

print(config.get({"city": "tokyo", "state": "kantō", 
                   "country": "japan", "continent": "asia"}))