WordPath

An idea of @cspaier

• How to start the API ?

Just run uvicorn main:app --reload
(you must have uvicorn installed)

how does it work ?

The server will calculate all "nearest words" (L¹) of the source and the objective, and cache them in preloaded.json. This file has been structured thanks a simple regex check:

For example, if the word is words, we will decompose in 5 regex checks:

• .ords
• w.rds
• wo.ds
• wor.s
• word. In regex, . means "Any character", so regex will check any word which have only one letter changed

• If any word is in both list, then the api found a word in 3 step.

• Else it will retrieve all "nearest words"(L²) of the "nearest words" list L¹.

• If any word L² is in the nearest words of the objective, then the api found a way to connect those two words.

• Else, the same process will be applied with Lⁿ⁻¹ and Lⁿ until n = maxLenght if lenght_limit is set to True (default: False)

If you add any words to words.txt, please set if preload to True (default: False). It will regenerate words.json (it can take severals minutes) Once words.json is updated, you can set preload to False

Endpoints

Get the shortest path between 2 words:

GET /path HTTP/1.1
Host: 127.0.0.1:8000
Content-Type: application/json
Content-Length: 77

{
    "starting": "pomme",
    "objective": "ville",
    "maxLenght": 7
}

Get the nearest words of a word:

GET /nearest-words HTTP/1.1
Host: 127.0.0.1:8000
Content-Type: application/json
Content-Length: 25

{
    "word": "ville"
}

Get two random words from the dictionary

GET /words HTTP/1.1
Host: 127.0.0.1:8000