ApiDoc Automation script created & mock portfolio data generator refectored.

[Purvanshsingh]

Fixes #15

As of now, we create all the classes and properties in ApiDoc using docwriter.py manually.
I have created the script which will parse the NonPerformingLoan.jsonld and can create hydra classes & properties for ApiDoc.

NonPerformingLoan.jsonld --> ApiDoc

Methods used to parse the jsonld-ld are defined in NPLVocab_parser.py
It can be used as:

import NPLVocab_parse as parser

npl_vocab = parser.get_npl_vocab()
classes = parser.get_all_classes(npl_vocab)
hydra_classes = parser.create_hydra_classes(classes)

The ApiDoc generated by the new script is tested with the postman. and all the tests are passing as well.

We just need to run docwriter.py to create new ApiDoc everytime it will generate classes & properties from JSON-LD.

mock_porfolio_generator.py is also refactored.
creating objects in sequence.

CounterParty --> Loan --> Collateral

[farazkhanfk7]

Changes look great. Also tested NPLVocab_parser.
Suggestion : get_npl_vocab function will give FileNotFoundError for NonPerformingLoan.jsonld. It only works when api_docwriter.py is ran from terminal ( cd : exact file path).
To solve file path issue you can use something like :

from pathlib import Path
from os.path import abspath, dirname

cwd_path = Path(dirname(dirname(abspath(__file__))))
vocab_file_path = cwd_path / "npl_vocab" / "NonPerformingLoan.jsonld"

[open-risk]

I would put the portfolio generator script in a separate directory. In principle this is not a test but a mock client.

[open-risk]

on the vocabulary parser, should there be different treatment of data properties versus object properties? actually not sure what the later (object properties) are mapped into. in any case data properties are the immediate focus

[Purvanshsingh]

@open-risk probably after merging this PR you can remove and put it in a separate directory.
As this commits consists some changes to portfolio generator as well.

[Purvanshsingh]

Object property in our JSON-LD is acting as FOREIGN KEY.
which is handled here :

https://github.com/Purvanshsingh/creditrisk-poc/blob/cebf08b38301f449566fda25f937b40ce8c65f8b/creditrisk_poc/api_doc/NPLVocab_parser.py#L52-L55

For object property the property_uri is the id of the linking class.

[Purvanshsingh]

@open-risk please do approve these changes so that we can merge this PR.

[open-risk]

Object property in our JSON-LD is acting as FOREIGN KEY.
which is handled here :

https://github.com/Purvanshsingh/creditrisk-poc/blob/cebf08b38301f449566fda25f937b40ce8c65f8b/creditrisk_poc/api_doc/NPLVocab_parser.py#L52-L55

That should probably always work if the input ontology is valid. The peculiar case is if the ontology has a class with an object property linking it to another class that has no data properties. Logically it is fine but it will be implemented as a FK to a table that has no data...