""" NOTE: this file represents the schema parsing logic for jsonschema_gentypes

"""

import logging

import contextlib

import re

from typing import Dict

from jsonschema_gentypes.cli import process_config

from jsonschema_gentypes import configuration

import referencing

import tempfile

import json

import referencing.exceptions

from polyapi.constants import JSONSCHEMA_TO_PYTHON_TYPE_MAP

def _cleanup_input_for_gentypes(input_data: Dict):

""" cleanup input_data in place to make it more suitable for jsonschema_gentypes

"""

for k, v in input_data.items():

if isinstance(v, dict):

_cleanup_input_for_gentypes(v)

elif k == "enum":

# jsonschema_gentypes doesn't like double quotes in enums

# TODO fix this upstream

for idx, enum in enumerate(v):

if isinstance(enum, str):

v[idx] = enum.replace('"', "'")

def _temp_store_input_data(input_data: Dict) -> str:

"""take in the input data and store it in a temporary json file"""

with tempfile.NamedTemporaryFile(

mode="w", delete=False, prefix="polyapi_", suffix=".json"

) as temp_file:

json.dump(input_data, temp_file)

return temp_file.name

def wrapped_generate_schema_types(type_spec: dict, root, fallback_type):

from polyapi.utils import pascalCase

if not root:

root = "List" if fallback_type == "List" else "Dict"

if type_spec.get("x-poly-ref") and type_spec["x-poly-ref"].get("path"):

# x-poly-ref occurs when we have an unresolved reference

# lets name the root after the reference for some level of visibility

root += pascalCase(type_spec["x-poly-ref"]["path"].replace(".", " "))

else:

# if we have no root, just add "My"

root = "My" + root

root = clean_title(root)

try:

return root, generate_schema_types(type_spec, root=root)

except RecursionError:

# some schemas are so huge, our library cant handle it

# TODO identify critical recursion penalty and maybe switch underlying logic to iterative?

return fallback_type, ""

except referencing.exceptions.CannotDetermineSpecification:

# just go with fallback_type here

# we couldn't match the right $ref earlier in resolve_poly_refs

# {'$ref': '#/definitions/FinanceAccountListModel'}

return fallback_type, ""

except:

logging.warning(f"WARNING parsing jsonschema failed: {type_spec}\nusing fallback type '{fallback_type}'")

return fallback_type, ""

def generate_schema_types(input_data: Dict, root=None):

"""takes in a Dict representing a schema as input then appends the resulting python code to the output file"""

_cleanup_input_for_gentypes(input_data)

tmp_input = _temp_store_input_data(input_data)

tmp_output = tempfile.NamedTemporaryFile(

mode="w", delete=False, prefix="polyapi_", suffix=".py"

).name

config: configuration.Configuration = {

"python_version": None, # type: ignore

"generate": [

{

"source": tmp_input,

"destination": tmp_output,

"root_name": root,

"api_arguments": {"get_name_properties": "UpperFirst"},

}

],

}

# jsonschema_gentypes prints source to stdout

# no option to surpress so we do this

with contextlib.redirect_stdout(None):

process_config(config, [tmp_input])

with open(tmp_output, encoding='utf-8') as f:

output = f.read()

output = clean_malformed_examples(output)

return output

# Matches commented example headers emitted by jsonschema-gentypes before a raw

# multiline JSON object/array body that is not commented out.

MALFORMED_EXAMPLE_HEADER_PATTERN = re.compile(

r"^\s*#\s*(?:\|\s*)?example:\s*([\[{])\s*$"

)

# Regex to fix invalid escape sequences in docstrings

INVALID_ESCAPE_PATTERNS = [

# Fix "\ " (backslash space) which is not a valid escape sequence

(re.compile(r'\\(\s)', re.DOTALL), r'\1'),

# Fix other common invalid escape sequences in docstrings

(re.compile(r'\\([^nrtbfav"\'\\])', re.DOTALL), r'\\\\\1'),

]

def clean_malformed_examples(example: str) -> str:

""" there is a bug in the `jsonschmea_gentypes` library where if an example from a jsonchema is an object,

it will break the code because the object won't be properly commented out. Also fixes invalid escape sequences.

"""

cleaned_lines = []

balance = 0

skipping_example = False

for line in example.splitlines(keepends=True):

if not skipping_example:

if MALFORMED_EXAMPLE_HEADER_PATTERN.match(line):

skipping_example = True

balance = line.count("{") + line.count("[") - line.count("}") - line.count("]")

continue

cleaned_lines.append(line)

continue

balance += line.count("{") + line.count("[") - line.count("}") - line.count("]")

if balance <= 0:

skipping_example = False

cleaned_example = "".join(cleaned_lines)

# Fix invalid escape sequences in docstrings

for pattern, replacement in INVALID_ESCAPE_PATTERNS:

cleaned_example = pattern.sub(replacement, cleaned_example)

return cleaned_example

def clean_title(title: str) -> str:

""" used by library generation, sometimes functions can be added with spaces in the title

or other nonsense. fix them!

"""

title = title.replace(" ", "")

# certain reserved words cant be titles, let's replace them

if title == "List":

title = "List_"

return title

def map_primitive_types(type_: str) -> str:

# Define your mapping logic here

return JSONSCHEMA_TO_PYTHON_TYPE_MAP.get(type_, "Any")

def is_primitive(type_: str) -> bool:

return type_ in JSONSCHEMA_TO_PYTHON_TYPE_MAP