Pydantic models aren't used even though it's hinted as a return type #101
Description
Well, the pydantic models which are defined and hinted as returntypes aren't actually instanciated. Instead most endpoints simply return the result of self__call which is either of dict or list. We can redefine the call-function to something like the following to serialize to a model. A problem which arises is that the test won't work with it's caching out of the box. Another problem, which arises are that some models aren't correctly implemented. For example the SearchResults model.
new __call method
P = TypeVar('P', bound=BaseModel)
def __call(
self, method: HttpMethod, path: str, options=None, return_content: bool = False,
pydantic_model: Optional[Type[P]] = None
) -> Union[P, list, dict, bytes, None]:
method_call = {
HttpMethod.GET: self._session.get,
HttpMethod.POST: self._session.post,
HttpMethod.PUT: self._session.put,
HttpMethod.DELETE: self._session.delete,
}.get(method)
if method_call is None:
raise ValueError(f"Unknown method type {method}")
data = {}
if method == HttpMethod.GET:
data["params"] = options
else:
data["json"] = options
response = method_call(
f"{BASE_URL}{path}",
headers={
"X-AuthenticationSession": self._authentication_session,
"X-SecurityToken": self._security_token,
},
**data,
)
response.raise_for_status()
# Some routes like add/remove instrument from a watch list
# only returns 200 OK with no further data about if the operation succeeded
if len(response.content) == 0:
return None
if return_content:
return response.content
if pydantic_model is None:
return response.json()
return pydantic_model.model_validate_json(response.content)Broken SearchResults
class SearchHit(BaseModel):
changePercent: float
currency: str
""" ISO 4217 """
lastPrice: float
flagCode: str
""" ISO 3166-1 alpha-2 """
tradable: bool
tickerSymbol: str
name: str
id: str
class SearchResult(BaseModel):
instrumentType: str
numberOfHits: int
topHits: List[SearchHit]
class SearchResults(BaseModel):
totalNumberOfHits: int