Add field usage as post endpoint #14
Conversation
| @@ -62,9 +93,17 @@ async def generate_csv(request, query: str, businessunitid: str): | |||
| async def inner(request, *old_args, **route_parameters): | |||
| kwargs = {} | |||
| name = None | |||
| if legacy or request.method == "GET": | |||
| parameters = request.args | |||
There was a problem hiding this comment.
what about both query args and json_body args ?
in this way it seems you cant have both?
i suggest:
parameters = request.args
try:
body_parameters = json.loads(request.body)
body_parameters.update(parameters) # query.args are higher priority than body params
parameters = body_parameters
except json.decoder.JSONDecodeError:
pass
else you loose the ability to:
@route("/age//lol")
@parse_params(request, age:int, country: str, name: str)
with request being:
url= /age/21/lol?country=dk
body={"name": "lolcat"}
There was a problem hiding this comment.
Yes I'll post an issue related to this saying it isn't currently implemented and if someone wants to do it. they can have a go at it.
There was a problem hiding this comment.
@sallas remember that not supprting path params together with json parameters atleast breaks your api's cus then you cant get businessunitid no more
There was a problem hiding this comment.
Good point. Path parameters must be supported in both cases. The intention was to not implement support for query params in a POST request
There was a problem hiding this comment.
i think the easiest is support path+query+json and maybe take inspiration in my comments.
and then maybe focus on priority:
- path
- query
- json
so path is most important, json least impotant
so big warning: dont use same names in path+query+json cuz then trouble.
but supporting all should make less complicated code i think
@jorgenpersson @sallas
There was a problem hiding this comment.
| for name, arg_type, default in parameters: | ||
| raw_value = request.args.get(name, None) | ||
| for name, arg_type, default in func_parameters: | ||
| raw_value = parameters.get(name, None) |
There was a problem hiding this comment.
maybe do something like:
query_parameters = request.args
json_parameters = json.loads(request.body)
...
NOTHING = object()
raw_value_origin = None:
raw_value = NOTHING
query_arg_value = query_parameters.get(name, None)
json_param_value = json_parameters.get(name, None)
if query_arg_value is not None:
raw_value_origin = "query"
raw_value = query_arg_value
elif json_param_value is not None:
raw_value_origin = "json"
raw_value = json_arg_value
cuz then you know where you got it from and the
elif name not in parameters:
can now be:
elif raw_value is NOTHING:
and the:
raise KeyError(
f"Missing required {'argument' if legacy else 'parameter'} {name}"
)
can now be:
raise KeyError(
f"Missing required {raw_value_origin} parameter {name}"
)
There was a problem hiding this comment.
I'll link to both of these comments in the issue. This could be enough for someone to figure out how to proceed. For now I don't want to spend the time taking a decision in something that seems like a half way there solution.
However, if we could somehow specify which arg is in the query vs in the body, that would be my preferred solution.
There was a problem hiding this comment.
👍
i could imagine something like:
def my_handler(request, varname: List[int])
would implicitly use the priority hierachy (so first path, then query, then json, (if name clashing)
and then give the following option to specify from where:
from sanicargs import types
def my_handler(request, varname: types.path.List[int])
or
def my_handler(request, varname: types.query.List[int])
or
def my_handler(request, varname: types.json.List[int])
then we can differentiate.
i will make a pr at some point on that issue
Pull request opened by github-pullrequestcreator.