#Let the random-forest model be: *rf_model*
from kale.common.serveutils import serve
kfserver = serve(rf_model) #model is now being deployed
#prepare data for prediction
data = [row.tolist() for _, row in
train_df[predictor_var].head(10).iterrows()]
data_json = json.dumps({"instances": data})
#prediciton using deployed model:
pred = kfserver.predict(data_json)
Question 1: The returned pred is class labels: 0/1. How to return probabilities?
I tried the following way after studying: kale.common.serveutils.predict
#let HOST be the host name of deployed model
#let the URL of calling the deployed model be: http://xxx:predict
headers = {"content-type": "application/json", "Host": HOST}
pred_2 = requests.post(url = URL, data=data_json, headers=headers)
Question 2: But not clear, where to set parameter, so pred_2 will return probabilities?
Related
I am new to Django so I figure this question could be a bit stupid.I have an api endpoint that returns a list of Doctors(and their details) and another that returns one doctor(and his details)-this is the call am trying to make.I think the issue I am having is with how to ref the pk in the request url.
As it is, when I test on postman I get the error {
"errors": "JSONDecodeError('Expecting value: line 1 column 1 (char 0)',)",
"status": "error"
}
I am almost certain the issue is in api_services.py.I really hope someone can just point it out to me.
views.py
`class FinanceDoctorsView(GenericAPIView):
authentication_classes = [TokenAuthentication]
permission_classes = [IsAuthenticated]
#classmethod
#encryption_check
def get(self, request, *args, **kwargs):
response = {}
pk = kwargs.get("pk")
try:
result = {}
auth = cc_authenticate()
res = getDoctorInfo(auth["key"], pk)
result = res
return Response(result, status=status.HTTP_200_OK)
except Exception as e:
error = getattr(e, "message", repr(e))
result["errors"] = error
result["status"] = "error"
return Response(result, status=status.HTTP_400_BAD_REQUEST)`
api_services.py
import requests
def getDoctorInfo(auth, params):
print("getting doctorInfo from Callcenter")
try:
headers = {
"Authorization": f'Token {auth}'
}
url = f'{CC_URL}/finance/doctor-info/<int:pk>'
res = requests.get(url, headers=headers)
print("returning doctorInfo response", res.status_code)
return res.json()
except ConnectionError as err:
print("connection exception occurred")
print(err)
return err
urls.py
path(
"doctor-info/<int:pk>", views.FinanceDoctorsView.as_view(), name="doctor_info"
),
I think in the api service file, you have made a typo
url = f'{CC_URL}/finance/doctor-info/<int:pk>'
Should had be
# as in the function you have defined params,
# and I think it could have been renamed as pk
url = f'{CC_URL}/finance/doctor-info/{params}'
I have this python script and I want to get Google Script equivalent but I do not know how to "pass" whatever needs to be passed between next get or post request once I log in.
import requests
import json
# login
session = requests.session()
data = {
'LoginName': 'name',
'Password': 'password'
}
session.post('https://www.web.com/en-CA/Login/Login', data=data)
session.get('https://www.web.com//en-CA/Redirect/?page=Dashboard')
# get customer table
data = {
'page': '1',
'pageSize': '100'
}
response = session.post('https://www.web.com/en-CA/Reporting', data=data)
print(response.json())
I wonder if there is an equivalent to .session() object from python's requests module. I did search google but could not find any working example. I am not a coder so I dot exactly know that that .session() object does. Would it be enough to pass headers from response when making new request?
UPDATE
I read in some other question that Google might be using for every single UrlFetchApp.fetch different IP so login and cookies might not work, I guess.
I believe your goal as follows.
You want to achieve your python script with Google Apps Script.
Issue and workaround:
If my understanding is correct, when session() of python is used, the multiple requests can be achieved by keeping the cookie. In order to achieve this situation using Google Apps Script, for example, I thought that the cookie is retrieved at 1st request and the retrieved cookie is included in the request header for 2nd request. Because, in the current stage, UrlFetchApp has no method for directly keeping cookie and using it to the next request.
From above situation, when your script is converted to Google Apps Script, it becomes as follows.
Sample script:
function myFunction() {
const url1 = "https://www.web.com/en-CA/Login/Login";
const url2 = "https://www.web.com//en-CA/Redirect/?page=Dashboard";
const url3 = "https://www.web.com/en-CA/Reporting";
// 1st request
const params1 = {
method: "post",
payload: {LoginName: "name", Password: "password"},
followRedirects: false
}
const res1 = UrlFetchApp.fetch(url1, params1);
const headers1 = res1.getAllHeaders();
if (!headers1["Set-Cookie"]) throw new Error("No cookie");
// 2nd request
const params2 = {
headers: {Cookie: JSON.stringify(headers1["Set-Cookie"])},
followRedirects: false
};
const res2 = UrlFetchApp.fetch(url2, params2);
const headers2 = res2.getAllHeaders();
// 3rd request
const params3 = {
method: "post",
payload: {page: "1", pageSize: "100"},
headers: {Cookie: JSON.stringify(headers2["Set-Cookie"] ? headers2["Set-Cookie"] : headers1["Set-Cookie"])},
followRedirects: false
}
const res3 = UrlFetchApp.fetch(url3, params3);
console.log(res3.getContentText())
}
By this sample script, the cookie can be retrieved from 1st request and the retrieved cookie can be used for next request.
Unfortunately, I have no information of your actual server and I cannot test for your actual URLs. So I'm not sure whether this sample script directly works for your server.
And, I'm not sure whether followRedirects: false in each request is required to be included. So when an error occurs, please remove it and test it again.
About the method for including the cookie to the request header, JSON.stringify might not be required to be used. But, I'm not sure about this for your server.
Reference:
Class UrlFetchApp
You might want to try this:
var nl = getNewLine()
function getNewLine() {
var agent = navigator.userAgent
if (agent.indexOf("Win") >= 0)
return "\r\n"
else
if (agent.indexOf("Mac") >= 0)
return "\r"
return "\r"
}
pagecode = 'import requests
import json
# login
session = requests.session()
data = {
\'LoginName\': \'name\',
\'Password\': \'password\'
}
session.post(\'https://www.web.com/en-CA/Login/Login\', data=data)
session.get(\'https://www.web.com//en-CA/Redirect/?page=Dashboard\')
# get customer table
data = {
\'page\': \'1\',
\'pageSize\': \'100\'
}
response = session.post(\'https://www.web.com/en-CA/Reporting\', data=data)
print(response.json())'
document.write(pagecode);
I used this program
Im having a hard time figuring out how to integrate this ajax request into my view. I'm still learning how to integrate django with ajax requests.
My first question would be: Does the ajax request need to have its own dedicated URL?
In my case I am trying to call it on a button to preform a filter(Preforms a query dependent on what is selected in the template). I have implemented this using just django but it needs to make new request everytime the user preforms a filter which I know is not efficient.
I wrote the most basic function using JQuery to make sure the communication is there. Whenever the user changed the option in the select box it would print the value to the console. As you will see below in the view, I would to call the ajax request inside this view function, if this is possible or the correct way of doing it.
JQuery - Updated
$("#temp").change( function(event) {
var filtered = $(this).val();
console.log($(this).val());
$.ajax({
url : "http://127.0.0.1:8000/req/ajax/",
type : "GET",
data : {
'filtered': filtered
},
dataType: 'json',
success: function(data){
console.log(data)
},
error: function(xhr, errmsg, err){
console.log("error")
console.log(error_data)
}
});
Views.py
def pending_action(request):
requisition_status = ['All', 'Created', 'For Assistance', 'Assistance Complete', 'Assistance Rejected']
FA_status = RequisitionStatus.objects.get(status='For Assistance')
current_status = 'All'
status_list = []
all_status = RequisitionStatus.objects.all()
status_list = [status.status for status in all_status]
# This is where I am handling the filtering currently
if request.GET.get('Filter') in status_list:
user_req_lines_incomplete = RequisitionLine.objects.filter(Q(parent_req__username=request.user) & Q(status__status=request.GET.get('Filter')))
current_status = request.GET.get('Filter')
else:
user_req_lines_incomplete = RequisitionLine.objects.filter(parent_req__username=request.user).exclude(status__status='Completed')
user_reqs = Requisition.objects.filter(par_req_line__in=user_req_lines_incomplete).annotate(aggregated_price=Sum('par_req_line__total_price'),
header_status=Max('par_req_line__status__rating'))
return render(request, 'req/pending_action.html', { 'user_reqs':user_reqs,
'user_req_lines_incomplete':user_req_lines_incomplete,
'requisition_status':requisition_status,
'current_status':current_status,
'FA_status':FA_status})
def filter_status(request):
status = request.GET.get('Filter')
data = {
'filtered': RequisitionLine.objects.filter(Q(parent_req__username=request.user) & Q(status__status=status)),
'current_status': status
}
return JsonResponse(data)
Urls.py
path('pending/', views.pending_action, name='pending_action')
First: you have to divide your template to unchangeable part and the part that you want to modify with your filter.
Second: for your goal you can use render_to_string. See the followning link https://docs.djangoproject.com/en/2.1/topics/templates/#usage
code example (views.py):
cont = {
'request': request, #important key-value
'your_models_instances': your_models_instances
}
html = render_to_string('your_filter_template.html', cont)
return_dict = {'html': html}
return JsonResponse(return_dict)
In your js file you need to determine relative url "{% url 'name in yours url file'%}"
And in success you need to add next line:
success: function(data){
$(".filter-block").html(data.html);
}
i hope it will help you! Good luck!
I am trying to load sample data from a REST API source that return XML inside my emberjs app but I am facing two problems:
The model name is always in plural, so instead of /sqlrest/CUSTOMER/3/ the code always generate /sqlrest/CUSTOMERS/3/
I know that DS.RESTAdaptor expects by default JSON format so I was wondering is there any way I can still get XML format and may be convert to JSON?
Thanks
Code I am using is as follows (This code I found in one of SO replies and altered to match the URL I am trying to access):
App.store = DS.Store.create({
revision: 11,
adapter: DS.RESTAdapter.create({
namespace: "sqlrest",
url: "http://www.thomas-bayer.com",
plurals: {
'customer': 'customer'
},
ajax: function (url, type, hash) {
hash.url = url;
hash.type = type;
hash.dataType = 'jsonp';
hash.contentType = 'application/json; charset=utf-8';
hash.context = this;
if (hash.data && type !== 'GET') {
hash.data = JSON.stringify(hash.data);
}
jQuery.ajax(hash);
},
})
});
and in route:
App.CustomersRoute = Ember.Route.extend({
model: function() {
//return App.Customer.find();
//New
return App.Customer.find(18);
}
});
Maybe you could look at ember-restless which allows XML consumption:
https://github.com/endlessinc/ember-restless
For pluralization, take a look in here:
https://github.com/emberjs/data/blob/v1.0.0-beta.6/packages/ember-data/lib/adapters/rest_adapter.js#L476
The only thing is, obviously if you're going to use ember-restless, you'll need to find the relative point in there that you need to override in a similar way (if it's possible to customize endpoints).
I need to make ajax submit to submit some data include a base64 string of the image, which is render from canvas.
When submit I look in the network panel of Chrome inspector and everything look fine, in "form data" it list all the data that I want to submit.
But in Grails I cannot get the data, there is nothing in the params, just the controller name and action name. Thus everything I get with simple params.dataName is null.
I guess there is something with the size of the post request, but I'm not so sure as I have done this before without ajax.
This is my code for upload with jquery ajax:
var imgBase64String = canvas.toDataURL("image/png");
imgBase64String = imgBase64String .replace('data:image/png;base64,', '');
var submitData = $(form).serializeArray();
submitData.push({name: "webImage", value: imgBase64String })
$.ajax({
type: 'POST',
url: '${createLink(action: 'myAction')}',
data: submitData,
dataType: "html",
success: function(data){//Success code},
});
UPDATE
My code on the server side, it fails at the simple step to retrieve params data:
def myAction= {
def paramData = params
log.info "paramData: " + paramData
def url = params.url
def email = params.email
def webImage = params.webImage
log.info "param: url = " + url
log.info "param: email = " + email
log.info "param: webImage = " + webImage
//Other implement code
}
And the output:
2012-10-08 16:31:28,988 [http-bio-8080-exec-5] INFO myController - paramData: [action:myAction, controller:myController]
2012-10-08 16:31:28,989 [http-bio-8080-exec-5] INFO myController - param: url = null
2012-10-08 16:31:28,989 [http-bio-8080-exec-5] INFO myController - param: email = null
2012-10-08 16:31:28,989 [http-bio-8080-exec-5] INFO myController - param: webImage = null
The size of the base64 image I'm trying to submit is 1998720, don't know if this matter.
Many thanks.
I believe you can simply pass canvas.toDataURL("image/png") into the data field in the $.ajax() method. Also use $.post() instead of $.ajax(). So your code should look like this in the js file:
$.post('/image/getCanvasImage', //this is your url
{
img : canvas.toDataURL('image/jpeg'),
email : email
}, function(data){
//whatever you wanna do with the returned data
}
);
Then in your action, import import sun.misc.BASE64Decoder package and you can write this code to save the canvas image:
def file = params.img.toString().substring((params.img.toString().indexOf(",")+1),params.img.toString().size())
byte[] decodedBytes = new BASE64Decoder().decodeBuffer(file)
def image = new File("mySavedImage.jpg")
image.setBytes(decodedBytes)
Should work!