I'm trying to serve 2 different api's with a different set of methods as documented here: https://cloud.google.com/endpoints/docs/deploying-apis-subdomains
via GCE, that are structured as below in my main.py
#endpoints.api(name = 'first', version = 'v1')
class firstApi(remote.Service):
#endpoints.method(
# Lots of stuff here
path = 'first'
# Lots more here
#endpoints.api(name = 'second', version = 'v1')
class secondApi(remote.Service):
#endpoints.method(
# Lots of stuff here
path = 'second'
# Lots more here
api = endpoints.api_server([firstApi, secondApi])
but when I generate firstv1openapi.json and secondv1openapi.json and deploy them, something strange happens. When I load it up in Google's API explorer, I see both methods under both API's
What is worse, is when I click on the 2nd API, and try any of the methods via the API explorer, the path always references the 1st API.
POST https://my_backend_here.appspot.com/_ah/api/first/v1/second
which as expected fails with a 404.
I've looked at both firstv1openapi.json and secondv1openapi.json to see if the api's cross reference each other at all, and they both look fine.
I'm out of trying different options, any help will be much appreciated.
I see that you're using the Python Endpoints Framework. The Framework currently works only in App Engine's Standard environment, not in GCE.
Related
I've just started on Google Cloud and I'm creating an iOS app to interact with Google Cloud services via a mobile backend. I'm using Python to write the backend for App Engine. I've gone through the tutorials in creating an API based on endpoints - but I have a question.
Do I have to create a Cloud Endpoints API, and then an app on App Engine? Basically, I want to be able to register accounts on my iOS app, call an API which then makes use of Google Datastore to store the account details. From looking at the tutorials (both the cloud endpoints one and then the guestbook one), am I meant to expose Google Datastore, cloud storage etc. within the endpoints api? Or does that link into another app where that is all done?
Sorry if this sounds a bit silly, but I just want to make sure!
Thanks in advance.
In a nutshell, your Cloud Endpoints API is your application. Some of the documentation regarding Cloud Endpoints can be a bit confusing (or vague), but on the server side it's essentially a bunch of Python decorators or Java annotations that allow you to expose your application logic as a REST API.
I find the Java implementation of Cloud Endpoints more intuitive than the Python one, which requires a bit more work to (de-)serialise your objects. You could look at endpoints_proto_datastore.ndb.EndpointsModel which might take some of the boilerplate stuff out of the equation (defining messages).
Essentially, when you write your API, each endpoint maps to a python function. Inside that function you can do what you like, but typically it will be either:
Deserialise your POSTed JSON, validate it, and write some entities to Datastore (or Cloud SQL, BigTable, wherever).
Read one or more entities from Datastore and serialize them to JSON and return them to the client.
For example, you might define your API (the whole collection of endpoint functions) as
#endpoints.api(name='cafeApi', version='v1', description='Cafe API', audiences=[endpoints.API_EXPLORER_CLIENT_ID])
class CafeApi(remote.Service):
# endpoints here
For example, you might have an endpoint to get nearby cafes:
#endpoints.method(GEO_RESOURCE, CafeListResponse, path='cafes/nearby', http_method='GET', name='cafes.nearby')
def get_nearby_cafes(self, request):
"""Get cafes close to specified lat,long"""
cafes = list()
for c in search.get_nearby_cafes(request.lat, request.lon):
cafes.append(c.response_message())
return CafeListResponse(cafes=cafes)
A couple of things to highlight here. With the Python Endpoints implementation, you need to define your resource and message classes - these are used to encapsulate request data and response bodies.
So, in the above example, GEO_RESOURCE encapsulates the fields required to make a GeoPoint (so we can search by location using Search API, but you might just search Datastore for Cafes with a 5-star rating):
GEO_RESOURCE = endpoints.ResourceContainer(
message_types.VoidMessage,
lat=messages.FloatField(1, required=True),
lon=messages.FloatField(2, required=True)
)
and the CafeListResponse would just encapsulate a list of CafeResponse objects (with Cloud Endpoints you return a single object):
class CafeListResponse(messages.Message):
locations = messages.MessageField(CafeResponse, 1, required=False, repeated=True)
where the CafeResponse is the message that defines how you want your objects (typically Datastore entities) serialised by your API. e.g.,
class LocationResponse(messages.Message):
id = messages.StringField(1, required=False)
coordinates = messages.MessageField(GeoMessage, 3, required=True)
name = messages.StringField(4, required=False)
With that endpoint signature, you can access it via an HTTP GET at /cafeApi/v1/cafes/nearby?lat=...&lon=... or via, say, the Javascript API client with `cafeApi.cafes.nearby(...).
Personally, I found Flask a bit more flexible with working with Python to create a REST API.
I have an API endpoint I'm trying to write unit tests for and I can't seem to figure out how to unit test the Python Google Cloud Storage client library calls (https://cloud.google.com/appengine/docs/python/googlecloudstorageclient/).
I was hoping to find a stub somewhere in the library and have it be as simple as unit testing the mail API would be (https://cloud.google.com/appengine/docs/python/tools/localunittesting?hl=en), but haven't found anything yet. Any idea how to go about this?
The list of available unit test does not list GCS. You can file a feature request on their GitHub to add that functionality.
In the mean time using the setUp for your tests to create files is probably your best bet.
Since we have been bitten by severalGoogle API transistions so far (blobstore, blobstore with gs:/, cloudstorage, google-clound-storage) we have long created our own thin wrapper around all GCS access. This also includes stubbing for tests, like this:
def open(path, mode='w', bucket=None, content_type=None):
if not bucket:
bucket = app_identity.get_default_gcs_bucket_name()
jsonpath = '/{}'.format(os.path.join(bucket, path))
jsonpath = jsonpath.replace('*', str(datetime.date.today()))
if os.environ.get(b'GAETK2_UNITTEST'):
LOGGER.info('running unittest, GCS disabled')
return StringIO()
return cloudstorage.open(jsonpath, mode, content_type=content_type)
Still a lot of work if you want to retrofit that on a big application. But might be worth it - the next Google API depreciation will come.
I want to produce a Google Apps document based on a (Google doc) template stored on the users Google Drive and some XML data held by a servlet running on Google App Engine.
Preferably I want to run as much as possible on the GAE. Is it possible to run Apps Service APIs on GAE or download/manipulate Google doc on GAE? I have not been able to find anything suitable
One alternative is obviously to implement the merge functionality using an Apps Script transferring the XML as parameters and initiate the script through http from GAE, but it just seem somewhat awkward in comparison.
EDIT:
Specifically I am looking for the replaceText script functionality, as shown in the Apps script snippet below, to be implemented in GAE. Remaining code is supported through Drive/Mail API, I guess..
// Get document template, copy it as a new temp doc, and save the Doc’s id
var copyId = DocsList.getFileById(providedTemplateId)
.makeCopy('My-title')
.getId();
var copyDoc = DocumentApp.openById(copyId);
var copyBody = copyDoc.getActiveSection();
// Replace place holder keys,
copyBody.replaceText("CustomerAddressee", fullName);
var todaysDate = Utilities.formatDate(new Date(), "GMT+2", "dd/MM-yyyy");
copyBody.replaceText("DateToday", todaysDate);
// Save and close the temporary document
copyDoc.saveAndClose();
// Convert temporary document to PDF by using the getAs blob conversion
var pdf = DocsList.getFileById(copyId).getAs("application/pdf");
// Attach PDF and send the email
MailApp.sendEmail({
to: email_address,
subject: "Proposal",
htmlBody: "Hi,<br><br>Here is my file :)<br>Enjoy!<br><br>Regards Tony",
attachments: pdf});
As you already found out, apps script is currently the only one that can access an api to modify google docs. All other ways cannot do it unless you export to another format (like pdf or .doc) then use libraries that can modify those, then reupload the new file asking to convert to a google doc native format, which in some cases would loose some format/comments/named ranges and other google doc features. So like you said, if you must use the google docs api you must call apps script (as a content service). Also note that the sample apps script code you show is old and uses the deptecated docsList so you need to port it to the Drive api.
Apps script pretty much piggy backs on top of the standard published Google APIs. Increasingly the behaviours are becoming more familiar.
Obviously apps script is js based and gae not. All the APIs apart from those related to script running are available in the standard gae client runtimes.
No code to check here so I'm afraid generic answer is all I have.
I see now it can be solved by using the Google Drive API to export (download) the Google Apps Doc file as PDF (or other formats) to GAE, and do simple replace-text editing using e.g. the iText library
In my case we work with other companies which would consume our APIs along with our internal javascript client. I think we need to create a web client id for javascript client. But when exposing APIs externally, is it correct to generate new web client id per company? If so do we have to update clientid each time and redeploy application?
I'm following this documentation and in their example client ids are hardcoded, if I need to give access to new 3rd party users, then I need to generate new client id for them but I'd expect to not redeploy application.
Update: I've created a feature request as per #Alex's suggestion below.
Unfortunately the docs at https://cloud.google.com/appengine/docs/python/endpoints/auth very specifically say, and I quote,
Because the allowed_client_ids must be specified at build time, you
must rebuild and redeploy your API backend after adding or changing
any client IDs in the authorized list of allowed_client_ids or
audiences
so it appears that your perfectly-reasonable use case is very explicitly not covered at this time.
I recommend you visit said page and enter a feature request via the "Write Feedback" link (around the upper right corner of the page) as well as entering a feature request on the Endpoints component of the App Engine feature tracker, https://code.google.com/p/googleappengine/issues/list?can=2&q=component=Endpoints&colspec=ID%20Type%20Component%20Status%20Stars%20Summary%20Language%20Priority%20Owner%20Log -- we monitor both, but with different processes, so trying both is best.
Sorry to be a bearer of bad news. For now, it seems the only workaround is to distribute to the other companies one of a bunch of client ids generated in advance (you can only change the valid bunch when you re-deploy, sigh) and perhaps add some extra, app-layer authorization check of your own -- exactly the kind of work endpoints should be doing on your behalf:-(.
You can use an asterisk as the client ID, that will allow any client to call it without redeploying your API backend. Not sure if this is a documented feature or not, but it works (at least) with both Python and Java.
#Api(name = "myapi",
version = "v1",
scopes = {"https://www.googleapis.com/auth/userinfo.email"},
description = "My flashy API",
clientIds = {"*"})
public class MyAPI { ... }
The CherryPy web server can supposedly be deployed in the Google App Engine.
Who has done it, and what was the experience like?
What special effort was required (configuration, etc.)?
Would you recommend it to others?
The article is a good example but its slightly out of date now as the patch is no longer required, the latest version of Cherrypy should run without it, I've gotten the sample below running in the development environment.
I've included cherrypy inside a zip file as the google app engine has a limit of one thousand files per application, it also makes it easier to deploy.
I'm also using the cherrypy dispatch handler to route the request.
import sys
sys.path.insert(0, 'cherrypy.zip')
import cherrypy
import wsgiref.handlers
class Root:
exposed = True
def GET(self):
return "give a basic description of the service"
d = cherrypy.dispatch.MethodDispatcher()
conf = {'/':
{
'request.dispatch': d
}
}
app = cherrypy.tree.mount(Root(), "/",conf)
wsgiref.handlers.CGIHandler().run(app)
So far I've not come across any particular issues but I have read some people have had issues with sessions.
See boodebr.org article (missing, but here on the Wayback machine) It works for me.
If you are looking for an example, look for the condition that accepts ServerMode.GAE in ServerInterface.auto in this example.
There is a good article on how to do this over here now here. I haven't actually tried this yet, I stuck with django on App Engine, but it seems to be a solid example.