How to mock elastic search python?

I'm going to give a very abstract answer because this applies to more than ES.

class ProductionCodeIWantToTest:  def __init__(self):    pass  def do_something(data):    es = ES() #or some database or whatever    es.post(data) #or the right syntax

Now I can't test this.With one small change, injecting a dependency:

class ProductionCodeIWantToTest:  def __init__(self, database):    self.database = database  def do_something(data):    database.save(data) #or the right syntax

Now you can use the real db:

es = ES() #or some database or whateverthing = ProductionCodeIWantToTest(es)

or test it

mock = #... up to you - just needs a save method so farthing = ProductionCodeIWantToTest(mock)

After looking at the decorator source code, the trick for me was to reference Elasticsearch with the module:

import elasticsearch...elasticsearch.Elasticsearch(...

instead of

from elasticsearch import Elasticsearch...Elasticsearch(...

You have to mock the attr or method you need, for example:

import mockwith mock.patch("elasticsearch.Elasticsearch.search") as mocked_search, \                mock.patch("elasticsearch.client.IndicesClient.create") as mocked_index_create:            mocked_search.return_value = "pipopapu"            mocked_index_create.return_value = {"acknowledged": True}

In order to know the path you need to mock, just explore the lib with your IDE. When you already know one you can easily find the others.