stranger6667 / pyoffers Goto Github PK

View Code? Open in Web Editor NEW

16.0 3.0 7.0 423 KB

Python client library for HasOffers API.

Home Page: https://pyoffers.readthedocs.io

License: MIT License

Python 100.00%

pyoffers's Introduction

PyOffers

Python client library for HasOffers API.

Installation

PyOffers can be obtained with pip:

$ pip install pyoffers

Usage example

Initialize API client:

>>> from pyoffers.api import HasOffersAPI
>>> hasoffers = HasOffersAPI(
    endpoint='https://api.hasoffers.com/Apiv3/json',
    network_token='<your_network_token>',
    network_id='<your_network_id>',
)

Execute queries:

   >>> # Get all offers with ID greater than 100, sorted by ID and with loaded `Country` data
   >>> hasoffers.offers.find_all(id__gt=100, sort='id', contain=['Country'])
   [<Offer: 102>,
<Offer: 104>,
<Offer: 106>,
<Offer: 108>,
<Offer: 110>,
<Offer: 112>]
   >>> # Get all clicks records for 2016-09-20
   >>> hasoffers.raw_logs.clicks.find_all('20160920')
   [<LogRecord: 7 (1027a606128bd067105f0b0921840f)>, ...]
   >>> # Get all conversions for specific offer
   >>> offer = hasoffers.offers.get_by_id(100)
   >>> offer.conversions.find_all()
   [<Conversion: 70532>]

Documentation

You can view documentation online at:

https://pyoffers.readthedocs.io

Or you can look at the docs/ directory in the repository.

Python support

PyOffers supports Python 3.5+.

pyoffers's People

Contributors

Stargazers

Watchers

Forkers

astrikov-d tonlam iamanikeev pepperstone jimmytuc seladb ourprofit

pyoffers's Issues

Do not touch cassette without changes

Support for filters

Support for this kind of things without manual building:

&filters[OR][status]=active&filters[OR][currency]=USD

may be something like:

api.conversions.filter_all(filters=Q(status='active') | Q(currency='USD'))

Refactor models instantiation

This is ugly:

if 'pageCount' in data:
    inner_data = data['data']
    result = []
    for item in inner_data.values():
        if 'Advertiser' in item:
            result.append(Advertiser(manager=self.advertisers, **item['Advertiser']))
        elif 'Offer' in item:
            result.append(Offer(manager=self.offers, **item['Offer']))
        elif 'Goal' in item:
            result.append(Goal(manager=self.goals, **item['Goal']))
    return result
elif 'Advertiser' in data:
    return Advertiser(manager=self.advertisers, **data['Advertiser'])
elif 'Offer' in data:
    return Offer(manager=self.offers, **data['Offer'])
elif 'Goal' in data:
    return Goal(manager=self.goals, **data['Goal'])

Better models representation

str(instance) and repr(instance) should show something useful

`fields` feature for `findAll` method

Multiple objects in response

Now it works only with single object. pagination is not handled

Generic methods

There are a lot of methods with same signatures and implementation. There should be a better way to work with them.

E.g. AdvertiserManager:

    def create(self, **kwargs):
        return self._call('create', data=kwargs)

GoalManager:

    def create(self, **kwargs):
        return self._call('create', data=kwargs)

It can be solved with inheritance, but it will lead to very large amount of base classes (I have to check it).

Possible solution - add all these methods to base class and propagate to child only members of generic_methods attribute:

generic_methods = ('find_all', 'create')

Add `conversions` manager for offer instances

Example:

instance = hasoffers.offers.find_by_id(1)
conversions = instance.conversions.find_all()

Should be helpful

Async requests

E.g. for logs, when we have to make a lot of requests to HasOffers API

Move objects instantiation to models

It would be better:

raw will not be required
logic will be separated

contain support for `findAll` queries

Add `findAll` for all available models if it applicable

Offer, Advertiser, Goal

Error when `find_all` returns single result

    119             self._managers[key].init_instance(item)
    120             for chunk in data
--> 121             for key, item in chunk.items()
    122         ]
    123         if len(initialized) == 1 and single:

KeyError: '73'

Add `findAllIds` method

This method returns list of ids, which leads to exception with current implementation

Error if there are no results in not paginated response

Logging before raise exception

https://github.com/Stranger6667/pyoffers/blob/master/pyoffers/api.py#L85

In case of 4XX, 5XX response code we will lose information about content and code of response.
I think that would be better write log before make raise_for_status().

Ability to get raw results from the responses

Add `getTargetCountries` method for Offer model

Multiple parameters feature is not working

prepare_query_params should return list of tuples instead of dict.
Because of this the following query will produce incorrect URL:

key=[1, 2, 3]

Url will contain only last value: key[]=3

In other words produced parameters may have identical keys

Case:
You have 1500 conversions for some offer and you want to grab them all. But you cant get them all in single pyoffers API call, because of max limit is 1000 items.
So, there should be an ability to get everything in one find_all call.
Possibly async request would be helpful here

Remove sensitive data from recorded session

Invalid querystring for lists

E.g.
If field fields has multiple values, then querystring should be:

fields[]=city&fields[]=country

AttributeError in `find_all` if there are no results

        if 'pageCount' in data:
>           return self.init_objects(*data['data'].values())
E           AttributeError: 'list' object has no attribute 'values'

pyoffers/api.py:93: AttributeError

Add tests for python 3.6

It is time to do it

Need check ``verify=False``

https://github.com/Stranger6667/pyoffers/blob/master/pyoffers/api.py#L66

May be it option (verify=False) not required here.

Improve sorting

HasOffers doc has a mistake.
Sorting key should be in the following format:

Target.key

It would be convenient if by default target will be prepended to key

Offer model

Methods:

create
addTargetCountry
addCategory
findById
update

Tests

OR queries

For now it is not supported, but it is easy to implement. But the main question is - how to design the interface?

Credentials cleaning is failing

Because responses from Amazon are not in gzip

Improved 'contain'

It would be better if related instances will be instantiated as attributes of returned object

Cache for `download_link`, `content`, `records` raw logs

VCR or betamax testing

It is hard to work with HasOffers API and save all responses manually

Add `RawLog` model

Methods

getDownloadLink
listDateDirs
listLogs

Not paginated result leads to exception

api.conversions.find_all(filters={'offer_id': 9})

Exception:

    112             self._managers[key].init_instance(item)
    113             for chunk in data
--> 114             for key, item in chunk.items()
    115         ]
    116         if len(initialized) == 1 and single:

KeyError: '55632'

api.offers.find_all(currency='CZK', sort='id', contain=['Country'])

Exception:

     13     def init_instance(self, data):
     14         if 'id' not in data:
---> 15             data = list(data.values())[0]
     16         return super().init_instance(data)

AttributeError: 'list' object has no attribute 'values'

Logs filtration

It is not very convenient to do like this to get log records:

In [2]: dirs = hasoffers.raw_logs.clicks.list_date_dirs()
In [3]: date_dir = dirs[0]
In [4]: logs = date_dir.list_logs()
In [5]: logs[0].records

It should be like this:

records = hasoffers.raw_logs.clicks.find_all(date='2016-09-09 09:00')

create
findAll
findById
update

Tests

Better interface for models

Something

offer.name

instead of

offer['name']

Managers are shared between API instances

Update docs

Add manual for Raw logs