search

facebookexperimental / Robyn

Robyn is an experimental, AI/ML-powered and open sourced Marketing Mix Modeling (MMM) package from Meta Marketing Science. Our mission is to democratise modeling knowledge, inspire the industry through innovation, reduce human bias in the modeling process & build a strong open source marketing science community.
https://facebookexperimental.github.io/Robyn/
MIT License
1.07k stars 323 forks source link

[Python API] robyn_refresh endpoint not yet develop? #993

Open rupatrawut opened 2 weeks ago

rupatrawut commented 2 weeks ago

Project Robyn

Describe issue

I would like to do robyn_refresh to builds on top of an selected initial model with new coming data. However, I can't find the Python API endpoint for this function.

Is there any plan to develop this function? Or is it possible to run R function for model_refresh from Python outcome?

Environment & Robyn version

rupatrawut commented 2 weeks ago

FYI, I try to save the model.json which is an input for robyn_refresh in Python using following code:

output_folder = Path(OutputCollect['plot_folder'][0])
for i in OutputCollect['clusters']['models']:
    print(i['solID'])
    write_robynmodel(sol=i['solID'], path=str(output_folder), InputJson=InputCollect, OutputJson=OutputCollect, OutputModels=OutputModels)

But got the MemoryError issue as below:

MemoryError Traceback (most recent call last) Cell In[34], line 6 4 for i in OutputCollect['clusters']['models']: 5 print(i['solID']) ----> 6 write_robynmodel(sol=i['solID'], path=str(output_folder), InputJson=InputCollect, OutputJson=OutputCollect, OutputModels=OutputModels) 8 # OutputCollect['allSolutions']

File ~\workspace\ml-marketing-mix-modeling\ml_mmm_robyn\robyn_api\python_helper.py:301, in write_robynmodel(sol, path, InputJson, OutputJson, OutputModels) 293 payload = { 294 'InputCollect' : json.dumps(InputJson), 295 'OutputCollect' : json.dumps(OutputJson), 296 'OutputModels' : json.dumps(OutputModels), 297 "jsonWriteArgs": json.dumps(writeArgs) 298 } 300 # Get response --> 301 respJson = robyn_api('robyn_write',payload=payload) 302 print('File written to path: ',updatedPath)

File ~\workspace\ml-marketing-mix-modeling\ml_mmm_robyn\robyn_api\python_helper.py:96, in robyn_api(argument, payload, api) 94 return respJson 95 else: ---> 96 response = requests.post(api.format(argument),data=payload) 97 respJson = json.loads(response.content.decode('utf-8')) 98 return respJson

File c:\Users\rupatrawut\workspace\ml-marketing-mix-modeling\conda-env\Lib\site-packages\requests\api.py:115, in post(url, data, json, kwargs) 103 def post(url, data=None, json=None, kwargs): 104 r"""Sends a POST request. 105 106 :param url: URL for the new :class:Request object. (...) 112 :rtype: requests.Response 113 """ --> 115 return request("post", url, data=data, json=json, **kwargs)

File c:\Users\rupatrawut\workspace\ml-marketing-mix-modeling\conda-env\Lib\site-packages\requests\api.py:59, in request(method, url, kwargs) 55 # By using the 'with' statement we are sure the session is closed, thus we 56 # avoid leaving sockets open which can trigger a ResourceWarning in some 57 # cases, and look like a memory leak in others. 58 with sessions.Session() as session: ---> 59 return session.request(method=method, url=url, kwargs)

File c:\Users\rupatrawut\workspace\ml-marketing-mix-modeling\conda-env\Lib\site-packages\requests\sessions.py:575, in Session.request(self, method, url, params, data, headers, cookies, files, auth, timeout, allow_redirects, proxies, hooks, stream, verify, cert, json) 562 # Create the Request. 563 req = Request( 564 method=method.upper(), 565 url=url, (...) 573 hooks=hooks, 574 ) --> 575 prep = self.prepare_request(req) 577 proxies = proxies or {} 579 settings = self.merge_environment_settings( 580 prep.url, proxies, stream, verify, cert 581 )

File c:\Users\rupatrawut\workspace\ml-marketing-mix-modeling\conda-env\Lib\site-packages\requests\sessions.py:486, in Session.prepare_request(self, request) 483 auth = get_netrc_auth(request.url) 485 p = PreparedRequest() --> 486 p.prepare( 487 method=request.method.upper(), 488 url=request.url, 489 files=request.files, 490 data=request.data, 491 json=request.json, 492 headers=merge_setting( 493 request.headers, self.headers, dict_class=CaseInsensitiveDict 494 ), 495 params=merge_setting(request.params, self.params), 496 auth=merge_setting(auth, self.auth), 497 cookies=merged_cookies, 498 hooks=merge_hooks(request.hooks, self.hooks), 499 ) 500 return p

File c:\Users\rupatrawut\workspace\ml-marketing-mix-modeling\conda-env\Lib\site-packages\requests\models.py:371, in PreparedRequest.prepare(self, method, url, headers, files, data, params, auth, cookies, hooks, json) 369 self.prepare_headers(headers) 370 self.prepare_cookies(cookies) --> 371 self.prepare_body(data, files, json) 372 self.prepare_auth(auth, url) 374 # Note that prepare_auth must be last to enable authentication schemes 375 # such as OAuth to work on a fully prepared request. 376 377 # This MUST go after prepare_auth. Authenticators could add a hook

File c:\Users\rupatrawut\workspace\ml-marketing-mix-modeling\conda-env\Lib\site-packages\requests\models.py:559, in PreparedRequest.prepare_body(self, data, files, json) 557 else: 558 if data: --> 559 body = self._encode_params(data) 560 if isinstance(data, basestring) or hasattr(data, "read"): 561 content_type = None

File c:\Users\rupatrawut\workspace\ml-marketing-mix-modeling\conda-env\Lib\site-packages\requests\models.py:129, in RequestEncodingMixin._encode_params(data) 124 for v in vs: 125 if v is not None: 126 result.append( 127 ( 128 k.encode("utf-8") if isinstance(k, str) else k, --> 129 v.encode("utf-8") if isinstance(v, str) else v, 130 ) 131 ) 132 return urlencode(result, doseq=True) 133 else:

MemoryError:

Which is wired because the model.json size when I ran the R code is just 20-30kb.

Environment:

gufengzhou commented 6 days ago

The refresh API endpoint is indeed not shipped yet. @yu-ya-tanaka do you want to have a look and talk to @ebyjose if you have time for this?