amirziai / sklearnflask Goto Github PK

Flask API for training and predicting using scikit learn models

License: MIT License

Python 82.80% Shell 17.20%

sklearnflask's Introduction

Flask API for scikit learn

A simple Flask application that can serve predictions from a scikit-learn model. Reads a pickled sklearn model into memory when the Flask app is started and returns predictions through the /predict endpoint. You can also use the /train endpoint to train/retrain the model. Any sklearn model can be used for prediction.

Dependencies

scikit-learn
Flask
pandas
numpy

pip install -r requirements.txt

Running API

python main.py <port>

Endpoints

/predict (POST)

Returns an array of predictions given a JSON object representing independent variables. Here's a sample input:

[
    {"Age": 85, "Sex": "male", "Embarked": "S"},
    {"Age": 24, "Sex": "female", "Embarked": "C"},
    {"Age": 3, "Sex": "male", "Embarked": "C"},
    {"Age": 21, "Sex": "male", "Embarked": "S"}
]

and sample output:

{"prediction": [0, 1, 1, 0]}

/train (GET)

Trains the model. This is currently hard-coded to be a random forest model that is run on a subset of columns of the titanic dataset.

/wipe (GET)

Removes the trained model.

sklearnflask's People

Contributors

Stargazers

Watchers

Forkers

jcollingosch toby5box queenb chihchengliang athuldilip dhrg randomm zulfiyagit jayinai nikhiljuturu xtuyaowu emdem bastinrobin huminpurin avinash-k gfilla cringe9000 fdaugs barretthugh darcy-buttrose steccami konark27 drwilliamssteven mallik-g vireshbackup leezqcst smrjans kris330 barbramagidi sigongli xinqiyang silverjimmy wubizhi foryforx jranimo sadi2268 zhangqiking karthiarjun harneet1 alinakhay apparazzi wn9081 sekharvg noskystar laziszaire pobch siosbon kashyap2108 arunrajeie muhammadsamir11 onetesseractinmultiverse amadyba dgreyling cxliuneu miyamotok0105 takahi-i ameyab rizplate pranith101 gidim dikchan jxz542189 catoa yangchih verenich jcardenasrdz robmarkcole xiaoqingwang songkunhuang sethips duannthor rabinandan lee-shining pedrocarvalhodev ptouati snowdj ufukhurriyetoglu kedardg jjaramillom douglaspetrin mirekphd jjwgit atiqahzainal navidre mcmellawatt ritvikkhanna09 gaohongwei unmay amala434 denniz-10 asna1005 xtjjyygy amanb87 huajunatuk jocelynttan sh-leon censhanhe mapmeld rathdebi dangkhoa0894

sklearnflask's Issues

After /wipe (GET), then the /predict (POST) also have a value output

To solve this problem, We should let clf=None, When we call /wipe(GET).
code show as below：
def wipe():
try:
shutil.rmtree('model')
os.makedirs(model_directory)
global clf
clf = None
return 'Model wiped'

except Exception as e:
    print(str(e))
    return 'Could not remove and recreate the model directory'

Thinks for sharing your code, have a good day.

Error converting prediction list to JSON array

Good work Amir, I am trying to run your code but got the following error, I am using python3 and installed all the requirements.
Intial Analysis:
The error is because prediction list contains int64 type values, but jsonifying this list is throwing this error
return jsonify({'prediction': prediction})
Suggested fix
prediction_str = [str(i) for i in prediction] return jsonify({'prediction': prediction_str})
Converting the int64 list to string array solled the error.
Error:
{
"error": "Object of type 'int64' is not JSON serializable",
"trace": "Traceback (most recent call last):\n File "main.py", line 42, in predict\n return jsonify({'prediction': prediction})\n File "C:\Users\vinay\Anaconda3\envs\flask-ml\lib\site-packages\flask\json.py", line 263, in jsonify\n (dumps(data, indent=indent, separators=separators), '\n'),\n File "C:\Users\vinay\Anaconda3\envs\flask-ml\lib\site-packages\flask\json.py", line 123, in dumps\n rv = _json.dumps(obj, **kwargs)\n File "C:\Users\vinay\Anaconda3\envs\flask-ml\lib\json\init.py", line 238, in dumps\n **kw).encode(obj)\n File "C:\Users\vinay\Anaconda3\envs\flask-ml\lib\json\encoder.py", line 201, in encode\n chunks = list(chunks)\n File "C:\Users\vinay\Anaconda3\envs\flask-ml\lib\json\encoder.py", line 430, in _iterencode\n yield from _iterencode_dict(o, _current_indent_level)\n File "C:\Users\vinay\Anaconda3\envs\flask-ml\lib\json\encoder.py", line 404, in _iterencode_dict\n yield from chunks\n File "C:\Users\vinay\Anaconda3\envs\flask-ml\lib\json\encoder.py", line 325, in _iterencode_list\n yield from chunks\n File "C:\Users\vinay\Anaconda3\envs\flask-ml\lib\json\encoder.py", line 437, in _iterencode\n o = _default(o)\n File "C:\Users\vinay\Anaconda3\envs\flask-ml\lib\site-packages\flask\json.py", line 80, in default\n return _json.JSONEncoder.default(self, o)\n File "C:\Users\vinay\Anaconda3\envs\flask-ml\lib\json\encoder.py", line 180, in default\n o.class.name)\nTypeError: Object of type 'int64' is not JSON serializable\n"
}

Add a License?

This is a really cool project, can you add a license to this?

The factor in README causing "Failed to decode JSON object" error.

Thank you very much for sharing your code!
But hey, there's a tiny flaw in your doc which is

.
Little things matter, right?

Error predict

Hi thanks for this repository, for some reason I have the following error:

{
"error": "No objects to concatenate",
"trace": "Traceback (most recent call last):\n File "main.py", line 32, in predict\n query = pd.get_dummies(pd.DataFrame(json_))\n File "C:\Users\AlinaK\Desktop\sklearnflask-master\venv\lib\site-packages\pandas\core\reshape.py", line 1071, in get_dummies\n result = concat(with_dummies, axis=1)\n File "C:\Users\AlinaK\Desktop\sklearnflask-master\venv\lib\site-packages\pandas\tools\merge.py", line 812, in concat\n copy=copy)\n File "C:\Users\AlinaK\Desktop\sklearnflask-master\venv\lib\site-packages\pandas\tools\merge.py", line 845, in init\n raise ValueError('No objects to concatenate')\nValueError: No objects to concatenate\n"
}

Do you know what might it be?

How can this be solved?

reindex columns

hey, thoguht to give you a heads up. When you do this operation:

for col in model_columns:
          if col not in query.columns:
               query[col] = 0

you can actually use the reindex method:
query = query.reindex(columns=model_columns, fill_value=0)