submit

Dockerfile

@@ -0,0 +1,20 @@
+# 
+FROM python:3.9
+
+# 
+WORKDIR /code
+
+# 
+COPY ./Pipfile /code/Pipfile
+
+COPY ./model_iLab_case.pkl /code/model_iLab_case.pkl
+
+#
+RUN pip install pipenv 
+RUN pipenv install
+
+# 
+COPY ./reco /code/reco
+
+# 
+CMD ["pipenv", "run", "uvicorn", "reco.api:app", "--host", "0.0.0.0", "--port", "80", "--reload"]

Pipfile

@@ -0,0 +1,15 @@
+[[source]]
+url = "https://pypi.org/simple"
+verify_ssl = true
+name = "pypi"
+
+[packages]
+fastapi = "*"
+uvicorn = "*"
+pydantic = "*"
+gensim = "==3.8.3"
+
+[dev-packages]
+
+[requires]
+python_version = "3.9"

README.md

@@ -0,0 +1,16 @@
+# DOCKERIZED WORD2VEC RECOMMENDATION SYSTEM MODEL
+
+About Running:
+
+```sh
+sudo docker build -t test .
+```
+```sh
+sudo docker run --net=host test
+```
+
+POSTMAN;
+
+POST REQUEST URL : http://0.0.0.0:80/predict 
+
+Example Request JSON Body; {"product_id" : "a368901216"}

reco/api.py

@@ -0,0 +1,35 @@
+from fastapi import FastAPI, Depends
+from pydantic import BaseModel
+from typing import List, Tuple
+
+from .model import get_model, Model
+
+app = FastAPI()
+
+
+class RecoRequest(BaseModel):
+    product_id: str
+
+
+class RecoResponse(BaseModel):
+    recos: List[List]
+# Senaryo2: RecoResponse class'i yoruma alinir.
+
+@app.post("/predict", response_model=RecoResponse)
+# Senaryo2: response_model = List[Tuple[str,float]] seklinde verilir. Return reco yapilir.
+def predict(request: RecoRequest, model: Model = Depends(get_model)):
+    reco = model.predict(request.product_id)
+    return RecoResponse(
+        recos=reco
+    )
+
+## DIGER CALISAN YOL ##
+
+"""
+class RecoResponse silinir.
+
+@app.post("/predict", response_model=List[Tuple[str,float]]) 
+def predict(request: RecoRequest, model: Model = Depends(get_model)):
+    reco = model.predict(request.product_id)
+    return reco
+"""

reco/model.py

@@ -0,0 +1,21 @@
+from gensim.models import Word2Vec
+import pickle
+
+
+class Model:
+    def __init__(self):
+        with open('model_iLab_case.pkl', 'rb') as f:
+            model_loaded = pickle.load(f)
+
+        self.model = model_loaded
+
+    def predict(self, product_id):
+        result = self.model.wv.most_similar(product_id, topn=15)
+        return result
+
+
+model = Model()
+
+
+def get_model():
+    return model