updating api

.gitignore

@@ -0,0 +1 @@
+__pycache__

Dockerfile

@@ -0,0 +1,16 @@
+# Read the doc: https://huggingface.co/docs/hub/spaces-sdks-docker
+# you will also find guides on how best to write your Dockerfile
+
+FROM python:3.9
+
+RUN useradd -m -u 1000 user
+USER user
+ENV PATH="/home/user/.local/bin:$PATH"
+
+WORKDIR /app
+
+COPY --chown=user ./requirements.txt requirements.txt
+RUN pip install --no-cache-dir --upgrade -r requirements.txt
+
+COPY --chown=user . /app
+CMD ["uvicorn", "api:app", "--host", "0.0.0.0", "--port", "7860"]

api.py

@@ -0,0 +1,21 @@
+import pandas as pd
+from fastapi import FastAPI, Response
+from fastapi.middleware.cors import CORSMiddleware 
+from routers.Prediction.prediction_route import router as predict_router
+
+
+app = FastAPI()
+
+app.add_middleware(
+    CORSMiddleware,
+    allow_origins=["*"],
+    allow_credentials=True,
+    allow_methods=["GET", "POST", "PUT", "DELETE"],
+    allow_headers=["*"],
+)
+
+app.include_router(predict_router)
+
+@app.get('/')
+def _default_router():
+    return Response('Server is running!')

lib/encoder.pkl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4aea61c11ea7f04fa415b58ab1de7afde98b47de958665b19f623fe9e4685f26
+size 1141

lib/model.pkl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:52f8f71d5b5eafa2bdf5921622bb442b5ce3a8708427a151eca3db3487671f45
+size 729566

requirements.txt

@@ -0,0 +1,4 @@
+xgboost
+fastapi
+uvicorn
+pandas

routers/Prediction/prediction_route.py

@@ -0,0 +1,55 @@
+from fastapi import APIRouter
+from pydantic import BaseModel
+from utils.ProcessingClass import PreProcessingClass
+import pickle
+
+class RequestType(BaseModel):
+    MONATSZAHL: str
+    AUSPRAEGUNG: str
+    JAHR: int
+    MONAT: str
+
+router = APIRouter(prefix='/predict')
+
+global encoder, xgb_model
+
+@router.on_event('startup')
+def _loadPickleFiles():
+
+    with open("lib\encoder.pkl", 'rb') as file:
+        global encoder
+        encoder = pickle.load(file)
+
+    with open("lib\model.pkl", 'rb') as file:
+        global xgb_model
+        xgb_model = pickle.load(file)
+
+    print("Pickle Files Loaded. Ready for Inference!")
+
+def _do_inference(df):
+
+    global xgb_model
+
+    return xgb_model.predict(df)
+
+
+@router.post("/")
+def predict(data: RequestType):
+
+    global encoder
+
+    pc = PreProcessingClass(
+        MONATSZAHL = data.MONATSZAHL,
+        AUSPRAEGUNG = data.AUSPRAEGUNG,
+        JAHR = data.JAHR,
+        MONAT = data.MONAT,
+        encoder = encoder
+    )
+
+    date_processed_df = pc._convert_date()
+
+    final_df = pc._one_hot(date_processed_df)
+
+    results = _do_inference(final_df)
+
+    return {"Final Predictions": results.tolist()[0]}

utils/ProcessingClass.py

@@ -0,0 +1,57 @@
+import pandas as pd
+
+class PreProcessingClass:
+    def __init__(self, MONATSZAHL, AUSPRAEGUNG, JAHR, MONAT, encoder):
+
+        self.parent_df = pd.DataFrame({
+            'MONATSZAHL': [MONATSZAHL],
+            'AUSPRAEGUNG': [AUSPRAEGUNG],
+            'JAHR': [JAHR],
+            'MONAT': [MONAT]
+        })
+
+        self.encoder = encoder
+
+    def _convert_date(self, column_name='MONAT', special_value='Summe'):
+
+        day_mapping = {
+            '01': 'January',
+            '02': 'February',
+            '03': 'March',
+            '04': 'April',
+            '05': 'May',
+            '06': 'June',
+            '07': 'July',
+            '08': 'August',
+            '09': 'September',
+            '10': 'October',
+            '11': 'November',
+            '12': 'December'
+        }
+
+        data_copy = self.parent_df.copy()
+        data_copy[column_name] = data_copy[column_name].apply(lambda x: day_mapping[x[4:]] if x != special_value else x)
+
+        return data_copy
+
+    def _one_hot(self, data):
+        
+        columns_to_encode = ['MONATSZAHL', 'AUSPRAEGUNG', 'JAHR', 'MONAT']
+
+        data_copy = data.copy()
+
+        encoded_columns = self.encoder.transform(data[columns_to_encode])
+
+        encoded_column_names = self.encoder.get_feature_names_out(columns_to_encode)
+
+        encoded_df = pd.DataFrame(encoded_columns, columns=encoded_column_names, index=data.index)
+
+        final_df = pd.concat([data.drop(columns=columns_to_encode),encoded_df], axis=1)
+
+        return final_df
+
+    
+
+    
+
+