addition of preprocessing endpoint

app.py

@@ -44,6 +44,10 @@ app = FastAPI(
 class PredictionFeatures(BaseModel):
     Text: str
 
+class PreprocessingFeatures(BaseModel):
+    tweet: str
+
+
 @app.get("/", tags=["Introduction Endpoints"])
 async def index():
     """
@@ -117,3 +121,37 @@ async def predict(predictionFeatures: PredictionFeatures):
 
     # Format and return the response
     return result
+
+
+
+@app.post("/preprocess", tags=["Machine Learning"])
+async def predict(preprocessingFeatures: PreprocessingFeatures):
+    """
+    This method will preprocess a raw tweet.This intermediate method is used as the preprocessing cannot be simply included in the prediction model
+ 
+    ### Input
+    - `preprocessingFeatures` (PreprocessingFeatures): An object containing the tweet to be preprocessed.
+      - `predictionFeatures` is a dictionnary with 'tweet' as only key
+      - The input text is provided as a string as a value of the 'tweet' key
+
+    ### Output
+    Returns a dictionary with the following keys:
+    - `tweet` (str): Initial tweet.
+    - `text_clean` (str): Preprocessed tweets after removal of punctation and stop words and text lemmatization .
+
+    """
+
+    # Convert input into a DataFrame
+    list_text = [preprocessingFeatures.tweet]
+
+    # Load model from MLflow
+    logged_model = 'runs:/08b34287ab0248f6b199121bff56fc86/text_preprocessor'
+    loaded_model = mlflow.pyfunc.load_model(logged_model)
+
+    df = pd.DataFrame(list_text,columns=['tweet'])
+
+    # Perform prediction
+    preprocessed_result = loaded_model.predict(pd.DataFrame(df))
+
+    # Format and return the response
+    return preprocessed_result