Add reported phishing to dataset

app.py

@@ -3,6 +3,7 @@ from fastapi.responses import JSONResponse
 from pydantic import  BaseModel
 from enum import Enum
 from transformers import pipeline
+from phishing_datasets import submit_entry
 
 app = FastAPI()
 
@@ -42,6 +43,7 @@ class OutputModel(BaseModel):
 
 pipe = pipeline(task="text-classification", model="mrm8488/bert-tiny-finetuned-sms-spam-detection")
 
+
 @app.get("/.well-known/apple-app-site-association", include_in_schema=False)
 def get_well_known_aasa():
     return JSONResponse(
@@ -57,9 +59,19 @@ def get_well_known_aasa():
     )
 
 @app.post("/predict")
-def predict(input: InputModel) -> OutputModel:
-    label = pipe(input.query.message.text)
+def predict(model: InputModel) -> OutputModel:
+    text = model.query.message.text
+    label = pipe(text)
     if label[0]['label'] == 'LABEL_1':
+        submit_entry(model.query.sender, model.query.message.text)
         return OutputModel(action=ActionModel.JUNK, sub_action=SubActionModel.NONE)
     else:
-        return OutputModel(action=ActionModel.NONE, sub_action=SubActionModel.NONE)
+        return OutputModel(action=ActionModel.NONE, sub_action=SubActionModel.NONE)
+
+class ReportModel(BaseModel):
+    sender: str
+    message: str
+
+@app.post("/report")
+def report(model: ReportModel):
+    submit_entry(model.sender, model.message)

phishing_datasets.py

@@ -0,0 +1,28 @@
+import pandas as pd
+from datasets import load_dataset, Dataset
+import os
+
+DATASET_NAME = os.getenv("DATASET_NAME")
+
+dataset = load_dataset(DATASET_NAME, split="train")
+df = pd.DataFrame(dataset)
+
+def submit_entry(sender, message):
+    """Adds a new SMS phishing report if it's not already in the dataset."""
+    global df
+
+    sender = sender.strip().replace(" ", "")  # Remove all spaces inside sender
+    message = message.strip()
+
+    # Check for duplicates
+    if ((df["sender"] == sender) & (df["message"] == message)).any():
+        "⚠️ This entry already exists in the dataset!"
+
+    # Append new entry
+    new_entry = pd.DataFrame([[sender, message]], columns=["sender", "message"])
+    df = pd.concat([df, new_entry], ignore_index=True)
+
+    new_dataset = Dataset.from_pandas(df)
+    new_dataset.push_to_hub(DATASET_NAME)
+
+    "✅ Submission saved successfully!"

requirements.txt

@@ -2,4 +2,6 @@ fastapi
 uvicorn[standard]
 pydantic
 transformers
-torch
+torch
+datasets
+pandas