Add Dockerfile.xgb

Author: christinaexyou

detectors/Dockerfile.xgb

@@ -0,0 +1,32 @@
+FROM registry.access.redhat.com/ubi9/ubi-minimal as base
+RUN microdnf update -y && \
+    microdnf install -y --nodocs \
+        python-pip python-devel && \
+    pip install --upgrade --no-cache-dir pip wheel && \
+    microdnf clean all
+RUN pip install --no-cache-dir torch
+
+# FROM icr.io/fm-stack/ubi9-minimal-py39-torch as builder
+FROM base as builder
+
+COPY ./common/requirements.txt .
+RUN pip install --no-cache-dir -r requirements.txt
+
+COPY ./xgb/requirements.txt .
+RUN pip install --no-cache-dir -r requirements.txt
+
+FROM builder
+
+
+WORKDIR /app
+ARG CACHEBUST=1
+RUN echo "$CACHEBUST"
+COPY xgb/build/model_artifacts /app/model_artifacts
+COPY ./common /common
+
+COPY  ./xgb/build/scheme.py /app
+COPY ./xgb/build/app.py /app
+COPY ./xgb/build/detector.py /app
+
+EXPOSE 8000
+CMD ["uvicorn", "app:app", "--workers", "4", "--host", "0.0.0.0", "--port", "8000", "--log-config", "/common/log_conf.yaml"]

detectors/xgb/build/app.py

@@ -5,12 +5,10 @@
 
 from fastapi import Header
 from prometheus_fastapi_instrumentator import Instrumentator
-
 sys.path.insert(0, os.path.abspath(".."))
 
+from common.app import DetectorBaseAPI as FastAPI
 from detector import Detector
-
-from detectors.common.app import DetectorBaseAPI as FastAPI
 from detectors.common.scheme import (
     ContentAnalysisHttpRequest,
     ContentsAnalysisResponse,

detectors/xgb/build/detector.py

@@ -8,9 +8,9 @@
 from detectors.common.scheme import (
     ContentAnalysisHttpRequest,
     ContentAnalysisResponse,
-    ContentAnalysisHttpResponse,
 )
 import pickle as pkl
+from base_detector_registry import BaseDetectorRegistry
 
 try:
     from common.app import logger
@@ -21,7 +21,7 @@
 class Detector:
     def __init__(self):
         # initialize the detector
-        model_files_path = os.path.abspath(s.path.join(os.sep, "app", "model_artifacts"))
+        model_files_path = os.path.abspath(os.path.join(os.sep, "app", "model_artifacts"))
         if not os.path.exists(model_files_path):
             model_files_path = os.path.join("build", "model_artifacts")
         logger.info(model_files_path)
@@ -39,7 +39,7 @@ def __init__(self):
             self.batch_size = 8
         logger.info("Detector initialized.")
 
-    def run(self, request: ContentAnalysisHttpRequest) -> ContentsAnalysisHttpResponse:
+    def run(self, request: ContentAnalysisHttpRequest) -> ContentAnalysisResponse:
         if hasattr(request, "detection_type") and request.detection_type != "spamCheck":
             logger.warning(f"Unsupported detection type: {request.detection_type}")
 

detectors/xgb/build/train.py

@@ -70,7 +70,7 @@ def preprocess_text(X):
                     subsample=grid_search.best_params_['subsample'],
                     random_state=42
         )
-        clf.fit(X, y)
+        clf.fit(X_vec, y)
 
         print(f"Saving training artifacts to {artifact_path}...")
         pickle.dump(vectorizer, open(f'{artifact_path}/vectorizer.pkl', 'wb'))

detectors/xgb/requirements.txt

@@ -1,7 +1,7 @@
-xgboost==3.0.2
+xgboost
 torch==2.4.0
 pandas==2.2.2
 numpy==1.26.4
 datasets
 nltk==3.9.1
-scikit-learn==1.7.0
+scikit-learn

tests/detectors/xgb/test_xgb.py

@@ -0,0 +1,29 @@
+import pytest
+from fastapi.testclient import TestClient
+
+class TestXGBDetectors:
+    @pytest.fixture
+    def client(self):
+        from detectors.xgb.build.app import app
+        from detectors.xgb.build.detector import Detector
+
+        app.set_detector(Detector(), "detector")
+        return TestClient(app)
+
+    @pytest.mark.parametrize(
+        "content,expected",
+        [
+            (["Congratulations! You've won a $1000 Walmart gift card. Click here to claim now."], True),
+            (["Don't forget to bring your notebook to class tomorrow."], False),
+        ]
+    )
+
+    def test_xgb_detectors(self, client, content, expected):
+        payload = {
+            "content": [content],
+        }
+        resp = client.post("api/v1/text/contexts", json=payload)
+        assert resp.status_code == 200
+        assert len(resp.json()[0]) > 0
+        assert resp.json()[0][0]['spam_check'] == expected
+