microservices for spam test and question generation

microservices/app.py

@@ -0,0 +1,109 @@
+from flask import Flask, request, jsonify
+import google.generativeai as genai
+import re,json
+
+# Configure the Flask app
+app = Flask(__name__)
+
+# Configure Gemini API
+GENAI_API_KEY = "AIzaSyD9ZtwHeCQqd4Hdu2549K-xPNEY7G0C8rE"
+genai.configure(api_key=GENAI_API_KEY)
+
+@app.route('/test-spam', methods=['POST'])
+def test_spam():
+    try:
+        # Extract the test message from the POST request body
+        data = request.get_json()
+        if not data or "test_message" not in data:
+            return jsonify({"error": "test_message is required in the request body"}), 400
+
+        test_message = data["test_message"]
+
+        # Construct a prompt for the Gemini API
+        prompt = (
+            f"Please act as a senior spam detector. Based on your knowledge and experience, "
+            f"evaluate the following message for spam: \"{test_message}\". "
+            f"Provide a spam score from 1 to 10, where 1 means not spam and 10 means highly suspicious. "
+            f"Only return the spam score as a number."
+        )
+
+        # Use the Gemini API to generate content
+        model = genai.GenerativeModel("gemini-1.5-flash")
+        response = model.generate_content(prompt)
+
+        # Extract the spam score from the response
+        if response and response.candidates:
+            spam_score_str = response.candidates[0].content.parts[0].text.strip()
+        else:
+            return jsonify({"error": "Invalid response structure from Gemini API"}), 500
+
+        return jsonify({"spam_score": spam_score_str}), 200
+
+    except Exception as e:
+        return jsonify({"error": str(e)}), 500
+
+def parse_questions(response_text):
+    # Extract JSON content from the response_text
+    match = re.search(r"```json\n(.*?)\n```", response_text, re.DOTALL)
+    if not match:
+        raise ValueError("JSON content not found in the response_text.")
+
+    # Parse the JSON content
+    json_content = match.group(1)
+    questions_data = json.loads(json_content)
+
+    # Extract and format questions
+    parsed_questions = []
+    for item in questions_data["questions"]:
+        question = {
+            "question": item["question"],
+            "options": item["options"],
+            "correct_answer": item["correct_answer"]
+        }
+        parsed_questions.append(question)
+
+    return parsed_questions
+
+@app.route('/generate-questions', methods=['POST'])
+def generate_questions():
+    try:
+        # Extract the topic (string_message) from the POST request body
+        data = request.get_json()
+        if not data or "string_message" not in data:
+            return jsonify({"error": "string_message is required in the request body"}), 400
+
+        string_message = data["string_message"]
+
+        # Construct a prompt to generate questions
+        prompt = (
+            f"Act as a well-qualified person on the topic: {string_message}. "
+            f"Help me generate 10 different questions with 4 options for each question. "
+            f"One of the options should be correct, and it should be clear which option is correct."
+            f"The response should be just a plain JSON, nothin more, with given schema:"
+            f"{'{"questions": [{"question": "What is the capital of France?", "options": ["Paris", "London", "Berlin", "Madrid"], "correct_answer": "Paris"}, ...]}'}"
+            f"Please provide the questions in the format specified above."
+            f"you need to generate questions based on the topic provided."
+        )
+
+        # Use the Gemini API to generate content
+        model = genai.GenerativeModel("gemini-1.5-flash")
+        response = model.generate_content(prompt)
+
+        # Extract the response text
+        if response and response.candidates:
+            response_text = response.candidates[0].content.parts[0].text.strip()
+        else:
+            return jsonify({"error": "Invalid response structure from Gemini API"}), 500
+
+        # import pdb; pdb.set_trace()
+        # Parse the questions from the response text
+        questions = parse_questions(response_text)
+
+        # Return the questions in the desired format
+        return jsonify({"questions": questions}), 200
+
+    except Exception as e:
+        return jsonify({"error": str(e)}), 500
+
+if __name__ == '__main__':
+    app.run(debug=True)