Update app.py

app.py

@@ -7,6 +7,7 @@ from langchain import PromptTemplate
 from langchain import LLMChain
 from langchain_together import Together
 import re
+
 # Set the API key with double quotes
 os.environ['TOGETHER_API_KEY'] = "d88cb7414e4039a84d2ed63f1b47daaaa4230c4c53a422045d8a30a9a3bc87d8"
 
@@ -24,17 +25,17 @@ def YtToQuizz(link, difficulty_level):
     transcript = yt.get_transcript(video_id)
     data = ""
     for text in transcript:
-        data += text.get('text')
+        data += text.get('text') + " "
     summary = Summary_BART(data)
     
     mcq_template = """
-    Give a 10 different multiple-choice question MCQ related to the summary: {summary}
-    The difficulty level of the question should be: {difficulty_level}
-    Please provide the following in:
+    Generate 10 different multiple-choice questions (MCQs) related to the following summary: {summary}
+    The difficulty level of the questions should be: {difficulty_level}
+    Please provide the following for each question:
     1. Question
     2. Correct answer
     3. Three plausible incorrect answer options
-    4. Proper MCQ format
+    4. Format: "Question: <question text>\\nCorrect answer: <correct answer>\\nIncorrect answers: <option1>, <option2>, <option3>"
     """ 
     prompt = PromptTemplate(
         input_variables=['summary', 'difficulty_level'],
@@ -50,22 +51,24 @@ def YtToQuizz(link, difficulty_level):
 
     response_text = response['text']
 
-    # Extract MCQs, correct answers, and options
-    #questions = re.findall(r'Question: (.*?)\n', response_text)
-    #correct_answers = re.findall(r'Correct answer: (.*?)\n', response_text)
-    #options = re.findall(r'Options: (.*?)\n', response_text)
-    #all_options = [option.split(', ') for option in options]
+    # Extract MCQs
+    mcq_pattern = r'Question: (.*?)\nCorrect answer: (.*?)\nIncorrect answers: (.*?)(?:\n|$)'
+    mcqs = re.findall(mcq_pattern, response_text, re.DOTALL)
+
+    if len(mcqs) < 10:
+        return "Failed to generate 10 complete MCQs. Please try again."
 
-    return response_text
+    formatted_mcqs = []
+    for idx, mcq in enumerate(mcqs[:10]):
+        question, correct_answer, incorrect_answers = mcq
+        incorrect_answers = incorrect_answers.split(', ')
+        formatted_mcqs.append(f"Q{idx+1}: {question}\nA) {correct_answer}\nB) {incorrect_answers[0]}\nC) {incorrect_answers[1]}\nD) {incorrect_answers[2]}\n")
+
+    return "\n\n".join(formatted_mcqs)
 
 def main(link, difficulty_level):
-    #questions, options, correct_answers = YtToQuizz(link, difficulty_level)
-    # return {
-    #     "Questions": questions,
-    #     "Options": options,
-    #     "Correct Answers": correct_answers
-    # }
     return YtToQuizz(link, difficulty_level)
+
 iface = gr.Interface(
     fn=main,
     inputs=[
@@ -73,7 +76,7 @@ iface = gr.Interface(
         gr.components.Dropdown(["Easy", "Medium", "Hard"], label="Select difficulty level:")
     ],
     outputs=[
-        gr.components.Textbox(label="MCQs Output")
+        gr.components.Textbox(label="MCQs Output", lines=20)
     ],
     title="YouTube Video Subtitle to MCQs Quiz",
     description="Generate MCQs from YouTube video subtitles"