The old_question_solution_using_openai_api from aananda-giri

old_question_solution_using_openai_api's Introduction

Introduction

A project developed for Stranger Hacks hackathon
Demo
Tagline: Old questions solution

working

  * Enter url to old exam questions
  * It will give you questions and corresponding answers

Running Code:

Install Requirements: pip install -r requirements.txt
Install tesseract-ocr
Get OpenAI API Key from here and put it in file .env as:

# .env
OPENAI_API_KEY="<API-KEY>"

Run python main.py

urls:

* home: http://127.0.0.1:8001
* mannual questions: http://127.0.0.1:8001/mannual

Internal Working

* Download pdf from url given
* Convert pdf to images
* Extract text from images using `pytesseract`
* Extract questions from `pytesseract` output using `openai-api`
* Extract answers of questions output using `openai-api`
* Store Answers and put link to question_answers in home page

problems:

* openAI API Rate Limit: `RateLimitError: You exceeded your current quota, please check your plan and billing details.`

Stored Data Format

 * `qa.json` format
 * using random_id to allow multiple qa with same title
 * todo: store to database

        {
                "random_id": 
                [
                    {
                        "url":<url>,
                        "title":<title>",
                        "qa": [
                                {
                                    "question": "question1",
                                    "answer" : "answer1"
                                },
                                {
                                    "question": "question2",
                                    "answer": "answer2"
                                },
                    },

                    {
                        "url":"<url2>",
                        
                    },
                    
                ],
                
                "random_id2": 
                [
                    {
                        "url":"<url>",
                        "title":"<title>",
                        "qa": [
                                {
                                    "question": "question1",
                                    "answer" : "answer1"
                                },
                                {
                                    "question": "question2",
                                    "answer": "answer2"
                                },
                    },

                    {
                        "url":"<url2>",
                        
                    },
                    
                ],
            ],

        }