Table of Contents
- Introducing WebGPT-4 Anki Builder
- Why should you choose WebGPT-4 Anki Builder?
- Quick Start
- Check out the video!
- Sucessful usecases
- Installation
- Docker Deployment for Web Application
- Code Documentation
- How to Contribute
- Future RoadMap
- Contact us(Contributors)
- Project's Funding
- Acknowledgements
- Recommended citation
WebGPT-4 Anki Builder takes the innovative concept of Auto Anki with ChatGPT 4 and elevates it to new heights, introducing a dynamic web application platform. This advanced tool seamlessly turns various types of educational content, including web pages, into highly effective and precise flashcards, driven by the power of ChatGPT 4 and GPT-4's URL processing capabilities.
Designed for students, educators, and lifelong learners, WebGPT-4 Anki Builder is your go-to solution for transforming complex information into manageable study aids. Whether you're handling dense academic materials or seeking to grasp intricate concepts, our application is engineered to streamline your learning process.
The WebGPT-4 Anki Builder not only extracts crucial concepts from your lectures to generate targeted questions, ensuring a thorough grasp of the material, but also searches the web for the most relevant answers to guarantee the accuracy of your flashcards. It leverages Anki's robust features to allow users to rate the difficulty level of each question, aiding in the memorization of challenging concepts.
Building upon this, the application now enables the creation of flashcards directly from web content using GPT-4 integration, capturing the essence of diverse subject matters. It offers customizable flashcard creation, empowering users to tailor the number of flashcards to their learning needs. The responsive web design ensures a seamless user experience across various devices. With real-time status updates, learners stay informed throughout the flashcard creation process, and enhanced error reporting with informative pop-ups guides them through any issues encountered. Finally, the upgraded GPT-4 integration enhances the relevance and accuracy of the flashcards, enriching the learning experience.
Picture this: you're faced with an ocean of educational content, daunting and seemingly impenetrable. Then, along comes WebGPT-4 Anki Builder, a beacon of clarity in this sea of information. It's not just a study tool; it's your personal academic ally, transforming complex lecture materials into easily digestible flashcards. With the power of GPT-4, it sifts through content, pinpointing and extracting key concepts, and tailoring these into a format that makes learning not just effective, but enjoyable.
This tool is a breakthrough in adaptive learning. Whether cramming for exams, revisiting old topics, or delving into new ones, WebGPT-4 Anki Builder molds itself to fit your unique learning needs. It offers a responsive, seamless study experience across all devices, complemented by real-time updates that keep you informed of your progress. The real magic lies in its precision – enhanced error reporting and the accuracy of GPT-4 integration ensure that what you're studying is not just relevant but also meticulously curated for your success.
Dive into the world of WebGPT-4 Anki Builder, where your study sessions turn into a knowledge fiesta, and learning feels like unlocking cheat codes for your brain! 🚀🧠🎉
Step into the future with WebGPT-4 Anki Builder: Just upload your study materials or feed in a URL, and voilà – your personalized APKG file is crafted in a flash! Import it into Anki and make studying not just a breeze, but a thrilling adventure. Quick, effortless, and smart – it's learning, reimagined! 🌟📚💡
After you run the application:
-
The above two represents: 1. The updated Tkinter GUI 2. Web Application of the Auto Anki
-
How to use Tkinter GUI
-
For file upload:
- Choose the API source from the dropdown (Google or GPT-4)
- Enter the number of flashcards you want to generate(default: 10)
- Click on browse files to select the file you want to process (Type: PDF or DOCX)
- After selecting the file, it will be processed and apkg file is generated in anki-decks folder
-
For URL processing:
- Paste the url of the website you want to generate anki cards on, in the url input field
- Enter the number of flashcards you want to generate
- Click on the -> button to process the url
- Now, it will be processed and apkg file is generated in anki-decks folder
-
Take the apkg file and upload in anki software to view the ankicards
-
-
How to use Web Application
-
The toggle by default will be File
- For file upload, Choose the API source from the dropdown (Google or GPT-4)
- Enter the number of flashcards you want to generate
- Click on Choose File to select the file you want to process (Type: PDF or DOCX)
- After choosing the file, click on Process File
- After the status changes from Ready to File is processed, click on Download APKG File*
- The apkg file can be found in your downloads
-
Swithc toggle to URL for url processing
- Enter the url of the website you want to generate anki cards on, in the url input field
- Enter the number of flashcards you want to generate
- Click on Process URL
- After the status changes from Ready to File is processed, click on Download APKG File*
-
Take the apkg file and upload in anki software to view the ankicards
-
With the Anki application, you can open the apkg file generated and enjoy a new and enhanced studying experience.
Here's a series of pictures to demonstrate the features we've added:
Our latest update introduces a revolutionary URL processing feature. Now, users can create flashcards directly from any website. This enhancement is powered by the integration of GPT-4, ensuring highly accurate and relevant content extraction. This is a significant leap from the previous version, which was limited to generating flashcards only from PDF or DOCX files.
Flexibility is key in our new update. Users now have the option to specify the number of flashcards they wish to generate. This new feature contrasts with the earlier version's fixed flashcard output, offering a more tailored learning experience.
We've overhauled the user interface, transitioning from a non-responsive tkinter-based GUI to a fully responsive design. This update enhances user interaction, making the application more adaptable to various screen sizes and user needs.
Our new status bar provides real-time updates on the application's state. It indicates when the GUI is ready, when files are being processed, and when the processing is complete. This feature eliminates the need to manually check for file generation, streamlining the user experience.
In the screenshots, we can see the status bar shows the status of the application after the file is uploaded
-
- Processing file...
-
- File processed successfully.
To further aid user understanding, we've implemented a message box that displays errors or confirms successful file generation. This feature provides clear, immediate feedback about the application's status and any issues encountered during file processing, a significant improvement over the previous version's lack of in-app updates.
In the below example, we removed the API KEY and tried to run the url processing:
In response to the latest changes in OpenAI's API, we've completely rewritten our GPT integration, upgrading to the cutting-edge GPT-4 model. This overhaul was crucial not only for adhering to new API standards but also for significantly boosting the precision of flashcard generation. Our team re-engineered this aspect from the ground up, ensuring smooth and efficient conversion, thereby enhancing the overall functionality and user experience of our learning tool.
Elevating our toolkit to new heights, we proudly introduce the seventh major update — a shift to a dynamic web application, enhancing user engagement far beyond what the previous tkinter GUI could offer. This new web interface boasts a dynamic background, a toggle button for effortless switching between file and URL processing, and a direct download option for the generated APKG files post-processing. It's a comprehensive upgrade that not only makes the application more visually appealing and interactive but also significantly enriches the user's experience with its streamlined and intuitive functionality.
In a stride toward universal accessibility and ease of deployment, we've encapsulated the web application ui for anki builder within a Docker container. By placing the Docker image on Docker Hub, we've dramatically reduced build overhead and ensured cross-platform compatibility. Now, users from any environment can spin up WebGPT-4 Anki Builder with minimal setup, opening doors to a hassle-free, consistent flashcard creation experience across any platform. 🐳
ODFYETZ44INYKVH7.mp4
🎉️ 🎉️ Case Study: Web Development Mastery
User: Matthew
Background: Aspiring Web Developer
Scenario:
- Matthew, an aspiring web developer, found himself overwhelmed with the sheer volume of coding concepts and frameworks he needed to master.
- He turned to WebGPT-4 Anki Builder for help, using it to generate flashcards from various online programming tutorials and documentation.
- Matthew utilized the URL processing feature to directly create cards from web resources, making his study sessions more efficient and focused.
Result:
- In just a few weeks, Matthew observed a remarkable improvement in his understanding of complex web development concepts.
- He successfully built his first full-stack web application, crediting the clarity and retention provided by WebGPT-4 Anki Builder for his accelerated learning curve.
🎉️ 🎉️ Case Study: Mastering Kubernetes and Flask
User: Varun
Background: Computer Science Major
Scenario:
- Varun, a computer science student, was eager to learn Kubernetes and Flask to deploy his web application efficiently.
- Facing the complexity of Kubernetes and the nuances of Flask, Varun turned to WebGPT-4 Anki Builder for a streamlined learning approach.
- He used the tool to create flashcards from Kubernetes documentation and Flask tutorials, transforming extensive technical content into bite-sized, manageable learning segments.
Result:
- With the help of WebGPT-4 Anki Builder, Varun quickly grasped the intricacies of container orchestration with Kubernetes and web development with Flask.
- His newfound proficiency led to the successful deployment of his web application, showcasing a robust backend managed by Flask and scalable deployment handled by Kubernetes.
- Clone the repository
git clone https://github.com/smarabattula/auto_anki - Set up a virtual environment (Optional, but highly recommended):
- Create the virtual environment:
python -m venv myenv - Activate the virtual environment:
source myenv/bin/activate(for MacOS) andsource myenv/Scripts/activate(for Windows). - Note: In future runs, you won't need to create the virtual environment again, just activate it.
- Create the virtual environment:
- Install all required packages:
pip install -r requirements.txt - Download a required model:
python -m spacy download en_core_web_lg - Add the Anki library as submodule in your project folder
git clone https://github.com/kerrickstaley/genanki - Navigate into the cloned directory and Install the Anki library:
cd genanki; python setup.py install - Navigate to Project folder again The command to run the project is:
- Web Application:
python3 code/ui1.py - Tkinter GUI :
python3 code/ui.py
- Web Application:
- Install the project as a Python package(Optional):
pip install . - Create .env file and insert
API_KEY='YOUR_OPENAI_KEY_FOR_GPT-4'
-
Create Docker Account (
https://www.docker.com/) -
Install Docker Dektop (
https://www.docker.com/products/docker-desktop/) -
In Search bar at top of docker desktop application search for
auto_anki -
Pull
msills23/auto_ankiimage -
Click the run action on the image
-
Expand the optional settings and enter
5000for the port number -
Run the new container
-
Go to containers tab and click on the square with an arrow through it to the right of the port numbers
This will open the GUI in your web browser
Documentation of the entire codebase is generated using Pycco. You can find the documentation here.
If you are a developer, and want to update documentation:
- Install Pycco
pip install pycco - Use Pycco to generate docs.
pycco auto_anki/**/*.py -p
For checking code coverage,
- Install Coverage (https://pypi.org/project/coverage/)
pip install coverage - For generating the report run
cd codepython -m coverage run -m unittest discover - For viewing the report run
python -m coverage report
| Name | Stmts | Miss | Cover |
|---|---|---|---|
| anki.py | 17 | 1 | 94% |
| extract_sizes.py | 49 | 16 | 67% |
| user_cli.py | 34 | 2 | 94% |
| wordprocessing.py | 125 | 4 | 97% |
| TOTAL | 238 | 23 | 90% |
We would be happy to receive contributions! If you'd like to, please go through our CONTRIBUTING.md.
For any feedback, issues, or bug reports, please create an issue here.
- Create an interactive flashcard viewer within the web application, allowing users to review and study the flashcards directly on the site.
- Implement features such as flipping the card for answers, progressing to the next card, and random shuffle.
- Introduce machine learning algorithms that adapt to individual user learning patterns and preferences.
- Provide personalized recommendations and flashcard adjustments based on user performance.
We are here to support you. Feel free to email us with any question or bug reports. We try our best to reply as soon as possible.
- Sasank Marabattula ([email protected])
- Varun Deepak Gudhe ([email protected])
- MatthewSills ([email protected])
Our project is currently not funded, and we operate on a volunteer and open-source basis , and currently,improvement of the project solely relies on the dedication of our team and contributions from the open-source community.
We have built this code on top of the stack from the project https://github.com/tran4code/auto_anki
You can cite us like this:
WebGPT-4 Anki Builder. Version 1.0. Sasank Marabattula, Varun Deepak Gudhe, MatthewSills. 2023. [https://github.com/varundeepakgudhe/auto_anki]