ESL-easel

an LLM-powered English PDF document analyzer that:

• summarizes the text
  • based on user's English language level
  • customize length of summary
• (optionally) shows new or "difficult" vocabulary
• (optionally) shows irregular verbs
• (optionally) shows and explain idioms like "break the ice" or "hit the sack"
• allows user to chat (ask questions on or perform operations about) the document
• allows for localization of user interface

Demo (for casual users)

Run the demo on esleasel.streamlit.app

(If the app has "fallen asleep" then please press the "Yes, get this app back up!" button to "wake" it up!)

Screenshots

Setup (for developers)

clone repo

Run the following commands on your command line:

% git clone https://github.com/joncoded/esleasel.git esleasel && cd esleasel
% pip install -r requirements.txt

.env file

Configure your .env file on the root folder (this must be done or the code will tell you to):

LLM_API_KEY=your_groq_api_key
PINECONE_API_KEY=your_pinecone_api_key
PINECONE_HOST=your_pinecone_host_url
PINECONE_INDEX=your_pinecone_index_name

If you don't have these you can get them for free at:

• groq
  • LLM_API_KEY : create api key > copy and paste an API key
    • you can later use the same API key for any large language model
• pinecone
  • PINECONE_API_KEY : on the sidebar
    • click "API keys"
    • click on the "+ API key" button
      • (modal will pop up) enter an API key name
      • click on the "Create key" button
      • copy and paste the API key
  • PINECONE_HOST : on the sidebar
    • click "Database"
    • click on the "Create index" button
      • leave everything default except for Dimensions: 768 (not 1024)
      • click on the "Create index" button
      • click the "Database" link on the sidebar again
      • copy and paste the host (looks like a URL)
  • PINECONE_INDEX :
    • if you got the host URL you will see the pinecone index name above it

Runtime!

Finally, back in the command line, run the app on your localhost with:

streamlit run app.py

The app will usually run with the url http://localhost:8501

(if port 8501 is already used, it will host the app on :8502 or so on...)

Contribute

Features

Make changes by following this procedure:

• clone the repository as mentioned above
• create a feature branch
  • git checkout -b feature/branch
• make and commit your changes
  • git commit -m "what you changed"
• push your changes
  • git push origin feature/branch
• open pull request
  • https://github.com/joncoded/esleasel/pulls

Also, feel free to raise any issues!

Localization

Translate the app into your language!

• edit the local.py dictionary file by adding a new language (see en (English) and fr (French) as examples
• go to app.py and check/change the lang_options variable
• optionally, make a pull request as if making a Feature contribution as listed above

Credits

• Groq for LLMs
• Pinecone for vector databases
• Langchain for its AI agent platform
• Streamlit for its super-easy deployment
• Google Gemini for recommending which LLM to use for analyzing English language PDFs
• Dr. Satyajit Pattnaik for technical guidance