Skip to content

dperique/LLM-text-handing

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

26 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

LLM text handling

termi-chat logo

These are programs that handle text using LLMs. We try hard to use quality opensource (e.g. llama3:8b) LLMs first and also paid models where appropriate.

Before running any of the programs, do this:

python -m env venv
source ./env/bin/activate
pip install -r requirements.txt
export PYTHONPATH=$(pwd)/python

Setup an ollama service or get an openai apikey.

NOTE: the summarizers support only ollama with llama3:8b for now; please ensure that your ollama API server is reachable at http://127.0.0.1:11434. The focus was on a local API server since summarization tends to be something with sensitive data.

Text summarizer

Taking inspiration from infiniteGPT and other chunking methods (e.g., llama-index), we have a text summarizer that can use ollama or openai. Use ollama for private summarization, hence the name osummarize

Run it like this:

./python osummarize/osummarize.py

Streamlit text summarizer

Using streamlit, we have a summarizer that can be used via a web interface as a "co-pilot" for your text summarization needs.

Run it like this:

streamlit run --server.port 8509 --server.headless True --theme.base dark sl_summarize/sl_osummary.py

Then browse to localhost:8509. You can create multiple tabs with that link so you can have multiple summarizations.

summarizer.png

HTML reader

This is an app that can read webpages. It is quite rudimentary but useful for gathering text from webpages (e.g., for summarization).

Run it like this:

python/reader/reader.py

Voice conversation chat

Taking inspiration from this youtube video showing chatgpt4o with whisper and tts-1 model, we make a simple, chat application using a popular chat application coding pattern. The program makes use of the ollama api and openai tts-1 model.

NOTE: pip install pyaudio did not work for me but conda install pyaudio did.

Run it like this:

export OPENAI_API_KEY=sk-... (fill in your API key)
python/voice-chat/voice-chat.py

Type check during development

We add type hints in the code so use this to ensure you comply:

mypy --install-types ;# You only need to do this once in your venv

Do this before committing to ensure you comply:

$ mypy python/sl_summarize/sl_osummary.py
Success: no issues found in 1 source file

$ mypy python/reader/reader.py
Success: no issues found in 1 source file

$ mypy python/voice-chat/voice_app.py
Success: no issues found in 1 source file

$ mypy python/osummarize/osummarize.py
Success: no issues found in 1 source file

or do this to check all in one command:

make lint

If there are exceptions, update the mypy.ini file accordingly. But please do so sparingly as we want to take advantage of typing checking as much as we can.

Releases

No releases published

Packages

No packages published