mindforger VS hunspell

Rear is a really interesting project with admirable goals. I believe this is just the beginning, but you have already done a great job!

I have been working on my note-taking application (https://github.com/dvorka/mindforger) for some time and wanted to go in the same direction. However, I gave up (for now). I used ggerganov/llama.cpp to host LLM models locally on a CPU-only machine with 32GB RAM, and used them for both RAG and note-taking use cases (like https://www.mindforger.com/index-200.html#llm). However, it did not work well for me - the performance was poor (high hardware utilization, long response times, failures, and crashes) and the actual responses were rarely useful (off-topic and impractical responses, hallucinations). I tried llama-2 7B with 4b quantization and a couple of similar models. Although I'm not happy about it, I switched to an online commercial LLM because it performs really well in terms of response quality, speed, and affordability. I frequently use the integrated LLM in my note-taking app as it can be used for many things.

Anyway, Reor "only" uses the locally hosted LLM in the generation phase of the RAG, which is a nicely constraint use case. I believe that a really lightweight LLM - I'm thinking about a tiny base model fine-tuned for summarization - could be the way to go (fast, non-hallucinating). I'm really curious to know if you have any suggestions or if you will have any in the future!

As for the vector DB, considering the resource-related problems I mentioned earlier, I was thinking about something similar to facebookresearch/faiss, which, unlike LanceDB, is not a fully-fledged vector DB. Have you made any experiments with similarity search projects or vector DBs? I would be interested in the trade-offs similar to small/large/hosted LLMs.

Overall, I think that both RAG with my personal notes as a corpus and a locally hosted generic purpose LLM for the use cases I mentioned above can take personal note-taking apps to a new level. This is the way! ;)

Good luck with your project!

Please share your suggestions, ideas or constructive criticism! You may install or update from GitHubreleases or PPA.

I finally managed to complete new MindForger release:

https://www.mindforger.com/NimbusnoteWikidpadBecause you mentioned writing:

Neat, the Similarity type reminds me of MindForger's Associations feature that also displays similarity scores between your current note and other existing notes

[Mindfrogger](https://github.com/dvorka/mindforger)

Mindforger: https://github.com/dvorka/mindforger/

Hunspell seems popular as well. I believe it's the one used by Firefox and LibreOffice, and I think it's the system spell checker in MacOS already? 🤷‍♂️

Almost all browsers use the Hunspell library for spell checking. You should investigate what methods it uses for stemming and suggesting corrections. How does that algorithm work for non-English languages? The main variation you will see between browsers in spelling suggestions is the base dictionary that is used.

I don't think so. Looking at the Chromium source it appears to use Hunspell. This is an okay spell checker, but not AI based AFAIK, only "Morphological analysis, stemming and generation".

Are you familiar with Hunspell? Dictionaries are comprehensive enough to be part of different Office Suites, so I don't see them as constricted to autocorrect.

If you're using Linux or MacOS, you should try hunspell.

i prefer and use hunspell

One can implement Huntspell which is what all browsers use (for example when typing in text areas). Is very simple and is C++.