cachho
cachho
With all the PRs and Issues, it has become obvious that we as maintainers need to decide on a way to configure embedchain.
### 🚀 The feature We should allow for a consistent way to update documents. I guess this could be through an id or through matching source. We have to decide...
### 🚀 The feature implement #89 the right way. Add a loader to embed local files. ### Motivation, pitch This is a simple function that should be provided. ### Alternatives...
### 🐛 Describe the bug This test has a lot of weird stuff going on, from duplicates to functions without purpose. Should definitely be fixed. https://github.com/embedchain/embedchain/blob/main/tests/vectordb/test_chroma_db.py
### 🐛 Describe the bug Just like OpenAI or GPT4All, Llama2 apparently needs a model to work. It could also be an issue with Replicate as the provider. 1. find...
### 🐛 Describe the bug Test code: ```python from embedchain import App from embedchain.config import ChromaDbConfig, AppConfig app = App(AppConfig(log_level="DEBUG"),chromadb_config=ChromaDbConfig(chroma_settings={"allow_reset": True})) app.db.reset() app.add("https://docs.zextras.com/carbonio-ce/html/general.html#what-is-product") ```
### Issue with current documentation: re: #508
### 🐛 Describe the bug If you use the chat method, the history is stored persistently in the `Llm` app-subclass. If `query` is used in the same script after `chat`,...
### Issue with current documentation: It looks like zsh is not compatible without putting quotes.