Metadata-Version: 2.1
Name: cybrex
Version: 1.0.5
Summary: Researching AI
Author: Interdimensional Walker
Project-URL: Homepage, https://github.com/nexus-stc/stc
Classifier: Programming Language :: Python :: 3.8
Requires-Python: >=3.8
Description-Content-Type: text/markdown
Requires-Dist: aiokit (>=1.2.3)
Requires-Dist: chromadb (==0.3.26)
Requires-Dist: izihawa-configurator (==1.0.4)
Requires-Dist: izihawa-utils (==1.1.1)
Requires-Dist: langchain (==0.0.222)
Requires-Dist: openai (==0.27.8)
Requires-Dist: pypdf (==3.12.0)
Requires-Dist: pyyaml (==6.0)
Requires-Dist: tiktoken (==0.4.0)
Requires-Dist: stc-geck (>=1.2.5)

# Cybrex AI

Cybrex AI integrates several strategies to boost the role of AI in science:

- IPFS is utilized to access the Standard Template Construct (STC).
- STC provides the raw documents for Cybrex.
- OpenAI constructs embeddings for these documents and Cybrex stores these embeddings locally in the Chroma database.
- These embeddings are then used to filter relevant documents before they are sent to OpenAI for Q&A and summarization.

## Install

```bash
pip install cybrex
```

Upon its initial launch, `cybrex` will create a `~/.cybrex` directory containing a `config.yaml` file and a `chroma` directory.
You can edit the config file to point to different IPFS addresses.

## Usage

```bash
# Summarize a document
OPENAI_API_KEY=... cybrex sum-doc --doi 10.1155/2022/7138756

# Question a document
OPENAI_API_KEY=... python3 cybrex chat-doc --doi 10.1155/2022/7138756 \
  --question "What is the antivirus effect of resveratrol?"
```
