# Chunks
A chunk is a basic building block of a knowledge-base. Ragit splits data files into chunks, and adds title and summary to each chunk. When it retrieves information, it scores, reranks, and retrieves chunks.
## Data Files
You build a knowledge-base from raw data files, like `.md`, `.txt`, `.pdf`... There are 3 kinds of files: unstaged files, staged files and processed files. If you're familiar with git, you must be familiar with the word "stage". Yes that's it.
Ragit tries to be as git-like as possible. Staged files are like those in git and processed files are like committed files in git. You run `rag add *.txt` or `rag add --all` to stage files. `rag add` respects `.ragignore` file in the root directory (where `.ragit/` exists). `rag build` pops files from the staging area, create chunks for the file, then mark the file as "processed".
If the file you're trying to `add` is already processed, its behavior depends on whether the file is modified since the last `rag build`. If it has been modified, it's staged again. The previously created chunks are not removed until `rag build` is run. If it hasn't been modified, it's not staged. It's not staged even when `--force` is set. In order to stage such file, you first have to run `rag remove <FILE>`.
Once a file is processed, `rag query` can use the chunks from the file. `rag query` doesn't care whether the original file exists or not. Chunks have all the information that `rag query` needs, and it doesn't try to look for the original file. That means you can delete the original files after `rag build` is complete.
## Data format
Chunks are saved in a content-addressable way. It's like git's object files. You can find the chunk files in `.ragit/chunks/`, a file per chunk. The first 2 characters of a chunk's uid is the directory name of the chunk file, and the remaining characters in uid consist its file name. For example, if its uid is `abcdef0123`, you'll find the chunk file at `.ragit/chunks/ab/cdef0123.chunk`.
There's no compact format like `.pack` in git.