priority 1
----------
- [x] add deltaEncode chunks function
    - [x] do not merge consecutive smaller chunks as these could be stored as chunks if no similar chunk is found. Thus it will need to be of `chunkSize` or less. Otherwise it could not be possibly used for deduplication.
    ```
    for each new chunk:
        find similar in sketchMap
        if exists:
            delta encode
        else:
            calculate fingerprint
            store in fingerprintMap
            store in sketchMap
    ```
- [ ] read from repo
    - [x] store recipe
    - [x] load recipe
    - [ ] read chunks in-order into a stream
    - [ ] read individual files
- [ ] properly store informations to be DNA encoded
    - [ ] tar source to keep files metadata ?
    - [ ] store chunks compressed
        - [ ] compress before storing
        - [ ] uncompress before loading
    - [ ] store compressed chunks into tracks ok trackSize (1024o)
- [ ] add chunk cache that would look like this:
    ```go
    type ChunkCache map[ChunkId][]byte // Do we really want to only keep the chunk content ?

    type Cache interface {
        Get(id ChunkId) Chunk
        Set(id ChunkId, Chunk)
    }
    ```

priority 2
----------
- [ ] use more the `Reader` API (which is analoguous to the `IOStream` in Java)
- [ ] refactor matchStream as right now it is quite complex