Add a pseudo-harvester to manually create datasets

Created by: adamreichold

Given a source definition like

[[sources]]
name = "manual"
type = "manual"
url = "file://localhost/usr/share/umwelt-info/manual_datasets.toml"

human-readable dataset descriptions like

[[datasets]]
id = "undine"
title = "Undine"
description = "Informationsplattform zu hydrologischen Extremereignissen (Hochwasser, Niedrigwasser)"
license = "Unknown"
contacts = []
tags = []
source_url = "https://undine.bafg.de/"

[[datasets.resources]]
type = "Unknown"
url = "https://undine.bafg.de/"

placed in /usr/share/umwelt-info/manual_datasets.toml is parsed and added to the index in binary form.