| { |
| "name": "Host", |
| "type": "record", |
| "namespace": "org.apache.nutch.storage", |
| "doc": "Host represents a store of webpages or other data which resides on a server or other computer so that it can be accessed over the Internet", |
| "fields": [ |
| { |
| "name": "metadata", |
| "type": { |
| "type": "map", |
| "values": ["null","bytes"] |
| }, |
| "doc": "A multivalued metadata container used for storing a wide variety of host metadata such as structured web server characterists etc", |
| "default": { |
| |
| } |
| }, |
| { |
| "name": "outlinks", |
| "type": { |
| "type": "map", |
| "values": ["null","string"] |
| }, |
| "doc": "Hyperlinks which direct outside of the current host domain these can used in a histogram style manner to generate host statistics", |
| "default": { |
| |
| } |
| }, |
| { |
| "name": "inlinks", |
| "type": { |
| "type": "map", |
| "values": ["null","string"] |
| }, |
| "doc": "Hyperlinks which link to pages within the current host domain these can used in a histogram style manner to generate host statistics", |
| "default": { |
| |
| } |
| } |
| ] |
| } |