Entity-Relationship Model of the BeeSpace Navigator Database September 2007
|
|  Yüklə 155 Kb.
|
Entity-Relationship Model of the BeeSpace Navigator Database – September 2007
This is the database scheme for BeeSpace v3. The 3 relational tables at the top are the session dynamic values; they are filled in as the user navigates through the system. The tables left to right store values for the user logins, the user objects (spaces and regions), and the user queries. The 2 relational tables at the bottom are the collection static values, they are filled in as new documents are imported into the system. The left table is the parsed entities from the contents while the right table is the documents with metadata. Adding entities is easy, requiring only an alter table and an add column command. The Java web application with the user interface also retrieves from other sources besides the MySQL database. The full-text indexing for search is supported by the Lemur/Indri toolkit. Temporary indexes in the local file system, such as the semantic graph for mutual information, are only dynamically generated, not permanently stored. User Table The User Table contains the directory of BeeSpace users. It stores both visible and hidden fields of a user's account.   
Column definitions: uid is auto-generated user id (PK) unique email of user given name family name user's organization password
user's state variable user's secret 
User Object Table The User Object Table stores user objects (e.g. regions/spaces) and their metadata.   
Column definitions: oid is auto-generated object id (PK) owner of this operation type code of object link to object user-assigned group user-assigned label user-assigned description text object size last create/update timestamp is-public flag uid is of User Object Table
Column descriptions: sid is auto-generated query id (PK) owner of this operation op code first operand second operand collection id last create/update timestamp 
Entity Table The Entity Table stores entities found from Jing's NER program. Fields offset and length are character based metrics for identifying the entity text. The entity field is rather redundant and was used for backtracing/debugging primarily.   
Column definitions: eid is auto-generated entity record id (PK) unique email of user field in record entity is found character offset into field of start pos character length of entity text the actual entity text (mainly for backtracing) cid is of NER Table
Column definitions: cid is auto-generated citation db id (PK) citation external docno title of citation abstract of citation publication source of citation author field of citation mesh field of citation chemicals field of citation keywords field of citation genes field of citation publication date 
p. of |
