uk.ac.man.documentparser.dataholders
Class Document
java.lang.Object
uk.ac.man.documentparser.dataholders.Document
- All Implemented Interfaces:
- java.io.Serializable
public class Document
- extends java.lang.Object
- implements java.io.Serializable
- See Also:
- Serialized Form
Constructor Summary |
Document(java.lang.String id,
java.lang.String title,
java.lang.String abs,
java.lang.String body,
java.lang.String raw,
Document.Text_raw_type raw_type,
java.lang.String year,
Journal journal,
Document.Type type,
Author[] authors,
java.lang.String volume,
java.lang.String issue,
java.lang.String pages,
java.lang.String xml,
ExternalID externalID)
|
Methods inherited from class java.lang.Object |
clone, equals, finalize, getClass, hashCode, notify, notifyAll, wait, wait, wait |
serialVersionUID
private static final long serialVersionUID
- See Also:
- Constant Field Values
ID
private java.lang.String ID
title
private java.lang.String title
abs
private java.lang.String abs
body
private java.lang.String body
rawContent
private java.lang.String rawContent
raw_type
private Document.Text_raw_type raw_type
year
private java.lang.String year
ignoreCoordinates
private boolean ignoreCoordinates
authors
private Author[] authors
journal
private Journal journal
externalID
private ExternalID externalID
type
private Document.Type type
volume
private java.lang.String volume
issue
private java.lang.String issue
pages
private java.lang.String pages
xml
private java.lang.String xml
Document
public Document(java.lang.String id,
java.lang.String title,
java.lang.String abs,
java.lang.String body,
java.lang.String raw,
Document.Text_raw_type raw_type,
java.lang.String year,
Journal journal,
Document.Type type,
Author[] authors,
java.lang.String volume,
java.lang.String issue,
java.lang.String pages,
java.lang.String xml,
ExternalID externalID)
getID
public java.lang.String getID()
isValid
public boolean isValid(int start,
int end)
toString
public java.lang.String toString()
- Overrides:
toString
in class java.lang.Object
toString
public java.lang.String toString(boolean simplify)
print
public void print()
prepareInsertStatements
public static java.sql.PreparedStatement prepareInsertStatements(java.sql.Connection conn,
boolean clear)
prepareInsertStatements
public static java.sql.PreparedStatement prepareInsertStatements(java.sql.Connection conn,
java.lang.String table,
boolean clear)
- Parameters:
conn
- the SQL connection used for the insertion
- Returns:
- 0: INSERT INTO articles_raw (xml, raw_text, date_inserted)... ; 1: "INSERT INTO articles (id_art, text_title, text_abstract, text_body, text_raw, text_raw_type, year, id_issn, article_type, authors, volume, issue, pages, id_ext, source)
saveToDB
public void saveToDB(java.sql.PreparedStatement pstmt)
getTitle
public java.lang.String getTitle()
getYear
public java.lang.String getYear()
hasTitle
public boolean hasTitle()
saveToTextFile
public void saveToTextFile(java.io.File file,
boolean simplify)
toHTML
public java.lang.String toHTML()
getRawContent
public java.lang.String getRawContent()
getMeshTaxIDs
public java.util.HashSet<java.lang.Integer> getMeshTaxIDs(java.util.HashMap<java.lang.String,java.lang.Integer> meshToTax)
isIgnoreCoordinates
public boolean isIgnoreCoordinates()
- Returns:
- the ignoreCoordinates
setIgnoreCoordinates
public void setIgnoreCoordinates(boolean ignoreCoordinates)
- Parameters:
ignoreCoordinates
- the ignoreCoordinates to set
getSerialVersionUID
public static long getSerialVersionUID()
- Returns:
- the serialVersionUID
getAbs
public java.lang.String getAbs()
- Returns:
- the abs
getBody
public java.lang.String getBody()
- Returns:
- the body
getRaw_type
public Document.Text_raw_type getRaw_type()
- Returns:
- the raw_type
getAuthors
public Author[] getAuthors()
- Returns:
- the authors
getJournal
public Journal getJournal()
- Returns:
- the journal
getExternalID
public ExternalID getExternalID()
- Returns:
- the externalIDs
getType
public Document.Type getType()
- Returns:
- the type
getVolume
public java.lang.String getVolume()
- Returns:
- the volume
getIssue
public java.lang.String getIssue()
- Returns:
- the issue
getPages
public java.lang.String getPages()
- Returns:
- the pages
setBody
public void setBody(java.lang.String body)
- Parameters:
body
- the body to set
setRawContent
public void setRawContent(java.lang.String rawContent)
- Parameters:
rawContent
- the rawContent to set
getXml
public java.lang.String getXml()
- Returns:
- the xml
setRaw_type
public void setRaw_type(Document.Text_raw_type raw_type)
- Parameters:
raw_type
- the raw_type to set
getDescription
public java.lang.String getDescription()
setTitle
public void setTitle(java.lang.String title)
setAbs
public void setAbs(java.lang.String abs)