Package org.apache.tika.eval.app

Class ExtractComparer

java.lang.Object

org.apache.tika.eval.app.ProfilerBase

org.apache.tika.eval.app.ExtractComparer

public class ExtractComparer
extends ProfilerBase

Nested Class Summary

Nested classes/interfaces inherited from class org.apache.tika.eval.app.ProfilerBase

ProfilerBase.EXCEPTION_TYPE, ProfilerBase.PARSE_ERROR_TYPE

Field Summary

Fields

Modifier and Type	Field	Description
static TableInfo	COMPARISON_CONTAINERS
static TableInfo	CONTENT_COMPARISONS
static TableInfo	CONTENTS_TABLE_A
static TableInfo	CONTENTS_TABLE_B
static TableInfo	EMBEDDED_FILE_PATH_TABLE_A
static TableInfo	EMBEDDED_FILE_PATH_TABLE_B
static TableInfo	EXCEPTION_TABLE_A
static TableInfo	EXCEPTION_TABLE_B
static TableInfo	EXTRACT_EXCEPTION_TABLE_A
static TableInfo	EXTRACT_EXCEPTION_TABLE_B
static TableInfo	PROFILES_A
static TableInfo	PROFILES_B
static TableInfo	REF_PAIR_NAMES
static TableInfo	TAGS_TABLE_A
static TableInfo	TAGS_TABLE_B

Fields inherited from class org.apache.tika.eval.app.ProfilerBase

FALSE, ID, MIME_TABLE, REF_EXTRACT_EXCEPTION_TYPES, REF_PARSE_ERROR_TYPES, REF_PARSE_EXCEPTION_TYPES, TRUE, writer

Constructor Summary

Constructors

Constructor	Description
ExtractComparer(Path inputDir, Path extractsA, Path extractsB, ExtractReader extractReader, IDBWriter writer)

Method Summary

Modifier and Type	Method	Description
protected void	compareFiles(org.apache.tika.eval.app.EvalFilePaths fpsA, org.apache.tika.eval.app.EvalFilePaths fpsB)
boolean	processFileResource(FetchKey fetchKey)
static void	USAGE()

Methods inherited from class org.apache.tika.eval.app.ProfilerBase

calcTextStats, closeWriter, getContent, getFileLength, getPathsFromExtractCrawl, getPathsFromSrcCrawl, getSourceFileLength, loadCommonTokens, setMaxContentLength, setMaxContentLengthForLangId, setMaxTokens, truncateContent, writeContentData, writeExceptionData, writeExtractException, writeProfileData

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Details

REF_PAIR_NAMES

public static TableInfo REF_PAIR_NAMES

COMPARISON_CONTAINERS

public static TableInfo COMPARISON_CONTAINERS

CONTENT_COMPARISONS

public static TableInfo CONTENT_COMPARISONS

PROFILES_A

public static TableInfo PROFILES_A

PROFILES_B

public static TableInfo PROFILES_B

EMBEDDED_FILE_PATH_TABLE_A

public static TableInfo EMBEDDED_FILE_PATH_TABLE_A

EMBEDDED_FILE_PATH_TABLE_B

public static TableInfo EMBEDDED_FILE_PATH_TABLE_B

CONTENTS_TABLE_A

public static TableInfo CONTENTS_TABLE_A

CONTENTS_TABLE_B

public static TableInfo CONTENTS_TABLE_B

TAGS_TABLE_A

public static TableInfo TAGS_TABLE_A

TAGS_TABLE_B

public static TableInfo TAGS_TABLE_B

EXCEPTION_TABLE_A

public static TableInfo EXCEPTION_TABLE_A

EXCEPTION_TABLE_B

public static TableInfo EXCEPTION_TABLE_B

EXTRACT_EXCEPTION_TABLE_A

public static TableInfo EXTRACT_EXCEPTION_TABLE_A

EXTRACT_EXCEPTION_TABLE_B

public static TableInfo EXTRACT_EXCEPTION_TABLE_B

Constructor Details

ExtractComparer

public ExtractComparer(Path inputDir,
 Path extractsA,
 Path extractsB,
 ExtractReader extractReader,
 IDBWriter writer)

Method Details

USAGE

public static void USAGE()
                  throws IOException

Throws:

IOException

processFileResource

public boolean processFileResource(FetchKey fetchKey)

Specified by:

processFileResource in class ProfilerBase

compareFiles

protected void compareFiles(org.apache.tika.eval.app.EvalFilePaths fpsA,
 org.apache.tika.eval.app.EvalFilePaths fpsB)
                     throws IOException

Throws:

IOException