Skip to content

Commit

Permalink
Improving goto-datetime
Browse files Browse the repository at this point in the history
  • Loading branch information
@renatoathaydes
renatoathaydes committed Oct 4, 2021
1 parent fc784ea commit a4adb36
Show file tree
Hide file tree
Showing 3 changed files with 31 additions and 19 deletions.
31 changes: 19 additions & 12 deletions src/main/java/com/athaydes/logfx/text/DateTimeFormatGuesser.java
View file
Original file line number Diff line number Diff line change
Expand Up @@ -31,10 +31,10 @@ public final class DateTimeFormatGuesser {

private static volatile DateTimeFormatGuesser STANDARD_INSTANCE;

static DateTimeFormatGuesser createStandard() {
private static DateTimeFormatGuesser createStandard() {
var prefixRegex = "[a-zA-Z\\[\\]_ -]{0,20}";
var time = "\\d{1,2}:\\d{1,2}:\\d{1,2}";
var timeMs = time + "(\\.\\d{1,6})?";
var timeMs = time + "(\\.\\d{1,9})?";

// 2017-09-11T18:13:57.483+02:00
var isoDateTime = prefixRegex + namedGroup( DATE_TIME_GROUP,
Expand All @@ -47,24 +47,31 @@ static DateTimeFormatGuesser createStandard() {
namedGroup( TIMEZONE_GROUP, "\\s+(GMT|[+-]\\d{4})" ) );

// Tue Aug 11 21:55:22 CEST 2020
String commonDateTime = prefixRegex + namedGroup( DATE_TIME_GROUP,
var commonDateTime = prefixRegex + namedGroup( DATE_TIME_GROUP,
"(\\w{3}\\s+)?\\w{3}\\s+\\d{1,2}\\s+" + time +
namedGroup( TIMEZONE_GROUP, "\\s+\\w{1,10}\\s+\\d{2,4}" ) );

// 10/Oct/2000:13:55:36 -0700
String ncsaCommonLogFormat = prefixRegex + namedGroup( DATE_TIME_GROUP,
var ncsaCommonLogFormat = prefixRegex + namedGroup( DATE_TIME_GROUP,
"\\w{1,10}[/\\-.]\\w{1,10}[/\\-.]\\d{2,4}([T\\s:]|\\s{1,2})" + time +
namedGroup( TIMEZONE_GROUP, "\\s{0,2}[+-]\\d{1,2}(:)?\\d{1,2}" ) + "?" );

// 2020-06-21 18:33:34-07
var appleDateTime = prefixRegex + namedGroup( DATE_TIME_GROUP,
"\\d{1,4}-\\d{1,2}-\\d{1,2}\\s+" + timeMs +
namedGroup( TIMEZONE_GROUP, "[+-]\\d{1,2}" ) );

return new DateTimeFormatGuesser( List.of(
new PatternBasedDateTimeFormatGuess( Pattern.compile( isoDateTime + ".*" ),
DateTimeFormatter.ofPattern( "yyyy-M-d'T'H:m:s[.SSS][z]" ) ),
new PatternBasedDateTimeFormatGuess( Pattern.compile( commonDateTime + ".*" ),
DateTimeFormatter.ofPattern( "[EE ]MMM dd HH:mm:ss[.SSS][ zzz][ yyyy]" ) ),
new PatternBasedDateTimeFormatGuess( Pattern.compile( ncsaCommonLogFormat + ".*" ),
DateTimeFormatter.ofPattern( "d/MMM/yyyy:H:m:s[:SSS][ Z]" ) ),
new PatternBasedDateTimeFormatGuess( Pattern.compile( rfc1123DateTime + ".*" ),
DateTimeFormatter.RFC_1123_DATE_TIME ) )
new PatternBasedDateTimeFormatGuess( "ISO", Pattern.compile( isoDateTime + ".*" ),
DateTimeFormatter.ofPattern( "yyyy-M-d'T'H:m:s[.SSS][.SS][.S][z]" ) ),
new PatternBasedDateTimeFormatGuess( "Common", Pattern.compile( commonDateTime + ".*" ),
DateTimeFormatter.ofPattern( "[EE ]MMM dd HH:mm:ss[.SSS][.SS][.S][ zzz][ yyyy]" ) ),
new PatternBasedDateTimeFormatGuess( "NCSA", Pattern.compile( ncsaCommonLogFormat + ".*" ),
DateTimeFormatter.ofPattern( "d/MMM/yyyy:H:m:s[:SSS][:SS][:S][ Z]" ) ),
new PatternBasedDateTimeFormatGuess( "RFC-1123", Pattern.compile( rfc1123DateTime + ".*" ),
DateTimeFormatter.RFC_1123_DATE_TIME ),
new PatternBasedDateTimeFormatGuess( "APPLE", Pattern.compile( appleDateTime + ".*" ),
DateTimeFormatter.ofPattern( "yyyy-M-d H:m:s[.SSS][.SS][.S][x]" ) ) )
);
}

Expand Down
1 change: 1 addition & 0 deletions src/main/java/com/athaydes/logfx/text/PatternBasedDateTimeFormatGuess.java
View file
Original file line number Diff line number Diff line change
Expand Up @@ -18,6 +18,7 @@
* The pattern must use the given group to extract the date-time.
*/
public record PatternBasedDateTimeFormatGuess(
String name,
Pattern linePattern,
DateTimeFormatter formatter
) implements DateTimeFormatGuess {
Expand Down
18 changes: 11 additions & 7 deletions src/test/groovy/com/athaydes/logfx/text/DateTimeFormatGuesserSpec.groovy
View file
Original file line number Diff line number Diff line change
Expand Up @@ -49,6 +49,7 @@ class DateTimeFormatGuesserSpec extends Specification {
where: 'valid log lines and their expected date-times'
lines << [
[ '2021-9-3T20:54:54.207+02:00 [file-change-watcher-1] INFO com.athaydes.logfx.file.FileChangeWatcher - Watching file ' ],
[ '2021-9-3T20:54:51.97+02:00' ],
[ '2017-09-11T18:13:57.483+02:00 TRACE {worker-1} com.acme.Log event' ],
[ '2017-9-14T19:23:53.499+02:00 [pool-6-thread-1] DEBUG com' ],
[ '2013-2-23T5:6:7 [pool-6-thread-1] DEBUG com' ],
Expand All @@ -59,10 +60,12 @@ class DateTimeFormatGuesserSpec extends Specification {
[ '[Fri, 09 Sep 2011 10:42:29 +0200] [core:error]' ],
[ '[Fri, 09 Sep 2011 10:42:29 -0530] [core:error]' ],
[ '[10/Oct/2000:13:55:36 -0700] INFO hello' ],
[ '2020-06-21 18:32:45-07 INFO hello' ],
]

expectedDateTimes << [
[ dateTime( '2021-09-03T20:54:54.207+02:00' ) ],
[ dateTime( '2021-09-03T20:54:51.97+02:00' ) ],
[ dateTime( '2017-09-11T18:13:57.483+02:00' ) ],
[ dateTime( '2017-09-14T19:23:53.499+02:00' ) ],
[ dateTime( '2013-02-23T05:06:07.000' ) ],
Expand All @@ -73,6 +76,7 @@ class DateTimeFormatGuesserSpec extends Specification {
[ dateTime( '2011-09-09T10:42:29+02:00' ) ],
[ dateTime( '2011-09-09T10:42:29-05:30' ) ],
[ dateTime( '2000-10-10T13:55:36-07:00' ) ],
[ dateTime( '2020-06-21T18:32:45-07:00' ) ],
]
}

Expand All @@ -83,7 +87,7 @@ class DateTimeFormatGuesserSpec extends Specification {
'This line has no date-time',
'2017-09-14T19:23:53.499+02:00 [pool-6-thread-1] DEBUG com',
'2013-02-23T05:06:07 [pool-6-thread-1] DEBUG com',
'INFO 2017-09-11T18:13:57.485+04:00 {worker-1} com.acme.Log event',
'2017-09-11T18:13:57.485+04:00 {worker-1} com.acme.Log event',
'INFO [Fri, 01 Sep 20017 22:02:55 GMT] - 22',
'Sun, 03 Nov 2019 14:22:00 GMT: msg',
'This line has no date-time',
Expand All @@ -98,7 +102,7 @@ class DateTimeFormatGuesserSpec extends Specification {

def patterns = [
// 2017-09-11T18:13:57.483+02:00
namedGroup( DATE_TIME_GROUP, "\\d{1,4}-\\d{1,4}-\\d{1,4}(T|\\s{1,2})$time$tz" ),
namedGroup( DATE_TIME_GROUP, "\\d{1,4}-\\d{1,4}-\\d{1,4}(T|\\s{1,2})$time$tz?" ),

// Tue, 3 Jun 2008 11:05:30 GMT
namedGroup( DATE_TIME_GROUP, "\\w{1,3},\\s+\\d{1,2}\\s+\\w{1,10}\\s+\\d{2,4}\\s+$time" +
Expand All @@ -113,10 +117,10 @@ class DateTimeFormatGuesserSpec extends Specification {
].collect { Pattern.compile( "${it}.*" ) }

def guesser = new DateTimeFormatGuesser( [
new PatternBasedDateTimeFormatGuess( patterns[ 0 ], DateTimeFormatter.ISO_DATE_TIME ),
new PatternBasedDateTimeFormatGuess( patterns[ 1 ], DateTimeFormatter.RFC_1123_DATE_TIME ),
new PatternBasedDateTimeFormatGuess( patterns[ 2 ], DateTimeFormatter.RFC_1123_DATE_TIME ),
new PatternBasedDateTimeFormatGuess( patterns[ 3 ],
new PatternBasedDateTimeFormatGuess( 'iso', patterns[ 0 ], DateTimeFormatter.ISO_DATE_TIME ),
new PatternBasedDateTimeFormatGuess( 'rfc1', patterns[ 1 ], DateTimeFormatter.RFC_1123_DATE_TIME ),
new PatternBasedDateTimeFormatGuess( 'rfc2', patterns[ 2 ], DateTimeFormatter.RFC_1123_DATE_TIME ),
new PatternBasedDateTimeFormatGuess( 'ncsa', patterns[ 3 ],
DateTimeFormatter.ofPattern( "d/MMM/yyyy:H:m:s[:SSS][ Z]", Locale.ENGLISH ) ),
] )

Expand All @@ -139,7 +143,7 @@ class DateTimeFormatGuesserSpec extends Specification {

def "A single line with a valid date-time is enough for the guesser to guess a formatter correctly"() {
given: 'The standard date-time format guesser'
def guesser = DateTimeFormatGuesser.createStandard()
def guesser = DateTimeFormatGuesser.standard()

and: 'Lots of lines, only one of which has a valid date-time'
def lines = generateRandomLines( 100 )
Expand Down

0 comments on commit a4adb36

Please sign in to comment.
-