[
https://issues.apache.org/jira/browse/OPENNLP-589?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17836992#comment-17836992
]
ASF GitHub Bot commented on OPENNLP-589:
----------------------------------------
rzo1 commented on code in PR #596:
URL: https://github.com/apache/opennlp/pull/596#discussion_r1564855870
##########
opennlp-tools/src/main/java/opennlp/tools/ml/maxent/RealBasicEventStream.java:
##########
@@ -49,13 +55,14 @@ public Event read() throws IOException {
}
private Event createEvent(String obs) {
- int lastSpace = obs.lastIndexOf(' ');
- if (lastSpace == -1)
+ int si = obs.indexOf(' ');
+ if (si == 0)
Review Comment:
Shouldn't that be `-1` ? Otherwise, it might produce an index out of bound
exception in the else branch (because nothing prevents the caller to just
insert `bla` as an element in the event stream (alhough it does not comply with
the text format).
> Text format of Events inconsistent across different implementations of
> EventStreamReaders
> -----------------------------------------------------------------------------------------
>
> Key: OPENNLP-589
> URL: https://issues.apache.org/jira/browse/OPENNLP-589
> Project: OpenNLP
> Issue Type: Bug
> Components: Machine Learning
> Affects Versions: maxent-3.0.3, 2.0.0, 2.1.0, 2.2.0, 2.3.0
> Reporter: Marcin Junczys-Dowmunt
> Assignee: Martin Wiesner
> Priority: Minor
> Fix For: 2.3.3
>
>
> BasicEventStream expects events to be written to text files as:
> context1 context2 context3 ... outcome
> FileEventStream expects events to be written to text files as:
> outcome context1 context2 context3 ...
> toString() of Event creates:
> outcome [context1 context2 context3 ...] (note the square brackets, which are
> part of context predicates when breaking on spaces).
> This is highly confusing and took me some time to understand. I guess this
> should be unified?
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
