[
https://issues.apache.org/jira/browse/OPENNLP-1973?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Richard Zowalla closed OPENNLP-1973.
------------------------------------
Resolution: Fixed
Fixed by https://github.com/apache/opennlp/pull/1312 (c47eeafef).
> Add a CompatibilityMode switch to keep the earlier output for models trained
> with older releases
> ------------------------------------------------------------------------------------------------
>
> Key: OPENNLP-1973
> URL: https://issues.apache.org/jira/browse/OPENNLP-1973
> Project: OpenNLP
> Issue Type: Improvement
> Reporter: Richard Zowalla
> Assignee: Kristian Rickert
> Priority: Major
> Fix For: 3.0.0
>
> Time Spent: 40m
> Remaining Estimate: 0h
>
> Split out of OPENNLP\-1928 \(PR #1275\).
> Adds {{opennlp.tools.util.CompatibilityMode}} to opennlp\-api with the values
> {{LEGACY}} and {{CURRENT}}, resolved from the system property
> {{opennlp.compat.mode}} at startup and overridable with {{setActive}} and
> {{reset}}. {{CURRENT}} is the default; {{LEGACY}} logs one warning per
> process. It is built like the whitespace mode switch and, like it, is
> scheduled for removal in 4.0.
> Classes whose output was corrected consult it to reproduce the earlier output
> for models that were trained on it. Users: the EmojiCharSequenceNormalizer
> fix \(RegEx removal part 8\) and the surrogate rule of OPENNLP\-1934.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
