[
https://issues.apache.org/jira/browse/TIKA-4545?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18044383#comment-18044383
]
Tilman Hausherr edited comment on TIKA-4545 at 12/11/25 11:23 AM:
------------------------------------------------------------------
Windows build fails in the tika pipes solr integration tests because of the
path problem we've had in the past. It can be fixed at 3 places in
TikaPipesSolrTestBase.getTikaConfig() like this
{code:java}
String res = json.replace("UPDATE_STRATEGY", updateStrategy.toString())
.replace("ATTACHMENT_STRATEGY", attachmentStrategy.toString())
.replaceAll("FETCHER_BASE_PATH",
Matcher.quoteReplacement(testFileFolder.toAbsolutePath().toString().replace("\\",
"/")))
.replace("PARSE_MODE", parseMode.name())
.replace("SOLR_URLS", solrUrls)
.replace("SOLR_ZK_HOSTS", solrZkHosts);
res = res.replace("TIKA_CONFIG",
tikaConfig.toAbsolutePath().toString().replace("\\", "/"));
Path log4jPropFile = pipesDirectory.resolve("log4j2.xml");
try (InputStream is =
this.getClass().getResourceAsStream("/pipes-fork-server-custom-log4j2.xml")) {
Files.copy(is, log4jPropFile,
java.nio.file.StandardCopyOption.REPLACE_EXISTING);
}
res = res.replace("LOG4J_PROPERTIES_FILE",
log4jPropFile.toAbsolutePath().toString().replace("\\", "/"));
{code}
Same in S3PipeIntegrationTest.java:
{code:java}
String pluginsConfig = pluginsTemplate
.replace("{TIKA_CONFIG}",
tikaConfigFile.getAbsolutePath().replace("\\", "/"))
.replace("{LOG4J_PROPERTIES_FILE}",
log4jPropFile.getAbsolutePath().replace("\\", "/"))
{code}
Same in TikaPipesKafkaTest.getTikaConfig, I realize now that a replace at the
end is fine too, like already done elsewhere.
was (Author: tilman):
Windows build fails in the tika pipes solr integration tests because of the
path problem we've had in the past. It can be fixed at 3 places in
TikaPipesSolrTestBase.getTikaConfig() like this
{code:java}
String res = json.replace("UPDATE_STRATEGY", updateStrategy.toString())
.replace("ATTACHMENT_STRATEGY", attachmentStrategy.toString())
.replaceAll("FETCHER_BASE_PATH",
Matcher.quoteReplacement(testFileFolder.toAbsolutePath().toString().replace("\\",
"/")))
.replace("PARSE_MODE", parseMode.name())
.replace("SOLR_URLS", solrUrls)
.replace("SOLR_ZK_HOSTS", solrZkHosts);
res = res.replace("TIKA_CONFIG",
tikaConfig.toAbsolutePath().toString().replace("\\", "/"));
Path log4jPropFile = pipesDirectory.resolve("log4j2.xml");
try (InputStream is =
this.getClass().getResourceAsStream("/pipes-fork-server-custom-log4j2.xml")) {
Files.copy(is, log4jPropFile,
java.nio.file.StandardCopyOption.REPLACE_EXISTING);
}
res = res.replace("LOG4J_PROPERTIES_FILE",
log4jPropFile.toAbsolutePath().toString().replace("\\", "/"));
{code}
Same in S3PipeIntegrationTest.java:
{code:java}
String pluginsConfig = pluginsTemplate
.replace("{TIKA_CONFIG}",
tikaConfigFile.getAbsolutePath().replace("\\", "/"))
.replace("{LOG4J_PROPERTIES_FILE}",
log4jPropFile.getAbsolutePath().replace("\\", "/"))
{code}
> Fully integrate new json based deserializer in 4.x
> --------------------------------------------------
>
> Key: TIKA-4545
> URL: https://issues.apache.org/jira/browse/TIKA-4545
> Project: Tika
> Issue Type: Task
> Reporter: Tim Allison
> Priority: Major
> Fix For: 4.0.0
>
>
> Follow on for TIKA-4544.
> Steps:
> * Add annotations to components (parsers, etc.) and unit tests to confirm
> they work (finished this today)
> * Modify components (parsers etc), at least a few of them so that they are
> actually configurable. We don't have to modify all, just the most important
> ones PDFParser, tesseract, MSOffice, and others???
> * Move to tika-config.json in tika-pipes client/server, tika-async-cli,
> tika-app and tika-server one by one
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
