This is an automated email from the ASF dual-hosted git repository.
tballison pushed a commit to branch main
in repository https://gitbox.apache.org/repos/asf/tika.git
The following commit(s) were added to refs/heads/main by this push:
new 7b36986bb7 TIKA-4885: clean up tika cli (#3144)
7b36986bb7 is described below
commit 7b36986bb7ce035aca23892052cbdff3e5a1f4e6
Author: Tim Allison <[email protected]>
AuthorDate: Tue Sep 8 16:26:02 2026 -0400
TIKA-4885: clean up tika cli (#3144)
---
CHANGES.txt | 3 ++
.../src/main/java/org/apache/tika/cli/TikaCLI.java | 58 +++++++---------------
.../test/java/org/apache/tika/cli/TikaCLITest.java | 16 ++++++
3 files changed, 37 insertions(+), 40 deletions(-)
diff --git a/CHANGES.txt b/CHANGES.txt
index 673a175fe8..95b46ec3af 100644
--- a/CHANGES.txt
+++ b/CHANGES.txt
@@ -1,5 +1,8 @@
Release 4.1.0 - unreleased
+ * tika-app's -m and --json wrote metadata at the SAX endDocument event
+ thereby dropping keys added by a parser after the endDocument (TIKA-4885).
+
* "content-enrichers": _mime-exclude now matches the real media type, so
excluding image/tiff also drops a legacy image/ocr-tiff engine (it was
a no-op). The zero-media-types check at config load asks the engine
diff --git a/tika-app/src/main/java/org/apache/tika/cli/TikaCLI.java
b/tika-app/src/main/java/org/apache/tika/cli/TikaCLI.java
index 1089229760..e6ba42a9d4 100644
--- a/tika-app/src/main/java/org/apache/tika/cli/TikaCLI.java
+++ b/tika-app/src/main/java/org/apache/tika/cli/TikaCLI.java
@@ -1384,41 +1384,28 @@ public class TikaCLI {
}
}
+ /** Swallows the document; {@link #output()} writes the metadata after the
parse. */
private static class NoDocumentMetHandler extends DefaultHandler {
protected final Metadata metadata;
protected PrintWriter writer;
- private boolean metOutput;
-
public NoDocumentMetHandler(Metadata metadata, PrintWriter writer) {
this.metadata = metadata;
this.writer = writer;
- this.metOutput = false;
}
- @Override
- public void endDocument() {
+ public void output() throws IOException {
String[] names = metadata.names();
Arrays.sort(names);
- outputMetadata(names);
- writer.flush();
- this.metOutput = true;
- }
-
- public void outputMetadata(String[] names) {
for (String name : names) {
for (String value : metadata.getValues(name)) {
writer.println(name + ": " + value);
}
}
+ writer.flush();
}
-
- public boolean metOutput() {
- return this.metOutput;
- }
-
}
/**
@@ -1463,16 +1450,16 @@ public class TikaCLI {
private class OutputType {
public void process(TikaInputStream tis, OutputStream output, Metadata
metadata) throws Exception {
- Parser p = parser;
ContentHandler handler = getContentHandler(output, metadata);
- p.parse(tis, handler, metadata, context);
- // fix for TIKA-596: if a parser doesn't generate
- // XHTML output, the lack of an output document prevents
- // metadata from being output: this fixes that
- if (handler instanceof NoDocumentMetHandler) {
- NoDocumentMetHandler metHandler = (NoDocumentMetHandler)
handler;
- if (!metHandler.metOutput()) {
- metHandler.endDocument();
+ try {
+ parser.parse(tis, handler, metadata, context);
+ } finally {
+ // Metadata is written once parse() returns, never at
endDocument: parsers
+ // set keys after their SAX document ends (the PDF totals in a
finally block,
+ // tk:parsed-by-full-set), some never start a document
(TIKA-596), and a
+ // failed parse still has metadata worth showing before the
error.
+ if (handler instanceof NoDocumentMetHandler metHandler) {
+ metHandler.output();
}
}
}
@@ -1484,26 +1471,17 @@ public class TikaCLI {
}
- private class NoDocumentJSONMetHandler extends DefaultHandler {
-
- protected final Metadata metadata;
-
- protected PrintWriter writer;
+ private class NoDocumentJSONMetHandler extends NoDocumentMetHandler {
public NoDocumentJSONMetHandler(Metadata metadata, PrintWriter writer)
{
- this.metadata = metadata;
- this.writer = writer;
+ super(metadata, writer);
}
@Override
- public void endDocument() throws SAXException {
- try {
- JsonMetadata.setPrettyPrinting(prettyPrint);
- JsonMetadata.toJson(metadata, writer);
- writer.flush();
- } catch (IOException e) {
- throw new SAXException(e);
- }
+ public void output() throws IOException {
+ JsonMetadata.setPrettyPrinting(prettyPrint);
+ JsonMetadata.toJson(metadata, writer);
+ writer.flush();
}
}
}
diff --git a/tika-app/src/test/java/org/apache/tika/cli/TikaCLITest.java
b/tika-app/src/test/java/org/apache/tika/cli/TikaCLITest.java
index b5efaeb364..1b3469a5f0 100644
--- a/tika-app/src/test/java/org/apache/tika/cli/TikaCLITest.java
+++ b/tika-app/src/test/java/org/apache/tika/cli/TikaCLITest.java
@@ -279,6 +279,22 @@ public class TikaCLITest {
assertTrue(json.contains("\"tk:digest:MD2\":"));
}
+ /**
+ * Keys a parser sets after its SAX document ends (the PDF parser's
document totals in
+ * its finally block, tk:parsed-by-full-set from the composite) must still
reach -m and
+ * --json, which used to write at endDocument.
+ */
+ @Test
+ public void testMetadataSetAfterEndDocumentIsOutput() throws Exception {
+ String content = getParamOutContent("-m", resourcePrefix +
"testPDF_childAttachments.pdf");
+ assertTrue(content.contains("pdf:ocr-page-count: 0"), content);
+ assertTrue(content.contains("tk:parsed-by-full-set:"), content);
+
+ String json = getParamOutContent("--json", resourcePrefix +
"testPDF_childAttachments.pdf");
+ assertTrue(json.contains("\"pdf:ocr-page-count\":"), json);
+ assertTrue(json.contains("\"tk:parsed-by-full-set\":"), json);
+ }
+
/**
* Test for -json with prettyprint option
*
