From 1271de4de51368f3fdfe1390ad54577cfac8fc59 Mon Sep 17 00:00:00 2001
From: Matthew Blissett <mblissett@gbif.org>
Date: Mon, 19 Mar 2018 17:22:06 +0100
Subject: [PATCH] Tests for existing basic validation done when using
 ArchiveFactory.

Then (regarding https://github.com/gbif/dwca-io/issues/43) we can move this
out of ArchiveFactory.
---
 .../java/org/gbif/dwca/io/ArchiveFactory.java |  3 +-
 .../org/gbif/dwca/io/ArchiveFactoryTest.java  | 47 +++++++++++++++++++
 src/test/resources/invalid/empty/.keep        |  0
 .../extension-core-id-missing/media.txt       |  0
 .../extension-core-id-missing/meta.xml        | 20 ++++++++
 .../extension-encoding-missing/media.txt      |  0
 .../extension-encoding-missing/meta.xml       | 20 ++++++++
 .../invalid/extension-id-missing/media.txt    |  0
 .../invalid/extension-id-missing/meta.xml     | 19 ++++++++
 .../extension-location-missing/media.txt      |  0
 .../extension-location-missing/meta.xml       | 19 ++++++++
 .../invalid/extension-missing/meta.xml        | 20 ++++++++
 .../meta-file-encoding-missing/meta.xml       | 11 +++++
 .../meta-file-encoding-missing/taxa.txt       |  0
 .../meta-file-location-missing/meta.xml       | 10 ++++
 .../meta-file-location-missing/taxa.txt       |  0
 src/test/resources/invalid/meta-only/meta.xml | 10 ++++
 17 files changed, 177 insertions(+), 2 deletions(-)
 create mode 100644 src/test/resources/invalid/empty/.keep
 create mode 100644 src/test/resources/invalid/extension-core-id-missing/media.txt
 create mode 100644 src/test/resources/invalid/extension-core-id-missing/meta.xml
 create mode 100644 src/test/resources/invalid/extension-encoding-missing/media.txt
 create mode 100644 src/test/resources/invalid/extension-encoding-missing/meta.xml
 create mode 100644 src/test/resources/invalid/extension-id-missing/media.txt
 create mode 100644 src/test/resources/invalid/extension-id-missing/meta.xml
 create mode 100644 src/test/resources/invalid/extension-location-missing/media.txt
 create mode 100644 src/test/resources/invalid/extension-location-missing/meta.xml
 create mode 100644 src/test/resources/invalid/extension-missing/meta.xml
 create mode 100644 src/test/resources/invalid/meta-file-encoding-missing/meta.xml
 create mode 100644 src/test/resources/invalid/meta-file-encoding-missing/taxa.txt
 create mode 100644 src/test/resources/invalid/meta-file-location-missing/meta.xml
 create mode 100644 src/test/resources/invalid/meta-file-location-missing/taxa.txt
 create mode 100644 src/test/resources/invalid/meta-only/meta.xml

diff --git a/src/main/java/org/gbif/dwca/io/ArchiveFactory.java b/src/main/java/org/gbif/dwca/io/ArchiveFactory.java
index f710ab05e..880ba6a91 100644
--- a/src/main/java/org/gbif/dwca/io/ArchiveFactory.java
+++ b/src/main/java/org/gbif/dwca/io/ArchiveFactory.java
@@ -25,9 +25,8 @@
 
 /**
  * Factory used to build {@link Archive} object from a DarwinCore Archive file.
- * 
- * @author mdoering
  *
+ * Replaced by DwcFiles, except that this also does some validation of the archive structure.
  */
 public class ArchiveFactory {
 
diff --git a/src/test/java/org/gbif/dwca/io/ArchiveFactoryTest.java b/src/test/java/org/gbif/dwca/io/ArchiveFactoryTest.java
index 160f7b397..bed6d19d7 100644
--- a/src/test/java/org/gbif/dwca/io/ArchiveFactoryTest.java
+++ b/src/test/java/org/gbif/dwca/io/ArchiveFactoryTest.java
@@ -542,4 +542,51 @@ public void testOpenArchiveForGenericCore() throws IOException, UnsupportedArchi
     assertEquals(DwcLayout.FILE_ROOT, arch.getDwcLayout());
   }
 
+  /**
+   * Basic validation of archives, that the declared files exist and have basic, valid structure.
+   */
+  @Test
+  public void testInvalidArchives() throws IOException {
+    // Simple archive problems
+    try {
+      Archive arch = ArchiveFactory.openArchive(FileUtils.getClasspathFile("invalid/empty"));
+      fail("Empty archive should not be opened.");
+    } catch (UnsupportedArchiveException e) {}
+
+    try {
+      Archive arch = ArchiveFactory.openArchive(FileUtils.getClasspathFile("invalid/meta-file-encoding-missing"));
+      fail("Archive with missing file encoding in meta.xml should not be opened.");
+    } catch (UnsupportedArchiveException e) {}
+
+    try {
+      Archive arch = ArchiveFactory.openArchive(FileUtils.getClasspathFile("invalid/meta-file-location-missing"));
+      fail("Archive with missing file location in meta.xml should not be opened.");
+    } catch (UnsupportedArchiveException e) {}
+
+    // Extension archive problems
+    try {
+      Archive arch = ArchiveFactory.openArchive(FileUtils.getClasspathFile("invalid/extension-missing"));
+      fail("Archive with missing extension file should not be opened.");
+    } catch (UnsupportedArchiveException e) {}
+
+    try {
+      Archive arch = ArchiveFactory.openArchive(FileUtils.getClasspathFile("invalid/extension-encoding-missing"));
+      fail("Archive with missing extension file encoding in meta.xml should not be opened.");
+    } catch (UnsupportedArchiveException e) {}
+
+    try {
+      Archive arch = ArchiveFactory.openArchive(FileUtils.getClasspathFile("invalid/extension-location-missing"));
+      fail("Archive with missing extension file location in meta.xml should not be opened.");
+    } catch (UnsupportedArchiveException e) {}
+
+    try {
+      Archive arch = ArchiveFactory.openArchive(FileUtils.getClasspathFile("invalid/extension-core-id-missing"));
+      fail("Archive with extension lacking coreid in meta.xml should not be opened.");
+    } catch (UnsupportedArchiveException e) {}
+
+    try {
+      Archive arch = ArchiveFactory.openArchive(FileUtils.getClasspathFile("invalid/extension-id-missing"));
+      fail("Archive with extension and core missing id in meta.xml should not be opened.");
+    } catch (UnsupportedArchiveException e) {}
+  }
 }
diff --git a/src/test/resources/invalid/empty/.keep b/src/test/resources/invalid/empty/.keep
new file mode 100644
index 000000000..e69de29bb
diff --git a/src/test/resources/invalid/extension-core-id-missing/media.txt b/src/test/resources/invalid/extension-core-id-missing/media.txt
new file mode 100644
index 000000000..e69de29bb
diff --git a/src/test/resources/invalid/extension-core-id-missing/meta.xml b/src/test/resources/invalid/extension-core-id-missing/meta.xml
new file mode 100644
index 000000000..e7560624d
--- /dev/null
+++ b/src/test/resources/invalid/extension-core-id-missing/meta.xml
@@ -0,0 +1,20 @@
+<?xml version='1.0' encoding='utf-8'?>
+<archive xmlns="http://rs.tdwg.org/dwc/text/">
+	<core encoding="UTF-8" fieldsTerminatedBy="\t" linesTerminatedBy="\n" fieldsEnclosedBy='' ignoreHeaderLines="1" rowType="http://rs.tdwg.org/dwc/terms/Taxon">
+		<files>
+			<location>taxa.txt</location>
+		</files>
+		<id index="0"/>
+		<field index="1" term="http://rs.tdwg.org/dwc/terms/scientificName"/>
+		<field index="2" term="http://rs.tdwg.org/dwc/terms/taxonRank"/>
+	</core>
+
+	<extension encoding="UTF-8" fieldsTerminatedBy="\t" linesTerminatedBy="\n" fieldsEnclosedBy='' ignoreHeaderLines="1" rowType="http://rs.gbif.org/terms/1.0/Multimedia">
+		<files>
+			<location>media.txt</location>
+		</files>
+		<coreid index="0"/>
+		<field index="1" term="http://purl.org/dc/terms/identifier"/>
+		<field default="StillImage" term="http://purl.org/dc/terms/type"/>
+	</extension>
+</archive>
diff --git a/src/test/resources/invalid/extension-encoding-missing/media.txt b/src/test/resources/invalid/extension-encoding-missing/media.txt
new file mode 100644
index 000000000..e69de29bb
diff --git a/src/test/resources/invalid/extension-encoding-missing/meta.xml b/src/test/resources/invalid/extension-encoding-missing/meta.xml
new file mode 100644
index 000000000..7a24cedbc
--- /dev/null
+++ b/src/test/resources/invalid/extension-encoding-missing/meta.xml
@@ -0,0 +1,20 @@
+<?xml version='1.0' encoding='utf-8'?>
+<archive xmlns="http://rs.tdwg.org/dwc/text/">
+	<core encoding="UTF-8" fieldsTerminatedBy="\t" linesTerminatedBy="\n" fieldsEnclosedBy='' ignoreHeaderLines="1" rowType="http://rs.tdwg.org/dwc/terms/Taxon">
+		<files>
+			<location>taxa.txt</location>
+		</files>
+		<id index="0"/>
+		<field index="1" term="http://rs.tdwg.org/dwc/terms/scientificName"/>
+		<field index="2" term="http://rs.tdwg.org/dwc/terms/taxonRank"/>
+	</core>
+
+	<extension fieldsTerminatedBy="\t" linesTerminatedBy="\n" fieldsEnclosedBy='' ignoreHeaderLines="1" rowType="http://rs.gbif.org/terms/1.0/Multimedia">
+		<files>
+			<location>media.txt</location>
+		</files>
+		<coreid index="0"/>
+		<field index="1" term="http://purl.org/dc/terms/identifier"/>
+		<field default="StillImage" term="http://purl.org/dc/terms/type"/>
+	</extension>
+</archive>
diff --git a/src/test/resources/invalid/extension-id-missing/media.txt b/src/test/resources/invalid/extension-id-missing/media.txt
new file mode 100644
index 000000000..e69de29bb
diff --git a/src/test/resources/invalid/extension-id-missing/meta.xml b/src/test/resources/invalid/extension-id-missing/meta.xml
new file mode 100644
index 000000000..9f01650c4
--- /dev/null
+++ b/src/test/resources/invalid/extension-id-missing/meta.xml
@@ -0,0 +1,19 @@
+<?xml version='1.0' encoding='utf-8'?>
+<archive xmlns="http://rs.tdwg.org/dwc/text/">
+	<core encoding="UTF-8" fieldsTerminatedBy="\t" linesTerminatedBy="\n" fieldsEnclosedBy='' ignoreHeaderLines="1" rowType="http://rs.tdwg.org/dwc/terms/Taxon">
+		<files>
+			<location>taxa.txt</location>
+		</files>
+		<field index="1" term="http://rs.tdwg.org/dwc/terms/scientificName"/>
+		<field index="2" term="http://rs.tdwg.org/dwc/terms/taxonRank"/>
+	</core>
+
+	<extension encoding="UTF-8" fieldsTerminatedBy="\t" linesTerminatedBy="\n" fieldsEnclosedBy='' ignoreHeaderLines="1" rowType="http://rs.gbif.org/terms/1.0/Multimedia">
+		<files>
+			<location>media.txt</location>
+		</files>
+		<coreid index="0"/>
+		<field index="1" term="http://purl.org/dc/terms/identifier"/>
+		<field default="StillImage" term="http://purl.org/dc/terms/type"/>
+	</extension>
+</archive>
diff --git a/src/test/resources/invalid/extension-location-missing/media.txt b/src/test/resources/invalid/extension-location-missing/media.txt
new file mode 100644
index 000000000..e69de29bb
diff --git a/src/test/resources/invalid/extension-location-missing/meta.xml b/src/test/resources/invalid/extension-location-missing/meta.xml
new file mode 100644
index 000000000..975ea5154
--- /dev/null
+++ b/src/test/resources/invalid/extension-location-missing/meta.xml
@@ -0,0 +1,19 @@
+<?xml version='1.0' encoding='utf-8'?>
+<archive xmlns="http://rs.tdwg.org/dwc/text/">
+	<core encoding="UTF-8" fieldsTerminatedBy="\t" linesTerminatedBy="\n" fieldsEnclosedBy='' ignoreHeaderLines="1" rowType="http://rs.tdwg.org/dwc/terms/Taxon">
+		<files>
+			<location>taxa.txt</location>
+		</files>
+		<id index="0"/>
+		<field index="1" term="http://rs.tdwg.org/dwc/terms/scientificName"/>
+		<field index="2" term="http://rs.tdwg.org/dwc/terms/taxonRank"/>
+	</core>
+
+	<extension encoding="UTF-8" fieldsTerminatedBy="\t" linesTerminatedBy="\n" fieldsEnclosedBy='' ignoreHeaderLines="1" rowType="http://rs.gbif.org/terms/1.0/Multimedia">
+		<files>
+		</files>
+		<coreid index="0"/>
+		<field index="1" term="http://purl.org/dc/terms/identifier"/>
+		<field default="StillImage" term="http://purl.org/dc/terms/type"/>
+	</extension>
+</archive>
diff --git a/src/test/resources/invalid/extension-missing/meta.xml b/src/test/resources/invalid/extension-missing/meta.xml
new file mode 100644
index 000000000..e7560624d
--- /dev/null
+++ b/src/test/resources/invalid/extension-missing/meta.xml
@@ -0,0 +1,20 @@
+<?xml version='1.0' encoding='utf-8'?>
+<archive xmlns="http://rs.tdwg.org/dwc/text/">
+	<core encoding="UTF-8" fieldsTerminatedBy="\t" linesTerminatedBy="\n" fieldsEnclosedBy='' ignoreHeaderLines="1" rowType="http://rs.tdwg.org/dwc/terms/Taxon">
+		<files>
+			<location>taxa.txt</location>
+		</files>
+		<id index="0"/>
+		<field index="1" term="http://rs.tdwg.org/dwc/terms/scientificName"/>
+		<field index="2" term="http://rs.tdwg.org/dwc/terms/taxonRank"/>
+	</core>
+
+	<extension encoding="UTF-8" fieldsTerminatedBy="\t" linesTerminatedBy="\n" fieldsEnclosedBy='' ignoreHeaderLines="1" rowType="http://rs.gbif.org/terms/1.0/Multimedia">
+		<files>
+			<location>media.txt</location>
+		</files>
+		<coreid index="0"/>
+		<field index="1" term="http://purl.org/dc/terms/identifier"/>
+		<field default="StillImage" term="http://purl.org/dc/terms/type"/>
+	</extension>
+</archive>
diff --git a/src/test/resources/invalid/meta-file-encoding-missing/meta.xml b/src/test/resources/invalid/meta-file-encoding-missing/meta.xml
new file mode 100644
index 000000000..1e73bfa53
--- /dev/null
+++ b/src/test/resources/invalid/meta-file-encoding-missing/meta.xml
@@ -0,0 +1,11 @@
+<?xml version='1.0' encoding='utf-8'?>
+<archive xmlns="http://rs.tdwg.org/dwc/text/">
+	<core fieldsTerminatedBy="\t" linesTerminatedBy="\n" fieldsEnclosedBy='' ignoreHeaderLines="1" rowType="http://rs.tdwg.org/dwc/terms/Taxon">
+		<files>
+			<location>taxa.txt</location>
+		</files>
+		<id index="0"/>
+		<field index="1" term="http://rs.tdwg.org/dwc/terms/scientificName"/>
+		<field index="2" term="http://rs.tdwg.org/dwc/terms/taxonRank"/>
+	</core>
+</archive>
diff --git a/src/test/resources/invalid/meta-file-encoding-missing/taxa.txt b/src/test/resources/invalid/meta-file-encoding-missing/taxa.txt
new file mode 100644
index 000000000..e69de29bb
diff --git a/src/test/resources/invalid/meta-file-location-missing/meta.xml b/src/test/resources/invalid/meta-file-location-missing/meta.xml
new file mode 100644
index 000000000..d3d23cc64
--- /dev/null
+++ b/src/test/resources/invalid/meta-file-location-missing/meta.xml
@@ -0,0 +1,10 @@
+<?xml version='1.0' encoding='utf-8'?>
+<archive xmlns="http://rs.tdwg.org/dwc/text/">
+	<core encoding="UTF-8" fieldsTerminatedBy="\t" linesTerminatedBy="\n" fieldsEnclosedBy='' ignoreHeaderLines="1" rowType="http://rs.tdwg.org/dwc/terms/Taxon">
+		<files>
+		</files>
+		<id index="0"/>
+		<field index="1" term="http://rs.tdwg.org/dwc/terms/scientificName"/>
+		<field index="2" term="http://rs.tdwg.org/dwc/terms/taxonRank"/>
+	</core>
+</archive>
diff --git a/src/test/resources/invalid/meta-file-location-missing/taxa.txt b/src/test/resources/invalid/meta-file-location-missing/taxa.txt
new file mode 100644
index 000000000..e69de29bb
diff --git a/src/test/resources/invalid/meta-only/meta.xml b/src/test/resources/invalid/meta-only/meta.xml
new file mode 100644
index 000000000..e2c6f8c42
--- /dev/null
+++ b/src/test/resources/invalid/meta-only/meta.xml
@@ -0,0 +1,10 @@
+<?xml version='1.0' encoding='utf-8'?>
+<archive xmlns="http://rs.tdwg.org/dwc/text/">
+	<core encoding="UTF-8" fieldsTerminatedBy="\t" linesTerminatedBy="\n" fieldsEnclosedBy='' ignoreHeaderLines="1" rowType="http://rs.tdwg.org/dwc/terms/Taxon">
+		<files>
+			<location>taxa.txt</location>
+		</files>
+		<id index="0"/>
+		<field index="1" term="http://rs.tdwg.org/dwc/terms/scientificName"/>
+	</core>
+</archive>