feat(model): Detect and exclude binary license files

Adds logic to detect binary license files using Apache Tika. If a
non-text file is found during the license info creation process, a
warning is printed, and it is not included in the final report.

This avoids including binary files that previously caused the reporter
to enter an endless loop during report generation.

Signed-off-by: Julian Olderdissen <[email protected]>

Author: Juli0q

model/src/main/kotlin/utils/FileArchiver.kt

@@ -25,6 +25,7 @@ import kotlin.time.measureTime
 import kotlin.time.measureTimedValue
 
 import org.apache.logging.log4j.kotlin.logger
+import org.apache.tika.Tika
 
 import org.ossreviewtoolkit.model.KnownProvenance
 import org.ossreviewtoolkit.utils.common.FileMatcher
@@ -60,6 +61,8 @@ class FileArchiver(
         ignoreCase = true
     )
 
+    private val tika = Tika()
+
     /**
      * Return whether an archive corresponding to [provenance] exists.
      */
@@ -77,6 +80,11 @@ class FileArchiver(
             directory.packZip(zipFile, overwrite = true) { file ->
                 val relativePath = file.relativeTo(directory).invariantSeparatorsPath
 
+                if (!file.isValidTextFile()) {
+                    logger.warn { "Not adding file '$relativePath' to archive because it is not a text file." }
+                    return@packZip false
+                }
+
                 matcher.matches(relativePath).also { result ->
                     logger.debug {
                         if (result) {
@@ -125,4 +133,15 @@ class FileArchiver(
             logger.error { "Failed to unarchive data for $provenance: ${it.collectMessages()}" }
         }.isSuccess
     }
+
+    internal fun File.isValidTextFile(): Boolean {
+        val detected = tika.detect(this)
+        val invalidTypes = listOf(
+            "application/octet-stream",
+            "application/x-executable",
+            "application/x-sharedlib",
+            "application/x-object"
+        )
+        return invalidTypes.none { detected.startsWith(it, ignoreCase = true) }
+    }
 }

model/src/test/kotlin/utils/FileArchiverTest.kt

@@ -24,7 +24,9 @@ import io.kotest.core.test.TestCase
 import io.kotest.engine.spec.tempdir
 import io.kotest.matchers.file.aFile
 import io.kotest.matchers.file.exist
+import io.kotest.matchers.file.shouldContainFile
 import io.kotest.matchers.file.shouldContainNFiles
+import io.kotest.matchers.file.shouldNotContainFile
 import io.kotest.matchers.shouldBe
 import io.kotest.matchers.shouldNot
 
@@ -60,10 +62,11 @@ class FileArchiverTest : StringSpec() {
         storage = FileProvenanceFileStorage(LocalFileStorage(storageDir), FileArchiverConfiguration.ARCHIVE_FILENAME)
     }
 
-    private fun createFile(path: String) {
+    private fun createFile(path: String): File {
         val file = workingDir.resolve(path)
         file.parentFile.safeMkdirs()
         file.writeText(path)
+        return file
     }
 
     /**
@@ -162,5 +165,33 @@ class FileArchiverTest : StringSpec() {
             archiver.unarchive(targetDir, PROVENANCE) shouldBe true
             targetDir shouldContainNFiles 0
         }
+
+        "exclude basic binary license file" {
+            createFile("License").writeBytes(byteArrayOf(0xFF.toByte(), 0xD8.toByte()))
+
+            val archiver = FileArchiver.createDefault()
+
+            archiver.archive(workingDir, PROVENANCE)
+
+            val result = archiver.unarchive(targetDir, PROVENANCE)
+
+            result shouldBe true
+            targetDir.shouldNotContainFile("License")
+        }
+
+        "include utf8 file with japanese chars" {
+            createFile("License").writeText("ぁあぃいぅうぇえぉおかがきぎ", charset = Charsets.UTF_8)
+
+            val archiver = FileArchiver.createDefault()
+
+            archiver.archive(workingDir, PROVENANCE)
+
+            val result = archiver.unarchive(targetDir, PROVENANCE)
+
+            result shouldBe true
+            with(targetDir) {
+                shouldContainFile("License")
+            }
+        }
     }
 }