Added hash to CPG metadata (#19)

The hash is calculated as a checksum of all *.js input files.
This is done after transpiling. Otherwise, it may happen, that we end up with
no files for the hash calculation at all (e.g., for TS projects).

src/main/scala/io/shiftleft/js2cpg/core/Js2Cpg.scala

@@ -198,12 +198,13 @@ class Js2Cpg {
     val privateKeyFilePassPool = otherPools(3)
     val htmlAsConfigPassPool   = otherPools(4)
 
-    val cpg = newEmptyCpg(Some(config.outputFile))
+    val cpg  = newEmptyCpg(Some(config.outputFile))
+    val hash = FileUtils.md5(jsFilesWithRoot.map(_._1))
 
     new AstCreationPass(File(config.srcDir), jsFilesWithRoot, cpg, functionKeyPool, report)
       .createAndApply()
 
-    new JsMetaDataPass(cpg, metaDataKeyPool).createAndApply()
+    new JsMetaDataPass(cpg, metaDataKeyPool, hash).createAndApply()
     new BuiltinTypesPass(cpg, builtinTypesKeyPool).createAndApply()
     new DependenciesPass(cpg, config, dependenciesKeyPool)
       .createAndApply()

src/main/scala/io/shiftleft/js2cpg/cpg/passes/JsMetaDataPass.scala

@@ -1,19 +1,21 @@
 package io.shiftleft.js2cpg.cpg.passes
 
 import io.shiftleft.codepropertygraph.Cpg
-import io.shiftleft.codepropertygraph.generated.{Languages, nodes}
+import io.shiftleft.codepropertygraph.generated.Languages
+import io.shiftleft.codepropertygraph.generated.nodes.NewMetaData
 import io.shiftleft.passes.{CpgPass, DiffGraph, KeyPool}
 import org.slf4j.LoggerFactory
 
-class JsMetaDataPass(cpg: Cpg, keyPool: KeyPool) extends CpgPass(cpg, keyPool = Some(keyPool)) {
+class JsMetaDataPass(cpg: Cpg, keyPool: KeyPool, hash: String)
+    extends CpgPass(cpg, keyPool = Some(keyPool)) {
 
   private val logger = LoggerFactory.getLogger(getClass)
 
   override def run(): Iterator[DiffGraph] = {
     logger.debug(s"Generating meta-data.")
 
     val diffGraph = DiffGraph.newBuilder
-    val metaNode  = nodes.NewMetaData().language(Languages.JAVASCRIPT)
+    val metaNode  = NewMetaData().language(Languages.JAVASCRIPT).hash(hash)
     diffGraph.addNode(metaNode)
     Iterator(diffGraph.build())
   }

src/main/scala/io/shiftleft/js2cpg/io/FileUtils.scala

@@ -5,14 +5,15 @@ import better.files.File
 import java.io.Reader
 import java.math.BigInteger
 import java.nio.charset.{CharsetDecoder, CodingErrorAction}
-import java.nio.file.{Files, FileVisitResult, Path, SimpleFileVisitor}
+import java.nio.file.{FileVisitResult, Files, Path, SimpleFileVisitor}
 import io.shiftleft.js2cpg.core.Config
 import io.shiftleft.js2cpg.io.FileDefaults._
 import org.slf4j.LoggerFactory
 
 import java.nio.file.attribute.BasicFileAttributes
+import java.security.{DigestInputStream, MessageDigest}
 import scala.collection.concurrent.TrieMap
-import scala.collection.{mutable, SortedMap}
+import scala.collection.{SortedMap, mutable}
 import scala.io.{BufferedSource, Codec, Source}
 import scala.jdk.CollectionConverters._
 
@@ -23,6 +24,18 @@ object FileUtils {
   // we only want to print excluded files and folders once (Path -> Reason as String)
   private val excludedPaths = TrieMap.empty[Path, String]
 
+  def md5(files: Seq[Path]): String = {
+    val md = MessageDigest.getInstance("MD5")
+    files.sortBy(_.toRealPath().toString).foreach { path =>
+      val dis = new DigestInputStream(Files.newInputStream(path), md)
+      while (dis.available() > 0) {
+        dis.read()
+      }
+      dis.close()
+    }
+    md.digest().map(b => String.format("%02x", Byte.box(b))).mkString
+  }
+
   def logAndClearExcludedPaths(): Unit = {
     excludedPaths.foreach {
       case (path, reason) =>

src/test/scala/io/shiftleft/js2cpg/cpg/passes/JsMetaDataPassTest.scala

@@ -13,7 +13,7 @@ class JsMetaDataPassTest extends AbstractPassTest {
     val cpg               = Cpg.emptyCpg
     val jsMetaDataKeyPool = new IntervalKeyPool(1, 100)
 
-    new JsMetaDataPass(cpg, jsMetaDataKeyPool).createAndApply()
+    new JsMetaDataPass(cpg, jsMetaDataKeyPool, "somehash").createAndApply()
 
     "create exactly 1 node" in {
       cpg.graph.V.asScala.size shouldBe 1
@@ -26,6 +26,10 @@ class JsMetaDataPassTest extends AbstractPassTest {
     "create a metadata node with correct language" in {
       cpg.metaData.language.l shouldBe List(Languages.JAVASCRIPT)
     }
+
+    "create a metadata node with a hash" in {
+      cpg.metaData.hash.l shouldBe List("somehash")
+    }
   }
 
 }