Remove certain maven metadata from uploaded dependencies. (aws#4704)

* Remove certain maven metadata from uploaded dependencies.

* Add test ensuring correct dependencies added to Q Code Transform upload zip.

* Address comments and fix ci.

* Add changelog.

---------

Co-authored-by: Leonardo Araneda Freccero <[email protected]>

Author: dhasani23

.changes/next-release/Bug Fix-ffcf9c8e-bd5e-4b34-a418-469717d913f6.json

@@ -0,0 +1,4 @@
+{
+	"type": "Bug Fix",
+	"description": "Amazon Q Code Transformation - Omit Maven metadata files when uploading dependencies to fix certain build failures in backend."
+}

packages/core/src/codewhisperer/service/transformByQ/transformApiHandler.ts

@@ -195,6 +195,27 @@ export async function uploadPayload(payloadFileName: string) {
     return response.uploadId
 }
 
+/**
+ * Array of file extensions used by Maven as metadata in the local repository.
+ * Files with these extensions influence Maven's behavior during compile time,
+ * particularly in checking the availability of source repositories and potentially
+ * re-downloading dependencies if the source is not accessible. Removing these
+ * files can prevent Maven from attempting to download dependencies again.
+ */
+const MavenExcludedExtensions = ['.repositories', '.sha1']
+
+/**
+ * Determines if the specified file path corresponds to a Maven metadata file
+ * by checking against known metadata file extensions. This is used to identify
+ * files that might trigger Maven to recheck or redownload dependencies from source repositories.
+ *
+ * @param path The file path to evaluate for exclusion based on its extension.
+ * @returns {boolean} Returns true if the path ends with an extension associated with Maven metadata files; otherwise, false.
+ */
+function isExcludedDependencyFile(path: string): boolean {
+    return MavenExcludedExtensions.some(extension => path.endsWith(extension))
+}
+
 /**
  * Gets all files in dir. We use this method to get the source code, then we run a mvn command to
  * copy over dependencies into their own folder, then we use this method again to get those
@@ -252,6 +273,9 @@ export async function zipCode(dependenciesFolder: FolderInfo) {
 
         if (dependencyFiles.length > 0) {
             for (const file of dependencyFiles) {
+                if (isExcludedDependencyFile(file)) {
+                    continue
+                }
                 const relativePath = path.relative(dependenciesFolder.path, file)
                 const paddedPath = path.join(`dependencies/${dependenciesFolder.name}`, relativePath)
                 zip.addLocalFile(file, path.dirname(paddedPath))

packages/core/src/test/codewhisperer/commands/transformByQ.test.ts

@@ -5,13 +5,16 @@
 
 import assert from 'assert'
 import * as vscode from 'vscode'
+import * as fs from 'fs-extra'
 import * as sinon from 'sinon'
+import { makeTemporaryToolkitFolder } from '../../../shared/filesystemUtilities'
 import * as model from '../../../codewhisperer/models/model'
 import * as startTransformByQ from '../../../codewhisperer/commands/startTransformByQ'
 import { HttpResponse } from 'aws-sdk'
 import * as codeWhisperer from '../../../codewhisperer/client/codewhisperer'
 import * as CodeWhispererConstants from '../../../codewhisperer/models/constants'
 import { getTestWindow } from '../../shared/vscode/window'
+import AdmZip from 'adm-zip'
 import { stopTransformByQMessage } from '../../../codewhisperer/models/constants'
 import { convertToTimeString, convertDateToTimestamp } from '../../../shared/utilities/textUtilities'
 import path from 'path'
@@ -26,6 +29,7 @@ import {
     pollTransformationJob,
     getHeadersObj,
     throwIfCancelled,
+    zipCode,
 } from '../../../codewhisperer/service/transformByQ/transformApiHandler'
 import {
     validateOpenProjects,
@@ -34,8 +38,15 @@ import {
 import { TransformationCandidateProject } from '../../../codewhisperer/models/model'
 
 describe('transformByQ', function () {
-    afterEach(function () {
+    let tempDir: string
+
+    beforeEach(async function () {
+        tempDir = await makeTemporaryToolkitFolder()
+    })
+
+    afterEach(async function () {
         sinon.restore()
+        await fs.remove(tempDir)
     })
 
     it('WHEN converting short duration in milliseconds THEN converts correctly', async function () {
@@ -199,4 +210,54 @@ describe('transformByQ', function () {
         }
         assert.deepStrictEqual(actual, expected)
     })
+
+    it(`WHEN zip created THEN dependencies contains no .sha1 or .repositories files`, async function () {
+        const m2Folders = [
+            'com/groupid1/artifactid1/version1',
+            'com/groupid1/artifactid1/version2',
+            'com/groupid1/artifactid2/version1',
+            'com/groupid2/artifactid1/version1',
+            'com/groupid2/artifactid1/version2',
+        ]
+        // List of files that exist in m2 artifact directory
+        const filesToAdd = [
+            '_remote.repositories',
+            'test-0.0.1-20240315.145420-18.pom',
+            'test-0.0.1-20240315.145420-18.pom.sha1',
+            'test-0.0.1-SNAPSHOT.pom',
+            'maven-metadata-test-repo.xml',
+            'maven-metadata-test-repo.xml.sha1',
+            'resolver-status.properties',
+        ]
+        const expectedFilesAfterClean = [
+            'test-0.0.1-20240315.145420-18.pom',
+            'test-0.0.1-SNAPSHOT.pom',
+            'maven-metadata-test-repo.xml',
+            'resolver-status.properties',
+        ]
+
+        m2Folders.forEach(folder => {
+            const folderPath = path.join(tempDir, folder)
+            fs.mkdirSync(folderPath, { recursive: true })
+            filesToAdd.forEach(file => {
+                fs.writeFileSync(path.join(folderPath, file), 'sample content for the test file')
+            })
+        })
+
+        const tempFileName = `testfile-${Date.now()}.zip`
+        model.transformByQState.setProjectPath(tempDir)
+        return zipCode({
+            path: tempDir,
+            name: tempFileName,
+        }).then(zipFile => {
+            const zip = new AdmZip(zipFile)
+            const dependenciesToUpload = zip.getEntries().filter(entry => entry.entryName.startsWith('dependencies'))
+            // Each dependency version folder contains each expected file, thus we multiply
+            const expectedNumberOfDependencyFiles = m2Folders.length * expectedFilesAfterClean.length
+            assert.strictEqual(expectedNumberOfDependencyFiles, dependenciesToUpload.length)
+            dependenciesToUpload.forEach(dependency => {
+                assert(expectedFilesAfterClean.includes(dependency.name))
+            })
+        })
+    })
 })