fix(amazonq): fix for /test project payload collection filter (#5305)

* Fix to decouple gitignore logic to run it module level.

Author: ashishrp-aws

.changes/next-release/bugfix-52963f8d-8f8e-4b53-b7dc-0ef4819b5976.json

@@ -0,0 +1,4 @@
+{
+  "type" : "bugfix",
+  "description" : "Amazon Q /test: Fixed an issue which incorrectly caused payload size exceeded exception when collecting project payload files"
+}

plugins/amazonq/codewhisperer/jetbrains-community/src/software/aws/toolkits/jetbrains/services/codewhisperer/codetest/sessionconfig/CodeTestSessionConfig.kt

@@ -18,7 +18,6 @@ import software.aws.toolkits.core.utils.createTemporaryZipFile
 import software.aws.toolkits.core.utils.debug
 import software.aws.toolkits.core.utils.getLogger
 import software.aws.toolkits.core.utils.putNextEntry
-import software.aws.toolkits.jetbrains.services.amazonq.FeatureDevSessionContext
 import software.aws.toolkits.jetbrains.services.codewhisperer.codescan.sessionconfig.Payload
 import software.aws.toolkits.jetbrains.services.codewhisperer.codescan.sessionconfig.PayloadContext
 import software.aws.toolkits.jetbrains.services.codewhisperer.codescan.sessionconfig.PayloadMetadata
@@ -33,6 +32,7 @@ import software.aws.toolkits.jetbrains.services.codewhisperer.language.programmi
 import software.aws.toolkits.jetbrains.services.codewhisperer.util.CodeWhispererConstants.CODE_SCAN_CREATE_PAYLOAD_TIMEOUT_IN_SECONDS
 import software.aws.toolkits.jetbrains.services.codewhisperer.util.CodeWhispererConstants.DEFAULT_CODE_SCAN_TIMEOUT_IN_SECONDS
 import software.aws.toolkits.jetbrains.services.codewhisperer.util.CodeWhispererConstants.DEFAULT_PAYLOAD_LIMIT_IN_BYTES
+import software.aws.toolkits.jetbrains.services.codewhisperer.util.GitIgnoreFilteringUtil
 import software.aws.toolkits.jetbrains.services.codewhisperer.util.isWithin
 import software.aws.toolkits.resources.message
 import java.io.File
@@ -53,8 +53,6 @@ class CodeTestSessionConfig(
         project.guessProjectDir() ?: error("Cannot guess base directory for project ${project.name}")
     }
 
-    private val featureDevSessionContext = FeatureDevSessionContext(project)
-
     val fileIndex = ProjectRootManager.getInstance(project).fileIndex
 
     /**
@@ -186,15 +184,16 @@ class CodeTestSessionConfig(
 
         moduleLoop@ for (module in project.modules) {
             val changeListManager = ChangeListManager.getInstance(module.project)
-            if (module.guessModuleDir() != null) {
-                stack.push(module.guessModuleDir())
+            module.guessModuleDir()?.let { moduleDir ->
+                val gitIgnoreFilteringUtil = GitIgnoreFilteringUtil(moduleDir)
+                stack.push(moduleDir)
                 while (stack.isNotEmpty()) {
                     val current = stack.pop()
 
                     if (!current.isDirectory) {
                         if (current.isFile && current.path != selectedFile?.path &&
                             !changeListManager.isIgnoredFile(current) &&
-                            runBlocking { !featureDevSessionContext.ignoreFile(current) } &&
+                            runBlocking { !gitIgnoreFilteringUtil.ignoreFile(current) } &&
                             runReadAction { !fileIndex.isInLibrarySource(current) }
                         ) {
                             if (willExceedPayloadLimit(currentTotalFileSize, current.length)) {
@@ -217,7 +216,7 @@ class CodeTestSessionConfig(
                     } else {
                         // Directory case: only traverse if not ignored
                         if (!changeListManager.isIgnoredFile(current) &&
-                            runBlocking { !featureDevSessionContext.ignoreFile(current) } &&
+                            runBlocking { !gitIgnoreFilteringUtil.ignoreFile(current) } &&
                             !traversedDirectories.contains(current) && current.isValid &&
                             runReadAction { !fileIndex.isInLibrarySource(current) }
                         ) {
@@ -242,21 +241,14 @@ class CodeTestSessionConfig(
         return PayloadMetadata(files, currentTotalFileSize, currentTotalLines, maxCountLanguage.toTelemetryType())
     }
 
-    fun getPath(root: String, relativePath: String = ""): Path? = try {
-        Path.of(root, relativePath).normalize()
-    } catch (e: Exception) {
-        LOG.debug { "Cannot find file at path $relativePath relative to the root $root" }
-        null
-    }
-
     fun getRelativePath(): Path? = try {
         selectedFile?.path?.let { Path.of(projectRoot.path).relativize(Path.of(it)).normalize() }
     } catch (e: Exception) {
         LOG.debug { "Cannot calculate relative path of $selectedFile with respect to $projectRoot" }
         null
     }
 
-    fun File.toVirtualFile() = LocalFileSystem.getInstance().findFileByIoFile(this)
+    private fun File.toVirtualFile() = LocalFileSystem.getInstance().findFileByIoFile(this)
 
     companion object {
         private val LOG = getLogger<CodeTestSessionConfig>()

plugins/amazonq/codewhisperer/jetbrains-community/src/software/aws/toolkits/jetbrains/services/codewhisperer/util/GitIgnoreFilteringUtil.kt

@@ -0,0 +1,102 @@
+// Copyright 2025 Amazon.com, Inc. or its affiliates. All Rights Reserved.
+// SPDX-License-Identifier: Apache-2.0
+
+package software.aws.toolkits.jetbrains.services.codewhisperer.util
+
+import com.intellij.openapi.vfs.VfsUtil
+import com.intellij.openapi.vfs.VirtualFile
+import kotlinx.coroutines.async
+import kotlinx.coroutines.withContext
+import kotlin.coroutines.coroutineContext
+
+class GitIgnoreFilteringUtil(private val moduleDir: VirtualFile) {
+    private var ignorePatternsWithGitIgnore = emptyList<Regex>()
+    private val additionalGitIgnoreRules = setOf(
+        ".aws-sam",
+        ".gem",
+        ".git",
+        ".gitignore",
+        ".gradle",
+        ".hg",
+        ".idea",
+        ".project",
+        ".rvm",
+        ".svn",
+        "*.zip",
+        "*.bin",
+        "*.png",
+        "*.jpg",
+        "*.svg",
+        "*.pyc",
+        "license.txt",
+        "License.txt",
+        "LICENSE.txt",
+        "license.md",
+        "License.md",
+        "LICENSE.md",
+        "node_modules",
+        "build",
+        "dist",
+        "annotation-generated-src",
+        "annotation-generated-tst"
+    )
+
+    init {
+        ignorePatternsWithGitIgnore = try {
+            buildList {
+                addAll(additionalGitIgnoreRules.map { convertGitIgnorePatternToRegex(it) })
+                addAll(parseGitIgnore())
+            }.mapNotNull { pattern ->
+                runCatching { Regex(pattern) }.getOrNull()
+            }
+        } catch (e: Exception) {
+            emptyList()
+        }
+    }
+
+    private fun parseGitIgnore(): Set<String> {
+        val gitignoreFile = moduleDir.findChild(".gitignore")
+        return gitignoreFile?.let {
+            if (it.isValid && it.exists()) {
+                it.inputStream.bufferedReader().readLines()
+                    .filter { line ->
+                        line.isNotBlank() && !line.startsWith("#")
+                    }
+                    .map { pattern ->
+                        convertGitIgnorePatternToRegex(pattern.trim())
+                    }
+                    .toSet()
+            } else {
+                emptySet()
+            }
+        }.orEmpty()
+    }
+
+    // gitignore patterns are not regex, method update needed.
+    private fun convertGitIgnorePatternToRegex(pattern: String): String = pattern
+        .replace(".", "\\.")
+        .replace("*", ".*")
+        .let { if (it.endsWith("/")) "$it.*" else "$it/.*" } // Add a trailing /* to all patterns. (we add a trailing / to all files when matching)
+
+    suspend fun ignoreFile(file: VirtualFile): Boolean {
+        // this method reads like something a JS dev would write and doesn't do what the author thinks
+        val deferredResults = ignorePatternsWithGitIgnore.map { pattern ->
+            withContext(coroutineContext) {
+                // avoid partial match (pattern.containsMatchIn) since it causes us matching files
+                // against folder patterns. (e.g. settings.gradle ignored by .gradle rule!)
+                // we convert the glob rules to regex, add a trailing /* to all rules and then match
+                // entries against them by adding a trailing /.
+                // TODO: Add unit tests for gitignore matching
+                val relative = getRelativePath(file)
+                async { pattern.matches("$relative/") }
+            }
+        }
+
+        // this will serially iterate over and block
+        // ideally we race the results https://github.com/Kotlin/kotlinx.coroutines/issues/2867
+        // i.e. Promise.any(...)
+        return deferredResults.any { it.await() }
+    }
+
+    private fun getRelativePath(file: VirtualFile): String = VfsUtil.getRelativePath(file, moduleDir).orEmpty()
+}