Fixed issues with progsnap converter

Author: Daniil.Karol

processing/src/main/kotlin/org/jetbrains/research/tasktracker/progsnap2/Main.kt

@@ -0,0 +1,57 @@
+package org.jetbrains.research.tasktracker.progsnap2
+
+import java.io.File
+
+fun main(args: Array<String>) {
+    val (trackedFilesDirectory, destinationDirectory) = gatherArguments(args)
+    with(trackedFilesDirectory) {
+        ProgsnapParser(
+            TaskTrackerData(
+                resolve(ACTIVITY_DATA_CSV),
+                resolve(DOCUMENT_DATA_CSV),
+                resolve(FILE_EDITOR_DATA_CSV),
+                resolve(RESEARCHES_CSV),
+                resolve(TOOL_WINDOW_DATA_CSV),
+                ProgsnapMeta(
+                    PROGSNAP_VERSION,
+                    CodeStateRepresentation.DIRECTORY,
+                    EventOrderScope.GLOBAL,
+                    true,
+                    "SubjectID"
+                )
+            )
+        ).convert(destinationDirectory)
+    }
+}
+
+private fun gatherArguments(args: Array<String>): Pair<File, File> {
+    require(args.size == 2) { "Exactly 2 arguments are required: <csvFolder> <existingFolder>." }
+
+    val csvFolder = File(args[0])
+    require(csvFolder.isDirectory) { "The first argument must be a directory containing CSV files." }
+    val missingFiles = expectedCsvFiles.filter { !File(csvFolder, it).exists() }
+    require(missingFiles.isEmpty()) {
+        "The following required CSV files are missing: ${missingFiles.joinToString(", ")}"
+    }
+
+    val existingDirectory = File(args[1])
+    require(existingDirectory.isDirectory) { "The destination directory argument must be an existing directory." }
+
+    return Pair(csvFolder, existingDirectory)
+}
+
+private const val ACTIVITY_DATA_CSV = "activitydata.csv"
+private const val DOCUMENT_DATA_CSV = "documentdata.csv"
+private const val FILE_EDITOR_DATA_CSV = "fileeditordata.csv"
+private const val RESEARCHES_CSV = "researches.csv"
+private const val TOOL_WINDOW_DATA_CSV = "toolwindowdata.csv"
+
+private val expectedCsvFiles = listOf(
+    ACTIVITY_DATA_CSV,
+    DOCUMENT_DATA_CSV,
+    FILE_EDITOR_DATA_CSV,
+    RESEARCHES_CSV,
+    TOOL_WINDOW_DATA_CSV
+)
+
+private const val PROGSNAP_VERSION = 5

processing/src/main/kotlin/org/jetbrains/research/tasktracker/progsnap2/ProgsnapParser.kt

@@ -9,13 +9,17 @@ import org.jetbrains.kotlinx.dataframe.io.writeCSV
 import java.io.File
 import java.time.OffsetDateTime
 import java.time.format.DateTimeFormatter
+import java.time.format.DateTimeParseException
 import kotlin.io.path.createDirectories
 import kotlin.io.path.exists
 
+@Suppress("TooManyFunctions")
 class ProgsnapParser(private val taskTrackerData: TaskTrackerData) {
 
     companion object {
-        private val DATE_FORMAT: DateTimeFormatter = DateTimeFormatter.ofPattern("yyyy-MM-dd HH:mm:ss.SSSSSS XXX")
+        private val DATE_FORMAT_WITH_MICROS = DateTimeFormatter.ofPattern("yyyy-MM-dd HH:mm:ss.SSSSSSXXX")
+        private val DATE_WITHOUT_MICROS = DateTimeFormatter.ofPattern("yyyy-MM-dd HH:mm:ssXXX")
+        private val DATE_FORMATS = listOf(DATE_FORMAT_WITH_MICROS, DATE_WITHOUT_MICROS)
         private val DIRECTORIES_DELIMITERS = arrayOf("/", "\\")
         private const val CODE_STATES_DIRECTORY_NAME = "CodeStates"
         private const val MAIN_TABLE_FILENAME = "MainTable.csv"
@@ -29,6 +33,17 @@ class ProgsnapParser(private val taskTrackerData: TaskTrackerData) {
         createMainTableFile(destinationDirectory, researchGroups)
     }
 
+    private fun parseDate(dateString: String): OffsetDateTime {
+        for (formatter in DATE_FORMATS) {
+            try {
+                return OffsetDateTime.parse(dateString, formatter)
+            } catch (_: DateTimeParseException) {
+                // Ignore and try the next format
+            }
+        }
+        error("Invalid date format: $dateString")
+    }
+
     private fun createMetaFile(destinationDirectory: File) {
         val metaDataFile = destinationDirectory.resolve(META_DATA_FILENAME)
         taskTrackerData.metaData.toDataFrame().writeCSV(metaDataFile)
@@ -63,13 +78,9 @@ class ProgsnapParser(private val taskTrackerData: TaskTrackerData) {
     private var researches =
         taskTrackerData.researches.toDataFrame()
             .convert("id").to<Int>()
-    private var surveyData =
-        taskTrackerData.surveyData.toDataFrame()
     private var toolWindowData =
         taskTrackerData.toolWindowData.toDataFrame()
             .add("data_type") { "toolWindowData" }
-    private var users =
-        taskTrackerData.users.toDataFrame()
 
     private fun File.toDataFrame() = DataFrame.readCSV(this)
 
@@ -78,7 +89,7 @@ class ProgsnapParser(private val taskTrackerData: TaskTrackerData) {
     private fun DataFrame<*>.convertDate() = this
         .dropNA("date")
         .update("date")
-        .with { OffsetDateTime.parse(it.toString(), DATE_FORMAT).toString() }
+        .with { parseDate(it.toString()).toString() }
         .convert("research_id").to<Int>()
         .fullJoin(researches) { "research_id" match "id" }
         .dropNA("date")
@@ -159,6 +170,9 @@ class ProgsnapParser(private val taskTrackerData: TaskTrackerData) {
                     "KeyPressed" -> "X-KeyPressed"
                     "KeyReleased" -> "X-KeyReleased"
                     "Action" -> "X-Action"
+                    "MouseClicked" -> "X-MouseClicked"
+                    "MouseMoved" -> "X-MouseMoved"
+                    "MouseWheel" -> "X-MouseWheel"
                     else -> error("Undefined activity data type has been detected in the data `$type`")
                 }
             }

processing/src/main/kotlin/org/jetbrains/research/tasktracker/progsnap2/TaskTrackerData.kt

@@ -9,9 +9,7 @@ data class TaskTrackerData(
     val documentData: File,
     val fileEditorData: File,
     val researches: File,
-    val surveyData: File,
     val toolWindowData: File,
-    val users: File,
     val metaData: ProgsnapMeta
 )