From dcbf6b5e513aee96d4a385e988af25853521d499 Mon Sep 17 00:00:00 2001
From: Mathew Ruberg <yezr@users.noreply.github.com>
Date: Mon, 9 Dec 2024 11:26:55 -0500
Subject: [PATCH] 898: Dominion CvrExport parsing (#899)

* get all CvrExport_N.json files to parse

* linter cleanup

* linter cleanup 2

* more concise regex to find all dominion files and sort them consistently

* complete rename of cvr_0 to cvr_6

* null check dominion cvr file list

---------

Co-authored-by: yezr <8996546+yezr@users.noreply.github.com>
---
 .../brightspots/rcv/DominionCvrReader.java    | 41 ++++++++++++-------
 .../{CvrExport_0.json => CvrExport_6.json}    |  0
 2 files changed, 27 insertions(+), 14 deletions(-)
 rename src/test/resources/network/brightspots/rcv/test_data/dominion_multi_file/dominion_multi_file_input_data/{CvrExport_0.json => CvrExport_6.json} (100%)

diff --git a/src/main/java/network/brightspots/rcv/DominionCvrReader.java b/src/main/java/network/brightspots/rcv/DominionCvrReader.java
index bc1912454..6ed6cf86b 100644
--- a/src/main/java/network/brightspots/rcv/DominionCvrReader.java
+++ b/src/main/java/network/brightspots/rcv/DominionCvrReader.java
@@ -25,6 +25,8 @@
 import java.nio.file.Path;
 import java.nio.file.Paths;
 import java.util.ArrayList;
+import java.util.Arrays;
+import java.util.Comparator;
 import java.util.HashMap;
 import java.util.HashSet;
 import java.util.List;
@@ -230,30 +232,41 @@ private void validateNamesAreInContest(List<CastVoteRecord> castVoteRecords)
   private void gatherCvrsForContest(List<CastVoteRecord> castVoteRecords, String contestIdToLoad) {
     try {
       Path singleCvrPath = Paths.get(cvrPath, CVR_EXPORT);
-      Path firstCvrPath = Paths.get(cvrPath, String.format(CVR_EXPORT_PATTERN, 0));
+
       if (singleCvrPath.toFile().exists()) {
         HashMap json = JsonParser.readFromFile(singleCvrPath.toString(), HashMap.class);
         parseCvrFile(json, castVoteRecords, contestIdToLoad);
-      } else if (firstCvrPath.toFile().exists()) {
+      } else {
+        // We are expecting multiple CvrExport_N.json files
+        String regexPath = CVR_EXPORT_PATTERN.replaceAll("%d", "\\\\d+");
+        File cvrDirectory = new File(cvrPath);
+        File[] matchedCvrFileArray = cvrDirectory.listFiles((dir, name) -> name.matches(regexPath));
+
+        if (matchedCvrFileArray == null || matchedCvrFileArray.length == 0) {
+          String errorMessage = "Error parsing Dominion cast vote records:"
+                  + " CvrExport.json file(s) not located";
+          throw new FileNotFoundException(errorMessage);
+        }
+
+        List<File> matchedCvrFiles = Arrays.asList(matchedCvrFileArray);
+        matchedCvrFiles.sort(Comparator.comparing(File::getAbsolutePath));
+
         int recordsParsed = 0;
+        int filesParsed = 0;
         int recordsParsedAtLastLog = 0;
-        int cvrSequence = 0;
-        Path cvrFilePath = Paths.get(cvrPath, String.format(CVR_EXPORT_PATTERN, cvrSequence));
-        while (cvrFilePath.toFile().exists()) {
-          HashMap json = JsonParser.readFromFile(cvrFilePath.toString(), HashMap.class);
+
+        for (File file : matchedCvrFiles) {
+          HashMap json = JsonParser.readFromFile(file.toString(), HashMap.class);
           recordsParsed += parseCvrFile(json, castVoteRecords, contestIdToLoad);
+          filesParsed++;
+
           if (recordsParsed - recordsParsedAtLastLog > 50000) {
-            Logger.info("Parsed %d records from %d files", recordsParsed, cvrSequence);
+            Logger.info("Parsed %d records from %d files", recordsParsed, filesParsed);
             recordsParsedAtLastLog = recordsParsed;
           }
-          cvrSequence++;
-          cvrFilePath = Paths.get(cvrPath, String.format(CVR_EXPORT_PATTERN, cvrSequence));
         }
-      } else {
-        throw new FileNotFoundException(
-            String.format(
-                "Error parsing cast vote record: neither %s nor %s exists",
-                singleCvrPath, firstCvrPath));
+
+        Logger.info("Parsed %d total records from %d total files", recordsParsed, filesParsed);
       }
     } catch (FileNotFoundException | CvrParseException exception) {
       Logger.severe("Error parsing cast vote record:\n%s", exception);
diff --git a/src/test/resources/network/brightspots/rcv/test_data/dominion_multi_file/dominion_multi_file_input_data/CvrExport_0.json b/src/test/resources/network/brightspots/rcv/test_data/dominion_multi_file/dominion_multi_file_input_data/CvrExport_6.json
similarity index 100%
rename from src/test/resources/network/brightspots/rcv/test_data/dominion_multi_file/dominion_multi_file_input_data/CvrExport_0.json
rename to src/test/resources/network/brightspots/rcv/test_data/dominion_multi_file/dominion_multi_file_input_data/CvrExport_6.json