[#944] Implement authorship analysis

frontend/src/components/c-segment.vue

@@ -1,6 +1,6 @@
 <template lang="pug">
 .segment(
- v-bind:class="{ untouched: !segment.knownAuthor, active: isOpen }",
+ v-bind:class="{ untouched: !segment.knownAuthor, active: isOpen, isNotFullCredit: !segment.isFullCredit }",
  v-bind:style="{ 'border-left': `0.25rem solid ${authorColors[segment.knownAuthor]}` }",
  v-bind:title="`Author: ${segment.knownAuthor || \"Unknown\"}`"
 )

frontend/src/types/types.ts

@@ -61,6 +61,7 @@ export interface Repo extends RepoRaw {
 
 export interface AuthorshipFileSegment {
  knownAuthor: string | null;
+ isFullCredit: boolean;
  lineNumbers: number[];
  lines: string[];
 }

frontend/src/types/zod/authorship-type.ts

@@ -4,6 +4,7 @@ const lineSchema = z.object({
  lineNumber: z.number(),
  author: z.object({ gitId: z.string() }),
  content: z.string(),
+ isFullCredit: z.boolean(),
 });
 
 const fileResult = z.object({

frontend/src/utils/segment.ts

@@ -1,12 +1,20 @@
 export default class Segment {
  knownAuthor: string | null;
 
+ isFullCredit: boolean;
+
  lineNumbers: Array<number>;
 
  lines: Array<string>;
 
- constructor(knownAuthor: string | null, lineNumbers: Array<number>, lines: Array<string>) {
+ constructor(
+ knownAuthor: string | null,
+ isFullCredit: boolean,
+ lineNumbers: Array<number>,
+ lines: Array<string>,
+ ) {
  this.knownAuthor = knownAuthor;
+ this.isFullCredit = isFullCredit;
  this.lineNumbers = lineNumbers;
  this.lines = lines;
  }

frontend/src/views/c-authorship.vue

@@ -511,6 +511,7 @@ export default defineComponent({
  splitSegments(lines: Line[]): { segments: Segment[]; blankLineCount: number; } {
  // split into segments separated by knownAuthor
  let lastState: string | null;
+ let lastCreditState: boolean;
  let lastId = -1;
  const segments: Segment[] = [];
  let blankLineCount = 0;
@@ -520,16 +521,19 @@ export default defineComponent({
  ? !this.isUnknownAuthor(line.author.gitId)
  : line.author.gitId === this.info.author;
  const knownAuthor = (line.author && isAuthorMatched) ? line.author.gitId : null;
+ const isFullCredit = line.isFullCredit;
 
- if (knownAuthor !== lastState || lastId === -1) {
+ if (knownAuthor !== lastState || lastId === -1 || (knownAuthor && isFullCredit !== lastCreditState)) {
  segments.push(new Segment(
  knownAuthor,
+ isFullCredit,
  [],
  [],
  ));
 
  lastId += 1;
  lastState = knownAuthor;
+ lastCreditState = isFullCredit;
  }
 
  const content = line.content || ' ';
@@ -989,6 +993,11 @@ export default defineComponent({
  padding-left: 2rem;
  word-break: break-word;
  }
+ &.isNotFullCredit {
+ .code {
+ background-color: mui-color('green', '100');
+ }
+ }
  &.untouched {
  $grey: mui-color('grey', '400');
  border-left: .25rem solid $grey;

src/main/java/reposense/RepoSense.java

@@ -79,7 +79,8 @@ public static void main(String[] args) {
  cliArguments.getSinceDate(), cliArguments.getUntilDate(),
  cliArguments.isSinceDateProvided(), cliArguments.isUntilDateProvided(),
  cliArguments.getNumCloningThreads(), cliArguments.getNumAnalysisThreads(),
- TimeUtil::getElapsedTime, cliArguments.getZoneId(), cliArguments.isFreshClonePerformed());
+ TimeUtil::getElapsedTime, cliArguments.getZoneId(), cliArguments.isFreshClonePerformed(),
+ cliArguments.isAuthorshipAnalyzed());
 
  FileUtil.zipFoldersAndFiles(reportFoldersAndFiles, cliArguments.getOutputFilePath().toAbsolutePath(),
  ".json");

src/main/java/reposense/authorship/AuthorshipReporter.java

@@ -11,7 +11,6 @@
 import reposense.model.RepoConfiguration;
 import reposense.system.LogsManager;
 
-
 /**
  * Generates the authorship summary data for each repository.
  */
@@ -29,24 +28,24 @@ public class AuthorshipReporter {
  private final FileInfoAnalyzer fileInfoAnalyzer = new FileInfoAnalyzer();
  private final FileResultAggregator fileResultAggregator = new FileResultAggregator();
 
-
  /**
  * Generates and returns the authorship summary for each repo in {@code config}.
+ * Further analyzes the authorship of each line in the commit if {@code shouldAnalyzeAuthorship} is true.
  */
- public AuthorshipSummary generateAuthorshipSummary(RepoConfiguration config) {
+ public AuthorshipSummary generateAuthorshipSummary(RepoConfiguration config, boolean shouldAnalyzeAuthorship) {
  List<FileInfo> textFileInfos = fileInfoExtractor.extractTextFileInfos(config);
 
  int numFiles = textFileInfos.size();
  int totalNumLines = textFileInfos.stream()
- .mapToInt(fileInfo -> fileInfo.getNumOfLines())
+ .mapToInt(FileInfo::getNumOfLines)
  .sum();
 
  if (totalNumLines > HIGH_NUMBER_LINES_THRESHOLD) {
  logger.warning(String.format(HIGH_NUMBER_LINES_MESSAGE, numFiles, totalNumLines));
  }
 
  List<FileResult> fileResults = textFileInfos.stream()
- .map(fileInfo -> fileInfoAnalyzer.analyzeTextFile(config, fileInfo))
+ .map(fileInfo -> fileInfoAnalyzer.analyzeTextFile(config, fileInfo, shouldAnalyzeAuthorship))
  .filter(Objects::nonNull)
  .collect(Collectors.toList());
 

src/main/java/reposense/authorship/FileInfoAnalyzer.java

@@ -12,6 +12,7 @@
 import java.util.logging.Logger;
 
 import reposense.authorship.analyzer.AnnotatorAnalyzer;
+import reposense.authorship.analyzer.AuthorshipAnalyzer;
 import reposense.authorship.model.FileInfo;
 import reposense.authorship.model.FileResult;
 import reposense.authorship.model.LineInfo;
@@ -44,10 +45,11 @@ public class FileInfoAnalyzer {
  /**
  * Analyzes the lines of the file, given in the {@code fileInfo}, that has changed in the time period provided
  * by {@code config}.
+ * Further analyzes the authorship of each line in the commit if {@code shouldAnalyzeAuthorship} is true.
  * Returns null if the file is missing from the local system, or none of the
  * {@link Author} specified in {@code config} contributed to the file in {@code fileInfo}.
  */
- public FileResult analyzeTextFile(RepoConfiguration config, FileInfo fileInfo) {
+ public FileResult analyzeTextFile(RepoConfiguration config, FileInfo fileInfo, boolean shouldAnalyzeAuthorship) {
  String relativePath = fileInfo.getPath();
 
  if (Files.notExists(Paths.get(config.getRepoRoot(), relativePath))) {
@@ -59,7 +61,7 @@ public FileResult analyzeTextFile(RepoConfiguration config, FileInfo fileInfo) {
  return null;
  }
 
- aggregateBlameAuthorModifiedAndDateInfo(config, fileInfo);
+ aggregateBlameAuthorModifiedAndDateInfo(config, fileInfo, shouldAnalyzeAuthorship);
  fileInfo.setFileType(config.getFileType(fileInfo.getPath()));
 
  AnnotatorAnalyzer.aggregateAnnotationAuthorInfo(fileInfo, config.getAuthorConfig());
@@ -71,6 +73,19 @@ public FileResult analyzeTextFile(RepoConfiguration config, FileInfo fileInfo) {
  return generateTextFileResult(fileInfo);
  }
 
+ /**
+ * Overloading method for test cases.
+ * <br>
+ * Analyzes the lines of the file, given in the {@code fileInfo}, that has changed in the time period provided
+ * by {@code config}, without further analyzing the authorship of each line in the commit.
+ * Does not further analyze the authorship of each line in the commit.
+ * Returns null if the file is missing from the local system, or none of the
+ * {@link Author} specified in {@code config} contributed to the file in {@code fileInfo}.
+ */
+ public FileResult analyzeTextFile(RepoConfiguration config, FileInfo fileInfo) {
+ return analyzeTextFile(config, fileInfo, false);
+ }
+
  /**
  * Analyzes the binary file, given in the {@code fileInfo}, that has changed in the time period provided
  * by {@code config}.
@@ -100,9 +115,8 @@ private FileResult generateTextFileResult(FileInfo fileInfo) {
  authorContributionMap.put(author, authorContributionMap.getOrDefault(author, 0) + 1);
  }
 
- return FileResult.createTextFileResult(
- fileInfo.getPath(), fileInfo.getFileType(), fileInfo.getLines(), authorContributionMap,
- fileInfo.exceedsFileLimit());
+ return FileResult.createTextFileResult(fileInfo.getPath(), fileInfo.getFileType(), fileInfo.getLines(),
+ authorContributionMap, fileInfo.exceedsFileLimit());
  }
 
  /**
@@ -139,8 +153,10 @@ private FileResult generateBinaryFileResult(RepoConfiguration config, FileInfo f
  * The {@code config} is used to obtain the root directory for running git blame as well as other parameters used
  * in determining which author to assign to each line and whether to set the last modified date for a
  * {@code lineInfo}.
+ * Further analyzes the authorship of each line in the commit if {@code shouldAnalyzeAuthorship} is true.
  */
- private void aggregateBlameAuthorModifiedAndDateInfo(RepoConfiguration config, FileInfo fileInfo) {
+ private void aggregateBlameAuthorModifiedAndDateInfo(RepoConfiguration config, FileInfo fileInfo,
+ boolean shouldAnalyzeAuthorship) {
  String blameResults;
 
  if (!config.isFindingPreviousAuthorsPerformed()) {
@@ -159,14 +175,14 @@ private void aggregateBlameAuthorModifiedAndDateInfo(RepoConfiguration config, F
  String authorName = blameResultLines[lineCount + 1].substring(AUTHOR_NAME_OFFSET);
  String authorEmail = blameResultLines[lineCount + 2]
  .substring(AUTHOR_EMAIL_OFFSET).replaceAll("<|>", "");
- Long commitDateInMs = Long.parseLong(blameResultLines[lineCount + 3].substring(AUTHOR_TIME_OFFSET)) * 1000;
+ long commitDateInMs = Long.parseLong(blameResultLines[lineCount + 3].substring(AUTHOR_TIME_OFFSET)) * 1000;
  LocalDateTime commitDate = LocalDateTime.ofInstant(Instant.ofEpochMilli(commitDateInMs),
  config.getZoneId());
  Author author = config.getAuthor(authorName, authorEmail);
 
  if (!fileInfo.isFileLineTracked(lineCount / 5) || author.isIgnoringFile(filePath)
  || CommitHash.isInsideCommitList(commitHash, config.getIgnoreCommitList())
- || commitDate.compareTo(sinceDate) < 0 || commitDate.compareTo(untilDate) > 0) {
+ || commitDate.isBefore(sinceDate) || commitDate.isAfter(untilDate)) {
  author = Author.UNKNOWN_AUTHOR;
  }
 
@@ -179,6 +195,13 @@ private void aggregateBlameAuthorModifiedAndDateInfo(RepoConfiguration config, F
  fileInfo.setLineLastModifiedDate(lineCount / 5, commitDate);
  }
  fileInfo.setLineAuthor(lineCount / 5, author);
+
+ if (shouldAnalyzeAuthorship && !author.equals(Author.UNKNOWN_AUTHOR)) {
+ String lineContent = fileInfo.getLine(lineCount / 5 + 1).getContent();
+ boolean isFullCredit = AuthorshipAnalyzer.analyzeAuthorship(config, fileInfo.getPath(), lineContent,
+ commitHash, author);
+ fileInfo.setIsFullCredit(lineCount / 5, isFullCredit);
+ }
  }
  }