Fix Plagiarism Workflow (#1881)

DevIos01 · web-flow · commit 5b7d8a509fb3 · 2024-07-02T15:48:56.000+02:00
diff --git a/.github/scripts/extract_percentages.py b/.github/scripts/extract_percentages.py
@@ -39,6 +39,7 @@ def write_to_markdown(file_path, lines):
     with open(file_path, 'w') as md_file:
         for line in lines:
             md_file.write(line + '\n')
+    log(f"Markdown file written to {file_path}")
 
 def main():
     if len(sys.argv) != 2:
@@ -58,5 +59,6 @@ def main():
         sys.exit(1)
     else:
         log("No high plagiarism percentages detected.")
+    log("Plagiarism report generation completed.")
 if __name__ == "__main__":
     main()
diff --git a/.github/workflows/check_plagiarism.yml b/.github/workflows/check_plagiarism.yml
@@ -12,43 +12,31 @@ jobs:
     steps:
       - name: Checkout code
         uses: actions/checkout@v4
-        if: always()
         with:
           fetch-depth: 0
 
       - name: Set up Python
-        if: always()
         uses: actions/setup-python@v5
         with:
           python-version: '3.10'
 
       - name: Install Compare50 && beautifulsoup4
-        if: always()
         run: pip install compare50 beautifulsoup4
 
       - name: Get list of changed files
-        if: always()
         id: changed-files
         run: |
           echo "Pull Request Base SHA: ${{ github.event.pull_request.base.sha }}"
           echo "Pull Request Head SHA: ${{ github.event.pull_request.head.sha }}"
-          echo "Running git diff..."
-          git diff --name-only --diff-filter=AM --find-renames --find-copies ${{ github.event.pull_request.base.sha }}..${{ github.event.pull_request.head.sha }}
-          echo "Filtering JS files in games/ directory..."
+          git diff --name-only --diff-filter=AM --find-renames --find-copies ${{ github.event.pull_request.base.sha }}..${{ github.event.pull_request.head.sha }} | grep 'games/.*\.js$' | xargs
           js_files=$(git diff --name-only --diff-filter=AM --find-renames --find-copies ${{ github.event.pull_request.base.sha }}..${{ github.event.pull_request.head.sha }} | grep 'games/.*\.js$' | xargs)
-          echo "Found JS files: $js_files"
-          if [ -z "$js_files" ]; then
-            echo "No JavaScript files found in the changes."
-          else
-            echo "FILES=$js_files" >> $GITHUB_ENV
-          fi
+          echo "FILES=$js_files" >> $GITHUB_ENV
 
       - name: Run Plagiarism Detection Script
         if: env.FILES != ''
         run: python .github/scripts/plagiarism_check.py "${{ env.FILES }}" games output_dir saved_dir
 
       - name: Extract and Display Similarity Percentages
-        if: always()
         run: python .github/scripts/extract_percentages.py saved_dir/
         id: extract-percentages
 
@@ -79,4 +67,4 @@ jobs:
 
       - name: Check for High Plagiarism Percentages
         if: always() && steps.extract-percentages.outcome == 'failure'
-        run: echo "Plagiarism percentage over threshold detected."
+        run: echo "Plagiarism percentage over threshold detected."
diff --git a/.github/workflows/workflow_run.yml b/.github/workflows/workflow_run.yml
@@ -22,11 +22,10 @@ jobs:
       - name: List available artifacts with detailed logs
         uses: actions/github-script@v7
         with:
-          github-token: ${{secrets.GITHUB_TOKEN}}
+          github-token: ${{ secrets.GITHUB_TOKEN }}
           script: |
             const runId = ${{ github.event.workflow_run.id }};
             console.log(`Fetching artifacts for workflow run ID: ${runId}`);
-            
             const artifacts = await github.rest.actions.listWorkflowRunArtifacts({
               owner: context.repo.owner,
               repo: context.repo.repo,
@@ -40,68 +39,76 @@ jobs:
       - name: Download PR Number Artifact
         uses: actions/github-script@v7
         with:
-          github-token: ${{secrets.GITHUB_TOKEN}}
+          github-token: ${{ secrets.GITHUB_TOKEN }}
           script: |
             const fs = require('fs');
             const path = require('path');
             const runId = ${{ github.event.workflow_run.id }};
             const artifactName = 'pr-number';
-            
-            console.log(`Downloading artifact ${artifactName} from workflow run ID: ${runId}`);
-            const artifactData = await github.rest.actions.downloadArtifact({
-              owner: context.repo.owner,
-              repo: context.repo.repo,
-              artifact_id: (await github.rest.actions.listWorkflowRunArtifacts({
+            console.log(`Checking for artifact ${artifactName} from workflow run ID: ${runId}`);
+            try {
+              const artifacts = await github.rest.actions.listWorkflowRunArtifacts({
                 owner: context.repo.owner,
                 repo: context.repo.repo,
                 run_id: runId
-              })).data.artifacts.find(a => a.name === artifactName).id,
-              archive_format: 'zip',
-            });
-            
-            const artifactPath = path.join(process.env.GITHUB_WORKSPACE, `${artifactName}.zip`);
-            fs.writeFileSync(artifactPath, Buffer.from(artifactData.data));
-            
-            console.log(`Artifact ${artifactName} downloaded to ${artifactPath}`);
-            require('child_process').execSync(`unzip -o ${artifactPath} -d ${process.env.GITHUB_WORKSPACE}`);
-            const prNumber = fs.readFileSync(path.join(process.env.GITHUB_WORKSPACE, 'pr_number.txt'), 'utf8').trim();
-            console.log(`PR Number: ${prNumber}`);
+              });
+              const artifact = artifacts.data.artifacts.find(a => a.name === artifactName);
+              if (!artifact) {
+                console.log(`Artifact '${artifactName}' not found, skipping download.`);
+                return;
+              }
+              const artifactData = await github.rest.actions.downloadArtifact({
+                owner: context.repo.owner,
+                repo: context.repo.repo,
+                artifact_id: artifact.id,
+                archive_format: 'zip',
+              });
+              const artifactPath = path.join(process.env.GITHUB_WORKSPACE, `${artifactName}.zip`);
+              fs.writeFileSync(artifactPath, Buffer.from(artifactData.data));
+              console.log(`Artifact ${artifactName} downloaded to ${artifactPath}`);
+              require('child_process').execSync(`unzip -o ${artifactPath} -d ${process.env.GITHUB_WORKSPACE}`);
+              const prNumber = fs.readFileSync(path.join(process.env.GITHUB_WORKSPACE, 'pr_number.txt'), 'utf8').trim();
+              console.log(`PR Number: ${prNumber}`);
+            } catch (error) {
+              console.log(`Error occurred: ${error.message}`);
+              console.log('Continuing workflow execution despite the error.');
+            }
 
       - name: Download Plagiarism Report Artifact from Another Workflow Run
         uses: actions/github-script@v7
         with:
-          github-token: ${{secrets.GITHUB_TOKEN}}
+          github-token: ${{ secrets.GITHUB_TOKEN }}
           script: |
             const fs = require('fs');
             const path = require('path');
             const runId = ${{ github.event.workflow_run.id }};
             const artifactName = 'plagiarism-report';
-            
-            console.log(`Fetching artifacts for workflow run ID: ${runId}`);
-            const artifacts = await github.rest.actions.listWorkflowRunArtifacts({
-              owner: context.repo.owner,
-              repo: context.repo.repo,
-              run_id: runId,
-            });
-            
-            const artifact = artifacts.data.artifacts.find(a => a.name === artifactName);
-            if (!artifact) {
-              throw new Error(`Artifact with name ${artifactName} not found`);
+            console.log(`Downloading artifact '${artifactName}' from workflow run ID: ${runId}`);
+            try {
+              const artifacts = await github.rest.actions.listWorkflowRunArtifacts({
+                owner: context.repo.owner,
+                repo: context.repo.repo,
+                run_id: runId
+              });
+              const artifact = artifacts.data.artifacts.find(a => a.name === artifactName);
+              if (!artifact) {
+                console.log(`Artifact '${artifactName}' not found, skipping download.`);
+                return;
+              }
+              const artifactData = await github.rest.actions.downloadArtifact({
+                owner: context.repo.owner,
+                repo: context.repo.repo,
+                artifact_id: artifact.id,
+                archive_format: 'zip',
+              });
+              const artifactPath = path.join(process.env.GITHUB_WORKSPACE, `${artifactName}.zip`);
+              fs.writeFileSync(artifactPath, Buffer.from(artifactData.data));
+              console.log(`Artifact ${artifactName} downloaded to ${artifactPath}`);
+              require('child_process').execSync(`unzip -o ${artifactPath} -d ${process.env.GITHUB_WORKSPACE}`);
+            } catch (error) {
+              console.log(`Error occurred: ${error.message}`);
+              console.log('Continuing workflow execution despite the error.');
             }
-            
-            console.log(`Downloading artifact ${artifactName} from workflow run ID: ${runId}`);
-            const artifactData = await github.rest.actions.downloadArtifact({
-              owner: context.repo.owner,
-              repo: context.repo.repo,
-              artifact_id: artifact.id,
-              archive_format: 'zip',
-            });
-            
-            const artifactPath = path.join(process.env.GITHUB_WORKSPACE, `${artifactName}.zip`);
-            fs.writeFileSync(artifactPath, Buffer.from(artifactData.data));
-            
-            console.log(`Artifact ${artifactName} downloaded to ${artifactPath}`);
-            require('child_process').execSync(`unzip -o ${artifactPath} -d ${process.env.GITHUB_WORKSPACE}`);
 
       - name: Check if Plagiarism Report Exists
         id: check-report
@@ -110,6 +117,7 @@ jobs:
             echo "REPORT_EXISTS=true" >> $GITHUB_ENV
           else
             echo "REPORT_EXISTS=false" >> $GITHUB_ENV
+          fi
 
       - name: Unzip Plagiarism Report Artifact
         if: env.REPORT_EXISTS == 'true'
@@ -119,7 +127,7 @@ jobs:
         id: fetch-comments
         uses: actions/github-script@v7
         with:
-          github-token: ${{secrets.GITHUB_TOKEN}}
+          github-token: ${{ secrets.GITHUB_TOKEN }}
           script: |
             const fs = require('fs');
             const path = require('path');
@@ -137,21 +145,16 @@ jobs:
         if: env.REPORT_EXISTS == 'true' && steps.fetch-comments.outputs.result != 'true'
         uses: actions/github-script@v7
         with:
-          github-token: ${{secrets.GITHUB_TOKEN}}
+          github-token: ${{ secrets.GITHUB_TOKEN }}
           script: |
             const fs = require('fs');
             const path = require('path');
-            
             const prNumber = fs.readFileSync(path.join(process.env.GITHUB_WORKSPACE, 'pr_number.txt'), 'utf8').trim();
             const markdownPath = path.join(process.env.GITHUB_WORKSPACE, 'plagiarism-report.md');
             console.log(`Reading the Markdown report from: ${markdownPath}`);
             let markdownContent = fs.readFileSync(markdownPath, 'utf8');
-
-            markdownContent = '[Plagiarism Check Result]\n\n' + markdownContent;
-
             console.log("Fetching associated pull request...");
             console.log(`Found associated pull request: #${prNumber}`);
-
             console.log("Posting the Markdown content as a comment...");
             const commentResponse = await github.rest.issues.createComment({
               issue_number: prNumber,
