fix(AI translation): It should be able to see the input file now

Author: amilcarlucas

.github/prompts/ai-translation-user.prompt.yml

@@ -8,3 +8,6 @@ messages:
       Please read the translation file "{{translation_file}}" and translate all the strings from English to {{language}}.
 
       The file contains strings in the format "line_number:English text" - please translate only the text after the colon while preserving the exact line number and colon format.
+
+      Translation strings to process:
+      {{translation_content}}

.github/workflows/ai-translation.yml

@@ -207,7 +207,7 @@ jobs:
       matrix:
         include: ${{ fromJson(needs.extract_strings.outputs.translation-matrix) }}
       fail-fast: false  # Continue processing other languages even if one fails
-      max-parallel: 5   # Limit concurrent AI requests
+      max-parallel: 1   # Limit concurrent AI requests to avoid rate limiting
 
     steps:
       - name: Harden the runner (Audit all outbound calls)
@@ -236,6 +236,15 @@ jobs:
           echo "📊 Translation file size: $file_size bytes"
           echo "✅ Translation file validation completed successfully"
 
+      - name: Prepare translation content
+        id: translation_content
+        run: |
+          # Read file and indent each line by 2 spaces (except the first)
+          awk 'NR==1 {print} NR>1 {print "  " $0}' "${{ matrix.file }}" > indented_content.txt
+          echo "content<<EOF" >> $GITHUB_OUTPUT
+          cat indented_content.txt >> $GITHUB_OUTPUT
+          echo "EOF" >> $GITHUB_OUTPUT
+
       - name: Run AI translation
         id: ai_translate
         uses: actions/ai-inference@334892bb203895caaed82ec52d23c1ed9385151e # v2.0.4
@@ -246,10 +255,8 @@ jobs:
             language: ${{ matrix.language }}
             lang_code: ${{ matrix.lang_code }}
             translation_file: ${{ matrix.file }}
-          file_input: |
-            translation_content: ${{ matrix.file }}
-          enable-github-mcp: true
-          github-mcp-toolsets: "context,repos"
+            translation_content: |
+              ${{ steps.translation_content.outputs.content }}
           model: openai/gpt-4.1
           max-tokens: 8000
           token: ${{ secrets.AMC_COPILOT_TOKEN_CLASSIC }}
@@ -308,22 +315,32 @@ jobs:
           # Save the AI response back to the original translation file
           if [ "${{ steps.check_translation.outputs.output_method }}" = "file" ]; then
             echo "📄 Using response file: ${{ steps.ai_translate.outputs.response-file }}"
-            cp "${{ steps.ai_translate.outputs.response-file }}" "${{ matrix.file }}"
+            cp "${{ steps.ai_translate.outputs.response-file }}" "${{ matrix.file }}.raw"
           elif [ "${{ steps.check_translation.outputs.output_method }}" = "content" ]; then
             echo "📝 Using response content"
-            echo "${{ steps.ai_translate.outputs.response }}" > "${{ matrix.file }}"
+            echo "${{ steps.ai_translate.outputs.response }}" > "${{ matrix.file }}.raw"
           else
             echo "❌ Unexpected output method: ${{ steps.check_translation.outputs.output_method }}"
             exit 1
           fi
 
-          # Validate the saved file
+          # Clean up AI output: keep only lines starting with number followed by colon
+          echo "🧹 Cleaning AI output..."
+          grep -E '^[0-9]+:' "${{ matrix.file }}.raw" > "${{ matrix.file }}" || {
+            echo "❌ Failed to extract valid translations from AI output"
+            echo "Raw output preview:"
+            head -20 "${{ matrix.file }}.raw"
+            exit 1
+          }
+
+          # Validate the cleaned file
           if [ -f "${{ matrix.file }}" ] && [ -s "${{ matrix.file }}" ]; then
-            echo "✅ AI translation saved successfully for ${{ matrix.language }} (${{ matrix.file }})"
-            echo "📊 File size: $(wc -c < "${{ matrix.file }}") bytes"
-            echo "📊 Line count: $(wc -l < "${{ matrix.file }}") lines"
+            echo "✅ AI translation saved and cleaned successfully for ${{ matrix.language }}"
+            echo "📊 Raw file size: $(wc -c < "${{ matrix.file }}.raw") bytes"
+            echo "📊 Cleaned file size: $(wc -c < "${{ matrix.file }}") bytes"
+            echo "📊 Valid translation lines: $(wc -l < "${{ matrix.file }}") lines"
           else
-            echo "❌ Translation file is empty or missing: ${{ matrix.file }}"
+            echo "❌ Translation file is empty or missing after cleanup: ${{ matrix.file }}"
             exit 1
           fi
 
@@ -437,7 +454,10 @@ jobs:
 
       - name: Insert AI translations into .po files
         if: needs.extract_strings.outputs.translations-to-process == 'true'
+        shell: bash  # Don't use -e flag to prevent premature exit
         run: |
+          set -x  # Enable command tracing for debugging
+
           # Check if we have any translated files
           if ls missing_translations_*.txt 1> /dev/null 2>&1; then
             echo "📥 Processing AI translations..."
@@ -450,12 +470,12 @@ jobs:
               if [ -f "$file" ]; then
                 if grep -q "# Translation failed" "$file" 2>/dev/null; then
                   echo "⚠️ Found failed translation: $file"
-                  ((failed_translations++))
+                  failed_translations=$((failed_translations + 1))
                   # Remove failed translation files so they don't get processed
                   rm "$file"
                 else
                   echo "✅ Found successful translation: $file"
-                  ((successful_translations++))
+                  successful_translations=$((successful_translations + 1))
                 fi
               fi
             done
@@ -466,8 +486,39 @@ jobs:
 
             if [ $successful_translations -gt 0 ]; then
               echo "🔄 Processing successful translations with insert_missing_translations.py"
-              python insert_missing_translations.py
-              echo "✅ AI translations inserted into .po files"
+
+              # Show files that will be processed
+              echo "Files to process:"
+              ls -lh missing_translations_*.txt
+
+              # Validate file format before processing
+              echo "Validating translation file format..."
+              for file in missing_translations_*.txt; do
+                echo "Checking $file:"
+                if grep -qE '^[0-9]+:' "$file"; then
+                  echo "✅ File format is valid"
+                else
+                  echo "❌ ERROR: File $file does not contain valid translation lines (format: number:text)"
+                  echo "File contents:"
+                  cat "$file"
+                  exit 1
+                fi
+
+                # Show file preview
+                echo "First 5 lines of $file:"
+                head -5 "$file"
+                echo "---"
+              done
+
+              # Run with full error output captured
+              echo "Running insert_missing_translations.py..."
+              if python insert_missing_translations.py 2>&1; then
+                echo "✅ AI translations inserted into .po files"
+              else
+                exit_code=$?
+                echo "❌ insert_missing_translations.py failed with exit code $exit_code"
+                exit $exit_code
+              fi
             else
               echo "⚠️ No successful translations to process"
             fi
@@ -536,14 +587,7 @@ jobs:
 
             🤖 **AI-Powered Translation Applied with Enhanced Matrix Processing**:
             - Automatically extracted missing translations using `extract_missing_translations.py`
-            - Used GitHub Actions matrix strategy to process numbered files in parallel
             - Applied AI-powered translations using GitHub Models (GPT-4o) for multiple languages
-            - **GITHUB PROMPT.YML FORMAT**: Using official GitHub prompt.yml template format with separated files
-            - **SEPARATED PROMPT.YML FILES**: Organized in .github/prompts/ directory for better structure
-            - **CLEAN ORGANIZATION**: ai-translation-system.prompt.yml and ai-translation-user.prompt.yml separated from workflows
-            - **PERSONAL ACCESS TOKEN**: Using amilcarlucas PAT for GitHub MCP access instead of GITHUB_TOKEN
-            - **GITHUB MCP ENABLED**: AI can read translation files directly from repository using Model Context Protocol
-            - **FILE-BASED PROMPTS**: AI reads translation files directly instead of embedding content in YAML prompts
             - Supports processing unlimited translations per language with automatic chunking
             - Inserted translated strings into .po files using `insert_missing_translations.py`
             - Compiled binary .mo files for immediate use
@@ -553,23 +597,12 @@ jobs:
             **Languages processed**: Portuguese (pt), German (de), Italian (it), Japanese (ja), Chinese Simplified (zh_CN)
 
             **Enhanced Matrix Processing & Scaling**:
-            - ✅ **Parallel processing** of translation files for better performance
             - ✅ **Automatic chunking** when >50 strings per language (configurable)
             - ✅ **Robust error handling** for failed AI translation requests with detailed debugging
             - ✅ **File validation** before and after AI processing
             - ✅ **Consistent terminology** guidelines applied across all chunks for each language
             - Robust error handling for failed AI translation requests
 
-            **Technical Improvements Made**:
-            - 🔧 **Organized prompt structure**: Moved prompt files to .github/prompts/ directory to avoid confusion with workflows
-            - 🔧 **GitHub MCP enabled**: AI can read translation files directly from repository using Model Context Protocol
-            - 🔧 **Separated prompt architecture**: System prompt and user prompt in separate files for better maintainability
-            - 🔧 **File-based AI prompts**: AI reads translation files directly, eliminating YAML content embedding issues
-            - 🔧 **Reusable system prompts**: System prompt can be reused across different translation tasks
-            - 🔧 **Enhanced reliability**: No more YAML syntax issues from embedded content with special characters
-            - 🔧 **Better scalability**: File-based approach handles large translation batches without prompt size limits
-            - 🔧 **403 error fix**: Enabled GitHub MCP to resolve permission issues when reading repository files
-
             **Translation Guidelines Applied**:
             - Technical aviation/drone context preservation
             - Formal register for technical documentation