ci: check duplicates, validate ZIP validity

.github/workflows/check-duplicates.yml

@@ -0,0 +1,29 @@
+name: Check Duplicates
+
+on:
+  push:
+    branches: [main]
+    paths:
+      - 'documents/**'
+      - 'bin/check-duplicates.sh'
+      - '.github/workflows/check-duplicates.yml'
+  pull_request:
+    branches: [main]
+    paths:
+      - 'documents/**'
+      - 'bin/check-duplicates.sh'
+      - '.github/workflows/check-duplicates.yml'
+
+jobs:
+  check-duplicates:
+    name: Check for Duplicate Files
+    runs-on: ubuntu-latest
+
+    steps:
+      - name: Checkout repository
+        uses: actions/checkout@v4
+        with:
+          lfs: true
+
+      - name: Check for duplicates
+        run: ./bin/check-duplicates.sh

.github/workflows/check-zip-validity.yml

@@ -0,0 +1,36 @@
+name: Check ZIP Validity
+
+on:
+  push:
+    branches: [main]
+    paths:
+      - 'documents/**/*.odt'
+      - 'documents/**/*.docx'
+      - 'documents/**/*.epub'
+      - 'bin/check-zip-validity.sh'
+      - '.github/workflows/check-zip-validity.yml'
+  pull_request:
+    branches: [main]
+    paths:
+      - 'documents/**/*.odt'
+      - 'documents/**/*.docx'
+      - 'documents/**/*.epub'
+      - 'bin/check-zip-validity.sh'
+      - '.github/workflows/check-zip-validity.yml'
+
+jobs:
+  check-zip:
+    name: Validate ZIP Structure
+    runs-on: ubuntu-latest
+
+    steps:
+      - name: Checkout repository
+        uses: actions/checkout@v4
+        with:
+          lfs: true
+
+      - name: Install unzip
+        run: sudo apt-get update && sudo apt-get install -y unzip
+
+      - name: Check ZIP validity
+        run: ./bin/check-zip-validity.sh

bin/check-duplicates.sh

@@ -0,0 +1,36 @@
+#!/bin/bash
+# Check for duplicate files in the documents directory
+# SPDX-License-Identifier: MIT
+
+set -euo pipefail
+
+echo "Checking for duplicate files in documents/..."
+echo ""
+
+TMPFILE=$(mktemp)
+trap 'rm -f "$TMPFILE"' EXIT
+
+find documents/ -type f \( -name "*.odt" -o -name "*.docx" -o -name "*.pdf" -o -name "*.epub" -o -name "*.md" \) -exec sha256sum {} \; | sort >"$TMPFILE"
+
+DUPLICATES=$(awk '{print $1}' "$TMPFILE" | uniq -d)
+
+if [ -z "$DUPLICATES" ]; then
+  echo "✅ No duplicate files found"
+  exit 0
+fi
+
+echo "❌ Duplicate files detected:"
+echo ""
+
+FOUND_DUPLICATES=0
+for CHECKSUM in $DUPLICATES; do
+  echo "Checksum: $CHECKSUM"
+  grep "^$CHECKSUM" "$TMPFILE" | awk '{print "  - " $2}'
+  echo ""
+  FOUND_DUPLICATES=1
+done
+
+if [ $FOUND_DUPLICATES -eq 1 ]; then
+  echo "❌ Found duplicate files. Please remove duplicates before committing."
+  exit 1
+fi

bin/check-zip-validity.sh

@@ -0,0 +1,34 @@
+#!/bin/bash
+# Check if ODT/EPUB/DOCX files are valid ZIP archives
+# SPDX-License-Identifier: MIT
+
+set -euo pipefail
+
+echo "Checking if ODT/EPUB/DOCX files are valid ZIP archives..."
+echo ""
+
+INVALID=0
+TOTAL=0
+
+# Find all ODT, EPUB, and DOCX files
+while IFS= read -r -d '' FILE; do
+  TOTAL=$((TOTAL + 1))
+
+  if unzip -t "$FILE" >/dev/null 2>&1; then
+    echo "✓ Valid: $FILE"
+  else
+    echo "❌ Invalid: $FILE"
+    INVALID=$((INVALID + 1))
+  fi
+done < <(find documents/ -type f \( -name "*.odt" -o -name "*.docx" -o -name "*.epub" \) -print0)
+
+echo ""
+echo "Checked $TOTAL files"
+
+if [ $INVALID -eq 0 ]; then
+  echo "✅ All files are valid ZIP archives"
+  exit 0
+else
+  echo "❌ Found $INVALID invalid ZIP file(s)"
+  exit 1
+fi