First big push

Dockerfile

@@ -0,0 +1,21 @@
+FROM jbarlow83/ocrmypdf:latest
+
+# Install packages, including tzdata for timezone support
+RUN apt-get update && apt-get install -y python3-flask python3-pip zip tzdata
+
+WORKDIR /app
+
+# Copy application code and entrypoint script
+COPY app /app
+COPY entrypoint.sh /app/entrypoint.sh
+RUN chmod +x /app/entrypoint.sh
+
+# Install gunicorn using pip with the flag to override external management
+RUN pip3 install --break-system-packages gunicorn
+
+# Expose the application port
+EXPOSE 5000
+
+# Override the base image's ENTRYPOINT with our custom script
+ENTRYPOINT ["/app/entrypoint.sh"]
+CMD []

app/app.py

@@ -0,0 +1,120 @@
+import os
+import shutil
+import subprocess
+import uuid
+from flask import Flask, request, render_template, send_file, jsonify
+from werkzeug.utils import secure_filename
+
+app = Flask(__name__)
+
+BASE_UPLOAD_FOLDER = 'uploads'
+BASE_OUTPUT_FOLDER = 'output'
+os.makedirs(BASE_UPLOAD_FOLDER, exist_ok=True)
+os.makedirs(BASE_OUTPUT_FOLDER, exist_ok=True)
+
+
+def cleanup_session_folders(session_id):
+    """Remove any existing session folders to ensure a clean slate."""
+    upload_folder = os.path.join(BASE_UPLOAD_FOLDER, session_id)
+    output_folder = os.path.join(BASE_OUTPUT_FOLDER, session_id)
+    if os.path.exists(upload_folder):
+        shutil.rmtree(upload_folder)
+    if os.path.exists(output_folder):
+        shutil.rmtree(output_folder)
+
+
+def create_session_folders(session_id):
+    """
+    Clears previous session folders if they exist and creates new ones.
+    This ensures that files from previous uploads won't mix with new ones.
+    """
+    cleanup_session_folders(session_id)
+    upload_folder = os.path.join(BASE_UPLOAD_FOLDER, session_id)
+    output_folder = os.path.join(BASE_OUTPUT_FOLDER, session_id)
+    os.makedirs(upload_folder, exist_ok=True)
+    os.makedirs(output_folder, exist_ok=True)
+    return upload_folder, output_folder
+
+
+def run_ocrmypdf(input_pdf, output_pdf):
+    cmd = [
+        'ocrmypdf',
+        '--rotate-pages',
+        '--deskew',
+        '--output-type', 'pdfa',
+        '--jobs', '4',
+        input_pdf,
+        output_pdf
+    ]
+    result = subprocess.run(cmd, capture_output=True, text=True)
+    if result.returncode != 0:
+        if "PriorOcrFoundError" in result.stderr:
+            raise ValueError("This file already contains selectable text and was skipped.")
+        else:
+            raise RuntimeError(f"OCRmyPDF failed: {result.stderr}")
+
+
+@app.route('/')
+def index():
+    # Generate a new session id for each visitor.
+    session_id = str(uuid.uuid4())
+    return render_template('index.html', session_id=session_id)
+
+
+@app.route('/upload/<session_id>', methods=['POST'])
+def upload_files(session_id):
+    # Create fresh session folders, clearing any previous data.
+    upload_folder, output_folder = create_session_folders(session_id)
+
+    if 'files' not in request.files:
+        return jsonify({'error': 'No files were uploaded.'}), 400
+
+    files = request.files.getlist('files')
+    processed_files = []
+    skipped_files = []
+
+    for file in files:
+        if file.filename == '':
+            continue
+
+        filename = secure_filename(file.filename)
+        input_path = os.path.join(upload_folder, filename)
+        output_path = os.path.join(output_folder, filename)
+
+        file.save(input_path)
+
+        try:
+            run_ocrmypdf(input_path, output_path)
+            processed_files.append(filename)
+        except ValueError:
+            skipped_files.append(filename)
+        except Exception as e:
+            # If there's an error, cleanup the session folders
+            cleanup_session_folders(session_id)
+            return jsonify({'error': f"Failed to process {filename}: {str(e)}"}), 500
+
+    if not processed_files:
+        return jsonify({
+            'error': 'All files were skipped because they already contain selectable text.',
+            'skipped_files': skipped_files
+        })
+
+    # Create a zip file containing only the current upload's processed files.
+    zip_filename = os.path.join(output_folder, 'processed_files.zip')
+    subprocess.run(['zip', '-j', zip_filename] + [os.path.join(output_folder, f) for f in processed_files])
+
+    return jsonify({
+        'download_url': f'/download/{session_id}/processed_files.zip',
+        'processed_files': processed_files,
+        'skipped_files': skipped_files
+    })
+
+
+@app.route('/download/<session_id>/<filename>')
+def download_file(session_id, filename):
+    file_path = os.path.join(BASE_OUTPUT_FOLDER, session_id, filename)
+    return send_file(file_path, as_attachment=True)
+
+
+if __name__ == '__main__':
+    app.run(host='0.0.0.0', port=5000, debug=False)

app/templates/index.html

@@ -0,0 +1,153 @@
+<!DOCTYPE html>
+<html lang="en">
+<head>
+  <meta charset="UTF-8">
+  <title>OCRmyPDF WebUI</title>
+  <meta name="viewport" content="width=device-width, initial-scale=1">
+
+  <!-- Inline minimal CSS for styling (no external dependencies) -->
+  <style>
+    /* Basic reset */
+    * {
+      box-sizing: border-box;
+      margin: 0;
+      padding: 0;
+    }
+    body {
+      font-family: Arial, sans-serif;
+      background-color: #f8f9fa;
+      padding: 20px;
+    }
+    .container {
+      max-width: 800px;
+      margin: auto;
+      background: #fff;
+      padding: 20px;
+      border-radius: 0.3em;
+      box-shadow: 0 0 10px rgba(0,0,0,0.1);
+    }
+    h2 {
+      margin-bottom: 20px;
+    }
+    input[type="file"] {
+      margin-bottom: 10px;
+    }
+    .btn {
+      display: inline-block;
+      padding: 0.5rem 1rem;
+      font-size: 1rem;
+      font-weight: bold;
+      text-align: center;
+      color: #fff;
+      background-color: #007bff;
+      border: none;
+      border-radius: 0.25rem;
+      cursor: pointer;
+      text-decoration: none;
+    }
+    .btn:disabled {
+      opacity: 0.65;
+      cursor: not-allowed;
+    }
+    .alert {
+      padding: 0.75rem 1.25rem;
+      margin: 1rem 0;
+      border: 1px solid transparent;
+      border-radius: 0.25rem;
+    }
+    .alert-info {
+      color: #055160;
+      background-color: #cff4fc;
+      border-color: #b6effb;
+    }
+    .alert-success {
+      color: #0f5132;
+      background-color: #d1e7dd;
+      border-color: #badbcc;
+    }
+    .alert-danger {
+      color: #842029;
+      background-color: #f8d7da;
+      border-color: #f5c2c7;
+    }
+    .mt-2 {
+      margin-top: 0.5rem;
+    }
+    .mt-3 {
+      margin-top: 1rem;
+    }
+    .mt-5 {
+      margin-top: 3rem;
+    }
+    .mb-4 {
+      margin-bottom: 1.5rem;
+    }
+  </style>
+  
+  <!-- Inline JavaScript -->
+  <script>
+    async function uploadFiles(sessionId) {
+      const files = document.getElementById('fileInput').files;
+      const formData = new FormData();
+      const uploadBtn = document.getElementById('uploadBtn');
+      const status = document.getElementById('status');
+      const resultLink = document.getElementById('resultLink');
+      const processedList = document.getElementById('processedList');
+      const skippedList = document.getElementById('skippedList');
+
+      resultLink.style.display = 'none';
+      processedList.innerHTML = '';
+      skippedList.innerHTML = '';
+      status.innerHTML = '';
+
+      for (let i = 0; i < files.length; i++) {
+          formData.append('files', files[i]);
+      }
+
+      uploadBtn.disabled = true;
+      uploadBtn.innerHTML = 'Processing...';
+      status.innerHTML = '<div class="alert alert-info">Processing files, please wait...</div>';
+
+      try {
+          const response = await fetch(`/upload/${sessionId}`, {
+              method: 'POST',
+              body: formData
+          });
+
+          const result = await response.json();
+
+          if (result.download_url) {
+              status.innerHTML = '<div class="alert alert-success">Processing complete! Download your files below.</div>';
+              resultLink.href = result.download_url;
+              resultLink.style.display = 'block';
+
+              if (result.processed_files && result.processed_files.length > 0) {
+                  processedList.innerHTML = '<strong>Processed Files:</strong><br>' + result.processed_files.join('<br>');
+              }
+              if (result.skipped_files && result.skipped_files.length > 0) {
+                  skippedList.innerHTML = '<strong>Skipped Files (Already Contain Text):</strong><br>' + result.skipped_files.join('<br>');
+              }
+          } else {
+              status.innerHTML = '<div class="alert alert-danger">' + (result.error || 'Unknown error occurred.') + '</div>';
+          }
+      } catch (error) {
+          status.innerHTML = '<div class="alert alert-danger">Error: ' + error.message + '</div>';
+      } finally {
+          uploadBtn.disabled = false;
+          uploadBtn.innerHTML = 'Upload and Process';
+      }
+    }
+  </script>
+</head>
+<body>
+  <div class="container mt-5">
+    <h2 class="mb-4">OCRmyPDF WebUI</h2>
+    <input type="file" id="fileInput" multiple>
+    <button id="uploadBtn" class="btn" onclick="uploadFiles('{{ session_id }}')">Upload and Process</button>
+    <div id="status" class="mt-3"></div>
+    <div id="processedList" class="mt-2"></div>
+    <div id="skippedList" class="mt-2"></div>
+    <a id="resultLink" class="btn mt-3" href="#" style="display:none;">Download Processed Files (ZIP)</a>
+  </div>
+</body>
+</html>

docker-compose.yaml

@@ -0,0 +1,11 @@
+services:
+  webui:
+    build: .
+    ports:
+      - "5000:5000"
+#    volumes:
+#      - ./uploads:/app/uploads
+#      - ./output:/app/output
+    environment:
+      - GUNICORN_TIMEOUT=300
+      - TZ=Australia/Sydney

entrypoint.sh

@@ -0,0 +1,13 @@
+#!/bin/bash
+# Set timezone if TZ environment variable is provided
+if [ -n "$TZ" ]; then
+  ln -snf /usr/share/zoneinfo/$TZ /etc/localtime
+  echo "$TZ" > /etc/timezone
+fi
+
+# Set a default Gunicorn timeout if not provided via environment variable
+: ${GUNICORN_TIMEOUT:=30}
+echo "Starting Gunicorn with timeout ${GUNICORN_TIMEOUT} seconds..."
+
+# Start Gunicorn with the specified timeout
+exec gunicorn --timeout "$GUNICORN_TIMEOUT" --workers 4 --bind 0.0.0.0:5000 app:app