Add application file

document_processor.py

@@ -0,0 +1,41 @@
+from pathlib import Path
+import json
+import fitz
+
+
+class DocumentProcessor:
+
+    def process(self, file_path):
+
+        path = Path(file_path)
+
+        if path.suffix.lower() == ".pdf":
+            return self.process_pdf(path)
+
+        raise ValueError(
+            f"Unsupported file type: {path.suffix}"
+        )
+
+    def process_pdf(self, path):
+
+        document = fitz.open(path)
+
+        pages = []
+
+        for page_number, page in enumerate(document):
+
+            text = page.get_text()
+
+            pages.append({
+                "page_number": page_number + 1,
+                "text": text
+            })
+
+        return {
+            "document": {
+                "filename": path.name,
+                "format": "pdf",
+                "page_count": len(document)
+            },
+            "pages": pages
+        }

requirements.txt

@@ -0,0 +1,6 @@
+gradio
+docling
+pymupdf
+python-docx
+ebooklib
+beautifulsoup4