| import json |
| import tempfile |
| from pathlib import Path |
|
|
| import gradio as gr |
| import spaces |
|
|
| from document_processor import DocumentProcessor |
| from normalizer import DocumentNormalizer |
|
|
| processor = DocumentProcessor() |
| normalizer = DocumentNormalizer() |
|
|
| OUTPUT_DIR = Path(tempfile.gettempdir()) / "ai-tutor-output" |
| OUTPUT_DIR.mkdir(parents=True, exist_ok=True) |
|
|
|
|
| @spaces.GPU(duration=120) |
| def process_document(file): |
|
|
| if file is None: |
| raise gr.Error("Please upload a document.") |
|
|
| file_path = file |
|
|
| if hasattr(file, "name"): |
| file_path = file.name |
|
|
| path = Path(file_path) |
|
|
| try: |
| result = processor.process(str(path)) |
|
|
| normalized = normalizer.normalize( |
| docling_document=result["document"], |
| filename=path.name, |
| raw_json=result["raw_json"], |
| ) |
|
|
| output = normalized.model_dump() |
|
|
| base_name = path.stem |
| json_path = OUTPUT_DIR / f"{base_name}_normalized.json" |
| markdown_path = OUTPUT_DIR / f"{base_name}_markdown.md" |
| stats_path = OUTPUT_DIR / f"{base_name}_statistics.json" |
|
|
| json_path.write_text( |
| json.dumps(output, indent=2, ensure_ascii=False), |
| encoding="utf-8", |
| ) |
| markdown_path.write_text(result["markdown"], encoding="utf-8") |
| stats_path.write_text( |
| json.dumps(normalized.statistics, indent=2, ensure_ascii=False), |
| encoding="utf-8", |
| ) |
|
|
| download_files = [ |
| str(json_path), |
| str(markdown_path), |
| str(stats_path), |
| ] |
|
|
| return ( |
| json.dumps( |
| output, |
| indent=2, |
| ensure_ascii=False, |
| ), |
| result["markdown"], |
| json.dumps( |
| normalized.statistics, |
| indent=2, |
| ensure_ascii=False, |
| ), |
| download_files, |
| ) |
|
|
| except Exception as e: |
| raise gr.Error( |
| f"Document processing failed: {type(e).__name__}: {e}" |
| ) |
|
|
|
|
| with gr.Blocks( |
| title="AI Tutor Document Processor" |
| ) as demo: |
|
|
| gr.Markdown( |
| """ |
| # 📚 AI Tutor Document Processor |
| |
| Convert educational documents into a structured |
| AI-Tutor document model. |
| |
| **Supported:** PDF, DOCX, PPTX, HTML, Markdown, TXT |
| """ |
| ) |
|
|
| with gr.Row(): |
|
|
| file_input = gr.File( |
| label="Upload Document", |
| file_types=[ |
| ".pdf", |
| ".docx", |
| ".pptx", |
| ".html", |
| ".htm", |
| ".md", |
| ".txt", |
| ], |
| type="filepath", |
| ) |
|
|
| process_button = gr.Button( |
| "Process Document", |
| variant="primary", |
| ) |
|
|
| gr.Markdown("## Processing Statistics") |
|
|
| statistics_output = gr.Code( |
| label="Statistics", |
| language="json", |
| ) |
|
|
| gr.Markdown("## Download") |
|
|
| download_output = gr.File( |
| label="Download Processed Files", |
| file_count="multiple", |
| ) |
|
|
| with gr.Tabs(): |
|
|
| with gr.Tab("AI Tutor JSON"): |
|
|
| json_output = gr.Code( |
| label="Normalized Document", |
| language="json", |
| ) |
|
|
| with gr.Tab("Markdown"): |
|
|
| markdown_output = gr.Markdown() |
|
|
| process_button.click( |
| fn=process_document, |
| inputs=file_input, |
| outputs=[ |
| json_output, |
| markdown_output, |
| statistics_output, |
| download_output, |
| ], |
| api_name="process_document", |
| ) |
|
|
|
|
| if __name__ == "__main__": |
| demo.launch(show_error=True) |
|
|
