doctra 0.4.0__py3-none-any.whl → 0.4.1__py3-none-any.whl

doctra/cli/main.py

@@ -9,6 +9,7 @@ detection results, and analyze document structure from the command line.
 import click
 import os
 import sys
+import traceback
 from pathlib import Path
 from typing import Optional
 
@@ -25,6 +26,10 @@ except ImportError:
     from doctra.parsers.enhanced_pdf_parser import EnhancedPDFParser
     from doctra.parsers.table_chart_extractor import ChartTablePDFParser
 
+# Import additional modules
+from doctra.engines.layout.paddle_layout import PaddleLayoutEngine
+from doctra.engines.image_restoration import DocResEngine
+
 
 @click.group(invoke_without_command=True)
 @click.pass_context
@@ -247,7 +252,6 @@ def parse(pdf_path: Path, output_dir: Optional[Path], use_vlm: bool,
     except Exception as e:
         click.echo(f"❌ Error initializing parser: {e}", err=True)
         if verbose:
-            import traceback
             click.echo(traceback.format_exc(), err=True)
         sys.exit(1)
 
@@ -271,7 +275,6 @@ def parse(pdf_path: Path, output_dir: Optional[Path], use_vlm: bool,
     except Exception as e:
         click.echo(f"❌ Error during parsing: {e}", err=True)
         if verbose:
-            import traceback
             click.echo(traceback.format_exc(), err=True)
         sys.exit(1)
     finally:
@@ -394,7 +397,6 @@ def enhance(pdf_path: Path, output_dir: Optional[Path], restoration_task: str,
     except Exception as e:
         click.echo(f"❌ Error initializing enhanced parser: {e}", err=True)
         if verbose:
-            import traceback
             click.echo(traceback.format_exc(), err=True)
         sys.exit(1)
 
@@ -418,7 +420,6 @@ def enhance(pdf_path: Path, output_dir: Optional[Path], restoration_task: str,
     except Exception as e:
         click.echo(f"❌ Error during enhanced parsing: {e}", err=True)
         if verbose:
-            import traceback
             click.echo(traceback.format_exc(), err=True)
         sys.exit(1)
     finally:
@@ -526,7 +527,6 @@ def charts(pdf_path: Path, output_dir: Path, use_vlm: bool, vlm_provider: str,
     except Exception as e:
         click.echo(f"❌ Error during chart extraction: {e}", err=True)
         if verbose:
-            import traceback
             click.echo(traceback.format_exc(), err=True)
         sys.exit(1)
 
@@ -604,7 +604,6 @@ def tables(pdf_path: Path, output_dir: Path, use_vlm: bool, vlm_provider: str,
     except Exception as e:
         click.echo(f"❌ Error during table extraction: {e}", err=True)
         if verbose:
-            import traceback
             click.echo(traceback.format_exc(), err=True)
         sys.exit(1)
 
@@ -683,7 +682,6 @@ def both(pdf_path: Path, output_dir: Path, use_vlm: bool, vlm_provider: str,
     except Exception as e:
         click.echo(f"❌ Error during extraction: {e}", err=True)
         if verbose:
-            import traceback
             click.echo(traceback.format_exc(), err=True)
         sys.exit(1)
 
@@ -772,7 +770,6 @@ def visualize(pdf_path: Path, pages: int, columns: int, width: int,
     except Exception as e:
         click.echo(f"❌ Error creating visualization: {e}", err=True)
         if verbose:
-            import traceback
             click.echo(traceback.format_exc(), err=True)
         sys.exit(1)
 
@@ -805,7 +802,6 @@ def analyze(pdf_path: Path, dpi: int, min_score: float, layout_model: str, verbo
         click.echo(f"🔍 Analyzing: {pdf_path.name}")
 
         # Create layout engine for analysis only
-        from doctra.engines.layout.paddle_layout import PaddleLayoutEngine
 
         if verbose:
             click.echo(f"   Using model: {layout_model}")
@@ -903,7 +899,6 @@ def analyze(pdf_path: Path, dpi: int, min_score: float, layout_model: str, verbo
     except Exception as e:
         click.echo(f"❌ Error analyzing PDF: {e}", err=True)
         if verbose:
-            import traceback
             click.echo(traceback.format_exc(), err=True)
         sys.exit(1)
 
@@ -922,7 +917,6 @@ def info():
     click.echo("=" * 50)
 
     # Check Python version
-    import sys
     python_version = f"{sys.version_info.major}.{sys.version_info.minor}.{sys.version_info.micro}"
     click.echo(f"Python version: {python_version}")
 
@@ -1003,7 +997,6 @@ def info():
     # DocRes information
     click.echo("\nDocRes Image Restoration:")
     try:
-        from doctra.engines.image_restoration import DocResEngine
         docres = DocResEngine()
         click.echo(f"  ✅ DocRes available - {len(docres.get_supported_tasks())} restoration tasks")
         click.echo("  Tasks: dewarping, deshadowing, appearance, deblurring, binarization, end2end")

doctra/cli/utils.py

@@ -7,8 +7,10 @@ different CLI commands.
 
 import click
 import sys
+import traceback
 from typing import Optional, Dict, Any
 from pathlib import Path
+from doctra.utils.progress import create_beautiful_progress_bar, create_notebook_friendly_bar
 
 
 def validate_vlm_config(use_vlm: bool, vlm_api_key: Optional[str]) -> None:
@@ -58,7 +60,6 @@ def handle_exception(e: Exception, verbose: bool = False) -> None:
     """
     click.echo(f"❌ Error: {e}", err=True)
     if verbose:
-        import traceback
         click.echo(traceback.format_exc(), err=True)
     sys.exit(1)
 
@@ -271,8 +272,6 @@ def create_progress_callback(description: str, total: int):
     :return: Callable progress callback function that takes an integer
              representing the number of completed items
     """
-    import sys
-    from doctra.utils.progress import create_beautiful_progress_bar, create_notebook_friendly_bar
 
     # Enhanced environment detection
     is_notebook = "ipykernel" in sys.modules or "jupyter" in sys.modules

doctra/engines/image_restoration/docres_engine.py

@@ -18,6 +18,8 @@ import sys
 import cv2
 import numpy as np
 import torch
+import tempfile
+import time
 from pathlib import Path
 from typing import Union, List, Tuple, Optional, Dict, Any
 
@@ -308,8 +310,6 @@ class DocResEngine:
     
     def _run_single_task(self, img_array: np.ndarray, task: str, save_prompts: bool) -> Tuple[np.ndarray, Dict]:
         """Run a single restoration task"""
-        import tempfile
-        import time
         
         # Create temporary file for inference
         with tempfile.NamedTemporaryFile(suffix='.jpg', delete=False) as tmp_file:
@@ -322,7 +322,6 @@ class DocResEngine:
             os.chdir(str(docres_dir))
             
             # Set global DEVICE variable that DocRes inference expects
-            import torch
             import inference  # Import the inference module to set its global DEVICE
             inference.DEVICE = self.device
             
@@ -364,8 +363,6 @@ class DocResEngine:
     
     def _run_end2end_pipeline(self, img_array: np.ndarray, save_prompts: bool) -> Tuple[np.ndarray, Dict]:
         """Run the end2end pipeline: dewarping → deshadowing → appearance"""
-        import tempfile
-        import time
         
         intermediate_steps = {}
         
@@ -374,7 +371,6 @@ class DocResEngine:
         os.chdir(str(docres_dir))
         
         # Set global DEVICE variable that DocRes inference expects
-        import torch
         import inference  # Import the inference module to set its global DEVICE
         inference.DEVICE = self.device
         
@@ -482,7 +478,6 @@ class DocResEngine:
         """
         try:
             from PIL import Image
-            import numpy as np
             from doctra.utils.pdf_io import render_pdf_to_images
             
             # Generate output path if not provided

doctra/engines/vlm/outlines_types.py

@@ -1,17 +1,19 @@
-from pydantic import BaseModel
+from pydantic import BaseModel, Field
 
 class Chart(BaseModel):
     """
     Structured representation of a chart extracted from an image.
     
-    Contains the title, headers, and data rows extracted from a chart
-    using VLM (Vision Language Model) processing.
+    Includes a title, a short description, column headers, and data rows
+    identified using VLM (Vision Language Model) processing.
 
-    :param title: Title or caption of the chart
+    :param title: Title or caption of the chart (max 31 characters)
+    :param description: Short description of the chart (max 300 characters)
     :param headers: Column headers for the chart data
     :param rows: Data rows containing the chart values
     """
-    title: str
+    title: str = Field(max_length=31)
+    description: str = Field(max_length=300)
     headers: list[str]
     rows: list[list[str]]
 
@@ -19,13 +21,15 @@ class Table(BaseModel):
     """
     Structured representation of a table extracted from an image.
     
-    Contains the title, headers, and data rows extracted from a table
-    using VLM (Vision Language Model) processing.
+    Includes a title, a short description, column headers, and data rows
+    identified using VLM (Vision Language Model) processing.
 
-    :param title: Title or caption of the table
+    :param title: Title or caption of the table (max 31 characters)
+    :param description: Short description of the table (max 300 characters)
     :param headers: Column headers for the table data
     :param rows: Data rows containing the table values
     """
-    title: str
+    title: str = Field(max_length=31)
+    description: str = Field(max_length=300)
     headers: list[str]
     rows: list[list[str]]

doctra/engines/vlm/service.py

@@ -73,7 +73,7 @@ class VLMStructuredExtractor:
         Extract structured chart data from an image.
 
         :param image_path: Path to the chart image file
-        :return: Chart object containing extracted title, headers, and data rows
+        :return: Chart object containing extracted title, description, headers, and data rows
         :raises Exception: If image processing or VLM extraction fails
         """
         prompt_text = (
@@ -81,6 +81,7 @@ class VLMStructuredExtractor:
             "If the title is not present in the image, generate a suitable title. "
             "Ensure that the table represents the data from the chart accurately."
             "The number of columns in the headers must match the number of columns in each row."
+            "Also provide a short description (max 300 characters) of the chart."
         )
         return self._call(prompt_text, image_path, Chart)
 
@@ -89,7 +90,7 @@ class VLMStructuredExtractor:
         Extract structured table data from an image.
 
         :param image_path: Path to the table image file
-        :return: Table object containing extracted title, headers, and data rows
+        :return: Table object containing extracted title, description, headers, and data rows
         :raises Exception: If image processing or VLM extraction fails
         """
         prompt_text = (
@@ -97,5 +98,6 @@ class VLMStructuredExtractor:
             "Provide the headers and rows of the table, ensuring accuracy in the extraction. "
             "If the title is not present in the image, generate a suitable title."
             "The number of columns in the headers must match the number of columns in each row."
+            "Also provide a short description (max 300 characters) of the table."
         )
         return self._call(prompt_text, image_path, Table)

doctra/exporters/excel_writer.py

@@ -5,6 +5,7 @@ from typing import Dict, Any, List, Set
 import pandas as pd  # pip install pandas openpyxl
 from openpyxl.styles import PatternFill, Font, Alignment
 from openpyxl.utils import get_column_letter
+from openpyxl.worksheet.hyperlink import Hyperlink
 
 _INVALID_SHEET_CHARS = r'[:\\/*?\[\]]'  # Excel-invalid characters
 _MAX_SHEET_LEN = 31
@@ -85,6 +86,61 @@ def _autosize_columns(ws, df: pd.DataFrame) -> None:
         ws.column_dimensions[get_column_letter(i)].width = min(max(10, max_len + 2), 60)
 
 
+def _style_summary_sheet(ws, df: pd.DataFrame, sheet_mapping: dict = None) -> None:
+    """
+    Apply special styling to the summary sheet with text wrapping for descriptions.
+    Add hyperlinks to table titles that link to their corresponding sheets.
+    
+    :param ws: OpenPyXL worksheet object to style
+    :param df: Pandas DataFrame containing the summary data
+    :param sheet_mapping: Dictionary mapping table titles to their sheet names
+    :return: None
+    """
+    # Style header row
+    _style_header(ws, ncols=df.shape[1])
+    
+    # Apply text wrapping to all data cells
+    wrap_alignment = Alignment(wrap_text=True, vertical="top")
+    
+    # Apply wrapping to all data rows (skip header row)
+    for row_idx in range(2, len(df) + 2):  # Start from row 2 (after header)
+        for col_idx in range(1, df.shape[1] + 1):
+            cell = ws.cell(row=row_idx, column=col_idx)
+            cell.alignment = wrap_alignment
+            
+            # Add hyperlink to table title column (column A)
+            if col_idx == 1 and sheet_mapping:  # Table Title column
+                table_title = cell.value
+                if table_title and table_title in sheet_mapping:
+                    sheet_name = sheet_mapping[table_title]
+                    
+                    # Create hyperlink to the sheet using proper Excel format
+                    # Escape sheet name if it contains spaces or special characters
+                    if ' ' in sheet_name or any(char in sheet_name for char in ['[', ']', '*', '?', ':', '\\', '/']):
+                        hyperlink_ref = f"#'{sheet_name}'!A1"
+                    else:
+                        hyperlink_ref = f"#{sheet_name}!A1"
+                    
+                    # Use Hyperlink class with proper parameters
+                    cell.hyperlink = Hyperlink(ref=hyperlink_ref, target=hyperlink_ref)
+                    # Style the hyperlink
+                    cell.font = Font(color="0000FF", underline="single")
+    
+    # Set specific column widths for summary sheet
+    # Table Title column - narrower
+    ws.column_dimensions['A'].width = 30
+    # Description column - wider to accommodate wrapped text
+    ws.column_dimensions['B'].width = 60
+    # Page column - narrow for page numbers
+    ws.column_dimensions['C'].width = 10
+    # Type column - narrow for Table/Chart
+    ws.column_dimensions['D'].width = 12
+    
+    # Set row heights to accommodate wrapped text
+    for row_idx in range(2, len(df) + 2):
+        ws.row_dimensions[row_idx].height = 60  # Allow for multiple lines
+
+
 def _normalize_data(headers: List[str], rows: List[List]) -> tuple[List[str], List[List]]:
     """
     Normalize headers and rows to ensure consistent dimensions.
@@ -159,6 +215,31 @@ def write_structured_excel(excel_path: str, items: List[Dict[str, Any]]) -> str
     taken: Set[str] = set()
 
     with pd.ExcelWriter(excel_path, engine="openpyxl", mode="w") as writer:
+        # Create summary sheet first
+        summary_data = []
+        sheet_mapping = {}  # Map table titles to their sheet names
+        
+        for item in valid_items:
+            title = item.get("title") or "Untitled"
+            description = item.get("description") or "No description available"
+            page_number = item.get("page", "Unknown")
+            item_type = item.get("type", "Table")  # Default to "Table" if not specified
+            
+            
+            summary_data.append({
+                "Table Title": title,
+                "Description": description,
+                "Page": page_number,
+                "Type": item_type
+            })
+        
+        # Create summary sheet first (but without hyperlinks initially)
+        if summary_data:
+            summary_df = pd.DataFrame(summary_data)
+            summary_df.to_excel(writer, sheet_name="Table Summary", index=False)
+            taken.add("Table Summary")
+
+        # Process individual table sheets to build sheet mapping
         for item in valid_items:
             try:
                 title = item.get("title") or "Untitled"
@@ -166,6 +247,9 @@ def write_structured_excel(excel_path: str, items: List[Dict[str, Any]]) -> str
                 rows = item.get("rows") or []
 
                 sheet_name = _safe_sheet_name(title, taken)
+                
+                # Add to sheet mapping for hyperlinks
+                sheet_mapping[title] = sheet_name
 
                 # Normalize data to handle mismatched dimensions
                 normalized_headers, normalized_rows = _normalize_data(headers, rows)
@@ -194,4 +278,9 @@ def write_structured_excel(excel_path: str, items: List[Dict[str, Any]]) -> str
                 print(f"Error processing item '{item.get('title', 'Unknown')}': {e}")
                 continue
 
+        # Now add hyperlinks to the summary sheet (after all sheets are created)
+        if summary_data and sheet_mapping:
+            summary_ws = writer.sheets["Table Summary"]
+            _style_summary_sheet(summary_ws, summary_df, sheet_mapping)
+
     return excel_path

doctra/parsers/enhanced_pdf_parser.py

@@ -8,6 +8,7 @@ capabilities with DocRes image restoration for improved document processing.
 from __future__ import annotations
 import os
 import sys
+import numpy as np
 from typing import List, Dict, Any, Optional, Union
 from contextlib import ExitStack
 from PIL import Image
@@ -16,9 +17,17 @@ from tqdm import tqdm
 from doctra.parsers.structured_pdf_parser import StructuredPDFParser
 from doctra.engines.image_restoration import DocResEngine
 from doctra.utils.pdf_io import render_pdf_to_images
-from doctra.utils.constants import IMAGE_SUBDIRS
+from doctra.utils.constants import IMAGE_SUBDIRS, EXCLUDE_LABELS
 from doctra.utils.file_ops import ensure_output_dirs
 from doctra.utils.progress import create_beautiful_progress_bar, create_notebook_friendly_bar
+from doctra.parsers.layout_order import reading_order_key
+from doctra.utils.ocr_utils import ocr_box_text
+from doctra.exporters.image_saver import save_box_image
+from doctra.exporters.markdown_writer import write_markdown
+from doctra.exporters.html_writer import write_html, write_structured_html
+from doctra.exporters.excel_writer import write_structured_excel
+from doctra.utils.structured_utils import to_structured_dict
+from doctra.exporters.markdown_table import render_markdown_table
 
 
 class EnhancedPDFParser(StructuredPDFParser):
@@ -146,7 +155,7 @@ class EnhancedPDFParser(StructuredPDFParser):
         pil_pages = enhanced_pages
         
         # Continue with standard parsing logic
-        self._process_parsing_logic(pages, pil_pages, out_dir, pdf_filename)
+        self._process_parsing_logic(pages, pil_pages, out_dir, pdf_filename, pdf_path)
 
     def _process_pages_with_restoration(self, pdf_path: str, out_dir: str) -> List[Image.Image]:
         """
@@ -186,7 +195,6 @@ class EnhancedPDFParser(StructuredPDFParser):
                 for i, page_img in enumerate(original_pages):
                     try:
                         # Convert PIL to numpy array
-                        import numpy as np
                         img_array = np.array(page_img)
                         
                         # Apply DocRes restoration
@@ -219,21 +227,11 @@ class EnhancedPDFParser(StructuredPDFParser):
         print(f"✅ Image restoration completed. Enhanced pages saved to: {enhanced_dir}")
         return enhanced_pages
 
-    def _process_parsing_logic(self, pages, pil_pages, out_dir, pdf_filename):
+    def _process_parsing_logic(self, pages, pil_pages, out_dir, pdf_filename, pdf_path):
         """
         Process the parsing logic with enhanced pages.
         This is extracted from the parent class to allow customization.
         """
-        from doctra.utils.constants import EXCLUDE_LABELS
-        from doctra.parsers.layout_order import reading_order_key
-        from doctra.utils.ocr_utils import ocr_box_text
-        from doctra.exporters.image_saver import save_box_image
-        from doctra.exporters.markdown_writer import write_markdown
-        from doctra.exporters.html_writer import write_html
-        from doctra.exporters.excel_writer import write_structured_excel
-        from doctra.exporters.html_writer import write_structured_html
-        from doctra.utils.structured_utils import to_structured_dict
-        from doctra.exporters.markdown_table import render_markdown_table
         
         fig_count = sum(sum(1 for b in p.boxes if b.label == "figure") for p in pages)
         chart_count = sum(sum(1 for b in p.boxes if b.label == "chart") for p in pages)
@@ -285,6 +283,9 @@ class EnhancedPDFParser(StructuredPDFParser):
                                     chart = self.vlm.extract_chart(abs_img_path)
                                     item = to_structured_dict(chart)
                                     if item:
+                                        # Add page and type information to structured item
+                                        item["page"] = page_num
+                                        item["type"] = "Chart"
                                         structured_items.append(item)
                                         md_lines.append(
                                             render_markdown_table(item.get("headers"), item.get("rows"),
@@ -306,6 +307,9 @@ class EnhancedPDFParser(StructuredPDFParser):
                                     table = self.vlm.extract_table(abs_img_path)
                                     item = to_structured_dict(table)
                                     if item:
+                                        # Add page and type information to structured item
+                                        item["page"] = page_num
+                                        item["type"] = "Table"
                                         structured_items.append(item)
                                         md_lines.append(
                                             render_markdown_table(item.get("headers"), item.get("rows"),

doctra/parsers/structured_pdf_parser.py

@@ -163,6 +163,9 @@ class StructuredPDFParser:
                                     chart = self.vlm.extract_chart(abs_img_path)
                                     item = to_structured_dict(chart)
                                     if item:
+                                        # Add page and type information to structured item
+                                        item["page"] = page_num
+                                        item["type"] = "Chart"
                                         structured_items.append(item)
                                         md_lines.append(
                                             render_markdown_table(item.get("headers"), item.get("rows"),
@@ -184,6 +187,9 @@ class StructuredPDFParser:
                                     table = self.vlm.extract_table(abs_img_path)
                                     item = to_structured_dict(table)
                                     if item:
+                                        # Add page and type information to structured item
+                                        item["page"] = page_num
+                                        item["type"] = "Table"
                                         structured_items.append(item)
                                         md_lines.append(
                                             render_markdown_table(item.get("headers"), item.get("rows"),

doctra/parsers/table_chart_extractor.py

@@ -178,6 +178,9 @@ class ChartTablePDFParser:
                                 extracted_chart = self.vlm.extract_chart(chart_path)
                                 structured_item = to_structured_dict(extracted_chart)
                                 if structured_item:
+                                    # Add page and type information to structured item
+                                    structured_item["page"] = page_num
+                                    structured_item["type"] = "Chart"
                                     structured_items.append(structured_item)
                                     vlm_items.append({
                                         "kind": "chart",
@@ -221,6 +224,9 @@ class ChartTablePDFParser:
                                 extracted_table = self.vlm.extract_table(table_path)
                                 structured_item = to_structured_dict(extracted_table)
                                 if structured_item:
+                                    # Add page and type information to structured item
+                                    structured_item["page"] = page_num
+                                    structured_item["type"] = "Table"
                                     structured_items.append(structured_item)
                                     vlm_items.append({
                                         "kind": "table",

doctra/ui/app.py

@@ -2,6 +2,11 @@ import os
 import shutil
 import tempfile
 import re
+import traceback
+import pandas as pd
+import html as _html
+import base64
+import json
 from pathlib import Path
 from typing import Optional, Tuple, List, Dict, Any
 
@@ -9,6 +14,7 @@ import gradio as gr
 
 from doctra.parsers.structured_pdf_parser import StructuredPDFParser
 from doctra.parsers.table_chart_extractor import ChartTablePDFParser
+from doctra.utils.pdf_io import render_pdf_to_images
 
 
 def _gather_outputs(out_dir: Path, allowed_kinds: Optional[List[str]] = None, zip_filename: Optional[str] = None, is_structured_parsing: bool = False) -> Tuple[List[tuple[str, str]], List[str], str]:
@@ -100,7 +106,6 @@ def _parse_markdown_by_pages(md_content: str) -> List[Dict[str, Any]]:
     Parse markdown content and organize it by pages.
     Returns a list of page dictionaries with content, tables, charts, and figures.
     """
-    import re
     
     pages = []
     current_page = None
@@ -209,7 +214,6 @@ def run_full_parse(
     try:
         parser.parse(str(input_pdf))
     except Exception as e:
-        import traceback
         traceback.print_exc()
         # Safely encode error message for return value
         try:
@@ -325,8 +329,6 @@ def run_extract(
             if excel_filename:
                 excel_path = out_dir / excel_filename
                 if excel_path.exists():
-                    import pandas as pd
-                    import html as _html
                     
                     # Read Excel file and create HTML tables
                     xl_file = pd.ExcelFile(excel_path)
@@ -489,7 +491,6 @@ def build_demo() -> gr.Blocks:
 
             def parse_markdown_by_pages(md_content: str):
                 """Parse markdown content and organize it by pages."""
-                import re
                 
                 pages = []
                 current_page = None
@@ -548,7 +549,6 @@ def build_demo() -> gr.Blocks:
                     return "Page not found", None
                 
                 # Build HTML with inline base64 images, render markdown tables, and preserve paragraphs/line breaks
-                import html as _html, base64, re as _re
                 base_dir = None
                 try:
                     stem = Path(pdf_path).stem if pdf_path else ""
@@ -589,7 +589,7 @@ def build_demo() -> gr.Blocks:
                     stripped = line.strip()
                     if stripped.startswith('![') and ('](images/' in stripped or '](images\\' in stripped):
                         flush_paragraph()
-                        match = _re.match(r'!\[([^\]]+)\]\(([^)]+)\)', stripped)
+                        match = re.match(r'!\[([^\]]+)\]\(([^)]+)\)', stripped)
                         if match and base_dir is not None:
                             caption = match.group(1)
                             rel_path = match.group(2).replace('\\\\', '/').replace('\\', '/').lstrip('/')
@@ -646,7 +646,6 @@ def build_demo() -> gr.Blocks:
                 # Ensure page images are prepared
                 try:
                     if pdf_path and not page_images:
-                        from doctra.utils.pdf_io import render_pdf_to_images
                         tmp_img_dir = Path(tempfile.mkdtemp(prefix="doctra_pages_"))
                         pil_pages = render_pdf_to_images(pdf_path)
                         saved_paths: List[str] = []
@@ -726,7 +725,6 @@ def build_demo() -> gr.Blocks:
                     for page in pages_data:
                         for line in page['content']:
                             if line.strip().startswith('![') and ('](images/' in line or '](images\\' in line):
-                                import re
                                 match = re.match(r'!\[([^\]]+)\]\(([^)]+)\)', line.strip())
                                 if match:
                                     caption = match.group(1)
@@ -745,7 +743,6 @@ def build_demo() -> gr.Blocks:
                 saved_paths: List[str] = []
                 try:
                     if input_pdf_path:
-                        from doctra.utils.pdf_io import render_pdf_to_images
                         tmp_img_dir = Path(tempfile.mkdtemp(prefix="doctra_pages_"))
                         pil_pages = render_pdf_to_images(input_pdf_path)
                         for idx, (im, _, _) in enumerate(pil_pages, start=1):
@@ -759,7 +756,6 @@ def build_demo() -> gr.Blocks:
 
                 # Build initial HTML with inline images and proper blocks for first page
                 if pages_data:
-                    import html as _html, base64, re as _re
                     base_dir = None
                     try:
                         stem = Path(input_pdf_path).stem if input_pdf_path else ""
@@ -771,7 +767,7 @@ def build_demo() -> gr.Blocks:
                     for raw_line in pages_data[0]['content']:
                         line = raw_line.strip()
                         if line.startswith('![') and ('](images/' in line or '](images\\' in line):
-                            match = _re.match(r'!\[([^\]]+)\]\(([^)]+)\)', line)
+                            match = re.match(r'!\[([^\]]+)\]\(([^)]+)\)', line)
                             if match and base_dir is not None:
                                 caption = match.group(1)
                                 rel_path = match.group(2).replace('\\\\', '/').replace('\\', '/').lstrip('/')
@@ -874,7 +870,6 @@ def build_demo() -> gr.Blocks:
                     if not mapping.exists():
                         return gr.Dropdown(choices=[], value=None, visible=False)
                     
-                    import json
                     data = json.loads(mapping.read_text(encoding="utf-8"))
                     choices = []
                     
@@ -902,7 +897,6 @@ def build_demo() -> gr.Blocks:
                     if not mapping.exists():
                         return "", None
                     
-                    import json, html as _html
                     data = json.loads(mapping.read_text(encoding="utf-8"))
                     
                     for entry in data:

doctra/utils/structured_utils.py

@@ -14,7 +14,7 @@ def to_structured_dict(obj: Any) -> Optional[Dict[str, Any]]:
       - JSON string
       - dict
       - Pydantic BaseModel (v1 .dict() or v2 .model_dump())
-    Returns a normalized dict with keys: title, headers, rows — or None.
+    Returns a normalized dict with keys: title, description, headers, rows, page, type — or None.
     """
     if obj is None:
         return None
@@ -36,10 +36,13 @@ def to_structured_dict(obj: Any) -> Optional[Dict[str, Any]]:
 
     if isinstance(obj, dict):
         title = obj.get("title") or "Untitled"
+        description = obj.get("description") or ""
         headers = obj.get("headers") or []
         rows = obj.get("rows") or []
+        page = obj.get("page", "Unknown")
+        item_type = obj.get("type", "Table")
         if not isinstance(headers, list) or not isinstance(rows, list):
             return None
-        return {"title": title, "headers": headers, "rows": rows}
+        return {"title": title, "description": description, "headers": headers, "rows": rows, "page": page, "type": item_type}
 
     return None

doctra/version.py

@@ -1,2 +1,2 @@
 """Version information for Doctra."""
-__version__ = '0.4.0'
+__version__ = '0.4.1'

{doctra-0.4.0.dist-info → doctra-0.4.1.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: doctra
-Version: 0.4.0
+Version: 0.4.1
 Summary: Parse, extract, and analyze documents with ease
 Home-page: https://github.com/AdemBoukhris457/Doctra
 Author: Adem Boukhris

{doctra-0.4.0.dist-info → doctra-0.4.1.dist-info}/RECORD

@@ -1,11 +1,11 @@
 doctra/__init__.py,sha256=rNLCyODOpaPb_TTP6qmQnuWZJW9JPXrxg1IfKnvb1No,773
-doctra/version.py,sha256=PSDo-SLZhu8_cWgmtvzLjHyKr7C8D_F61M1tiywnuKY,62
+doctra/version.py,sha256=gJX4jQdS3czcKE2h1k17fJPgWzxHyGH2oFP9nW9cTLw,62
 doctra/cli/__init__.py,sha256=4PTujjYRShOOUlZ7PwuWckShPWLC4v4CYIhJpzgyv1k,911
-doctra/cli/main.py,sha256=6b415qg-8gV4M2Uf0WvdU_nFx65DYFgRu5Q3Ys_LvAo,43756
-doctra/cli/utils.py,sha256=IghiUZQCOmXODC5-5smHGz2KeV4xqbP4avmA1Mggln0,11800
+doctra/cli/main.py,sha256=_gvG8bm-Mn1tIEw6eJUgqz9dYEo9klXGiJDJzjqgPyo,43503
+doctra/cli/utils.py,sha256=w3Bxyzczcbl_cs1Cea8C3ehv7dkGl_wecprYZXrcGhk,11772
 doctra/engines/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 doctra/engines/image_restoration/__init__.py,sha256=vzcN6Rw7_U-5jIK2pdo2NlgqdLdXDShigrOGM7QLNEE,263
-doctra/engines/image_restoration/docres_engine.py,sha256=6j2LfoqirmEEmLTOsz8nkhqaHUQHjYbJr-2MR01i6Gc,21754
+doctra/engines/image_restoration/docres_engine.py,sha256=n9Pr0R7dbu_UHv51urGv_wC6ZYW-43bmXxiyTCOEOMo,21612
 doctra/engines/layout/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 doctra/engines/layout/layout_models.py,sha256=vuTzjWd3FD-SkFPngktmUVhOJ6Xvff6ufwFEq796PQs,3162
 doctra/engines/layout/paddle_layout.py,sha256=P2-Gk8wHpWoA5Jpmo_3OLI59zWq3HeAOBOUKKVdXu8I,6792
@@ -14,20 +14,20 @@ doctra/engines/ocr/api.py,sha256=YOBKDLExXpvSiOsc_TDJasaMPxzdVx1llQCtYlsruWo,128
 doctra/engines/ocr/path_resolver.py,sha256=2_7Nsekt3dCDU3oVsgdr62iMrlAhbGNfYwgh4G7S3pA,1492
 doctra/engines/ocr/pytesseract_engine.py,sha256=Imz2uwju6himkBiS8CH7DLxBRe-LtmMYZiOdb_6PoQw,2911
 doctra/engines/vlm/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-doctra/engines/vlm/outlines_types.py,sha256=qL-G6MNiA5mxp1qAPVEFhOANp4NqVt_MQKseJCr_xXE,970
+doctra/engines/vlm/outlines_types.py,sha256=fQK6ru7XiXHaa8JPpaTTBaTk_zQ93ZyhFp4SyAnUdVU,1337
 doctra/engines/vlm/provider.py,sha256=aE8Eo1U-8XqAimakNlT0-T4etIyCV8rZ3DwxdqbFeTc,3131
-doctra/engines/vlm/service.py,sha256=4ExDbLmyyC3ICXxr7OSIqvbOdrwbIJek-DE54vAUgDA,4151
+doctra/engines/vlm/service.py,sha256=nygxMe7uTq6Bv70ycBPL59F2a0ESp1Hix4j833p6rUM,4343
 doctra/exporters/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-doctra/exporters/excel_writer.py,sha256=U5Eb5SF7_ll1QveUapSWSkCRt3OEoisKEVUQ_7X8Wjo,7762
+doctra/exporters/excel_writer.py,sha256=rwyqlH73P7z413BELovQY_pS6IMkkqHEho6mbPrJ2Sk,11857
 doctra/exporters/html_writer.py,sha256=OlW24Eg5bZcjldRHtd3GDD7RrajuRXj43EJpXIJkYf8,38810
 doctra/exporters/image_saver.py,sha256=zsPoQ0CwoE643ui4iZMdXk96kv5mU8L_zC2JfF22N1A,1639
 doctra/exporters/markdown_table.py,sha256=4_OJIwG_WoIPYBzJx1njy_3tNVdkK6QKSP-P9r-b0zw,2030
 doctra/exporters/markdown_writer.py,sha256=L7EjF2MB8jYX7XkZ3a3NeeEC8gnb0qzRPTzIN9tdfuw,1027
 doctra/parsers/__init__.py,sha256=8M6LVzcWGpuTIK_1SMXML3ll7zK1CTHXGI5qXvqdm-A,206
-doctra/parsers/enhanced_pdf_parser.py,sha256=7KfkQexXTxbi8Naen7HFlFaeoEGpfdbYbvRqkTXw22A,18095
+doctra/parsers/enhanced_pdf_parser.py,sha256=NBBopYdSIHWd_O96J0qR3DqZvbAt3CfK1hwUkXu8540,18377
 doctra/parsers/layout_order.py,sha256=W6b-T11H907RZ2FaZwNvnYhmvH11rpUzxC5yLkdf28k,640
-doctra/parsers/structured_pdf_parser.py,sha256=QIZIS5SAaIdGiT8o7G_a4D-Cht7nVLGeSuVzqSYLn14,19160
-doctra/parsers/table_chart_extractor.py,sha256=kSubqX0n0kVu_3jzX6QUyKmEGs9sG3Bg9kzUzn2wPHo,13733
+doctra/parsers/structured_pdf_parser.py,sha256=AU6yLW2kpd8bxZjelmm73L4CVBysnVAdKxwPkTV1Fzk,19602
+doctra/parsers/table_chart_extractor.py,sha256=ePmk9m9n-mvkqOvxpWC42ElxbnKMmDnq-e6SWiNqgzA,14195
 doctra/third_party/docres/inference.py,sha256=krD5EQDiqki-5uTMqqHYivhL38sfSOhYgaihI751070,13576
 doctra/third_party/docres/utils.py,sha256=N0ZVmOTB3wsinFlYu5hT84C4_MhWGdc98T8LTG-S9dA,14566
 doctra/third_party/docres/data/MBD/MBD.py,sha256=-d6cVQX1FVcGmQ_yJ5l-PQ3xKmkmveQQBytZ38pEGfY,4653
@@ -49,7 +49,7 @@ doctra/third_party/docres/data/MBD/model/deep_lab_model/sync_batchnorm/unittest.
 doctra/third_party/docres/data/preprocess/crop_merge_image.py,sha256=f2NANY92s6IQ1hl1MAXfftFPIyIrj24O4TONjg7SXEc,4747
 doctra/third_party/docres/models/restormer_arch.py,sha256=BSwv_odCcp4HUZj3gv21e4IzFRBiyk8FjKAO8kF4YS8,12510
 doctra/ui/__init__.py,sha256=XzOOKeGSBnUREuDQiCIWds1asFSa2nypFQTJXwclROA,85
-doctra/ui/app.py,sha256=WpXUWHSs7wSYNjY4iBOZJHsKGQ88jDytvOFIjuhqAGE,44031
+doctra/ui/app.py,sha256=iFSAVZacL7iHB1SHhcUzperJGNQVWqUhvOYdlgjjt50,43623
 doctra/utils/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 doctra/utils/bbox.py,sha256=R2-95p0KiWvet3TH27TQVvCar7WJg6z0u3L21iEDF-A,674
 doctra/utils/constants.py,sha256=ZWOvNDrvETbQ_pxHiX7vUW4J5Oj8_qnov0QacUOBizI,189
@@ -59,9 +59,9 @@ doctra/utils/ocr_utils.py,sha256=Doa1uYBg3kRgRYd2aPq9fICHgHfrM_efdhZfI7jl6OM,780
 doctra/utils/pdf_io.py,sha256=c8EY47Z1iqVtlLFHS_n0qGuXJ5ERFaMUd84ivXV0b9E,706
 doctra/utils/progress.py,sha256=IKQ_YErWSEd4hddYMUiCORy0_kW4TOYJM891HUEq2_E,11901
 doctra/utils/quiet.py,sha256=5XPS-1CtJ0sVk6qgSQctdhr_wR8mP1xoJLoUbmkXROA,387
-doctra/utils/structured_utils.py,sha256=znC2zr80rZMfIV58lipZ8M4zPq6IF070pdwLBve1qiE,1251
-doctra-0.4.0.dist-info/licenses/LICENSE,sha256=HrhfyXIkWY2tGFK11kg7vPCqhgh5DcxleloqdhrpyMY,11558
-doctra-0.4.0.dist-info/METADATA,sha256=nlIT-QfxcwWi97jbQIastNHty8if3CyUv0LaDGiK7tk,28298
-doctra-0.4.0.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
-doctra-0.4.0.dist-info/top_level.txt,sha256=jI7E8jHci2gP9y0GYaWxlg9jG0O5n3FjHJJPLXDXMds,7
-doctra-0.4.0.dist-info/RECORD,,
+doctra/utils/structured_utils.py,sha256=vU84dsD8wIlTyMsA9hitorGH-eroQiVuWEpBTQBUT24,1478
+doctra-0.4.1.dist-info/licenses/LICENSE,sha256=HrhfyXIkWY2tGFK11kg7vPCqhgh5DcxleloqdhrpyMY,11558
+doctra-0.4.1.dist-info/METADATA,sha256=wXduiq7VJS5vf-TXdxpYFCKGfPyGYr5jGK0mwH3OjUw,28298
+doctra-0.4.1.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
+doctra-0.4.1.dist-info/top_level.txt,sha256=jI7E8jHci2gP9y0GYaWxlg9jG0O5n3FjHJJPLXDXMds,7
+doctra-0.4.1.dist-info/RECORD,,