release: prepare v0.7.0 - Split Table Merging Feature enhancement

[AdemBoukhris457]

Summary

This PR prepares the release of v0.7.0, which introduces automatic split table detection and merging capabilities. The release adds a feature that automatically detects and merges tables split across page boundaries, along with documentation enhancements and visualizations.

Release Highlights

Split table merging

• Two-phase detection (proximity + structural validation)
• LSD-based column detection for accurate table matching
• Configurable thresholds and confidence scoring
• Fallback mechanisms for borderless tables
• Automatic composite image creation

Documentation

• Mermaid flowcharts replacing ASCII art
• Step-by-step narrative explanations
• Complete API reference
• Troubleshooting guides

Changes Included in v0.7.0

Code changes

• Added SplitTableDetector class with detection logic
• Extended StructuredPDFParser with split table merging parameters:
  • merge_split_tables (default: False)
  • bottom_threshold_ratio, top_threshold_ratio
  • max_gap_ratio, column_alignment_tolerance
  • min_merge_confidence
• Implemented two-phase detection algorithm:
  • Phase 1: Proximity checks (position, overlap, gap, width)
  • Phase 2: Structural validation (column detection, alignment, confidence)
• Added fallback mechanisms for edge cases
• Integrated merge processing into parsing pipeline

Documentation

• Added split table merging guide (split-table-merging.md)
• Converted 8 ASCII diagrams to Mermaid flowcharts
• Added "Process Overview" narrative section
• Updated API reference with new parameters
• Fixed documentation links and navigation

Recent Merge Requests Included

This release incorporates the following merge requests:

• docs: Add Mermaid diagrams to split table merging documentation #71 - docs: Add Mermaid diagrams to split table merging documentation
• fix(docs): Fix broken link paths in split table merging documentation #70 - fix(docs): Fix broken link paths in split table merging documentation
• fix(mkdocs): Add split table merging docs to navigation and fix broken links #69 - fix(mkdocs): Add split table merging docs to navigation and fix broken links
• docs: Add split table merging documentation #68 - docs: Add split table merging documentation
• feat: Add automatic split table detection and merging #67 - feat: Add automatic split table detection and merging

Breaking Changes

None. This release is backward compatible. The split table merging feature is opt-in (default merge_split_tables=False) and does not affect existing functionality.

Release Notes

v0.7.0 adds automatic split table detection and merging. The parser can detect tables split across pages and merge them into single images for improved extraction.

The feature uses a two-phase approach: proximity detection filters candidate pairs, and structural validation confirms matches using computer vision. It includes configurable thresholds, confidence scoring, and fallbacks for borderless tables.

Documentation includes interactive Mermaid diagrams and a step-by-step narrative explaining the process. The feature is disabled by default and can be enabled via merge_split_tables=True.

---

Version bump: 0.6.2 → 0.7.0