docs: refresh project documentation

[GreatV]

Summary

• polish the root README with concise setup, current model coverage, and focused examples
• add a complete Cargo feature guide and align model documentation with the auto-download registry
• document the current oar-ocr-vl model lineup, native inference paths, environment variables, and runnable examples
• correct stale API names, predictor descriptions, PP-OCRv6 defaults, tokenizer mappings, dictionary compatibility, and model sizes
• make standalone VL commands include the ONNX Runtime setup required by the crate

Why

The documentation had drifted from the current APIs and model registry. In particular, several VL commands omitted download-binaries, external snippets relied on undeclared transitive dependencies, PP-OCRv3 and PP-OCRv6 asset mappings were incomplete, and a few predictor and formula-model descriptions overstated the implemented behavior.

Impact

Users now get setup commands that work on a clean machine, accurate model and dictionary selections, complete feature guidance, and examples that match the current classic and VLM APIs. Existing acknowledgment descriptions are preserved.

Validation

• cargo test --workspace --doc
• cargo check --workspace --examples
• RUSTDOCFLAGS='-D warnings' cargo doc --workspace --no-deps
• cargo check -p oar-ocr-vl --examples --no-default-features --features download-binaries
• cargo check -p oar-ocr-vl --examples --no-default-features --features cuda,download-binaries
• cargo test -p oar-ocr-core --features auto-download registry_
• verified all tracked Markdown links, anchors, and code fences
• verified every registered model asset is documented and displayed sizes match registry bytes
• cargo fmt --all --check
• git diff --check HEAD

[gemini-code-assist]

Code Review

This pull request comprehensively updates the documentation, examples, and configuration options across the workspace to support newer Vision-Language Models (VLMs) such as PaddleOCR-VL-1.6, MinerU2.5-Pro, and MinerU-Diffusion-V1. Key changes include the addition of a detailed Cargo features guide, new environment variables for VLM performance and debugging overrides, updated pre-trained model tables with precise file sizes, and revised example snippets that incorporate the download-binaries feature. Additionally, minor documentation refinements were made to the core OCR builder methods. No review comments were provided, so there is no feedback to address.

Important

The consumer version of Gemini Code Assist on GitHub is being sunset. Starting June 18, 2026, new organization installations will be blocked, and all code review activity will officially cease on July 17, 2026.
For more details on the timeline and next steps, please review the Help Documentation.