I want OCR-assisted matching for figure and table graphics in the full text parser.

I want OCR-assisted matching for figure and table graphics in the full text parser.: a task in MiMo-V2.6-RL-harbor-code: MiMo-V2.6-RL Code (Harbor) (Harbor dataset). I also want a public OpticalCharacterRecognitionGraphicMatcher(ocr model, max resolution=1024) matcher with get graphic…

The task

I also want a public OpticalCharacterRecognitionGraphicMatcher(ocr_model, max_resolution=1024) matcher with get_graphic_matches(semantic_graphic_list, candidate_semantic_content_list) so I can use this behavior directly. For a SemanticGraphic whose LayoutGraphic has a non-SVG local_file_path, the matcher should open…

Part of FineEnvs/MiMo-V2.6-RL-harbor-code.