Transforms complex documents like PDFs and Office docs into LLM-ready markdown/JSON for your Agentic workflows.
Python ai4sciencedocument-analysisdocxextract-datalayout-analysisocrparserpdfpdf-converterpdf-extractor-llmpdf-extractor-pretrainpdf-extractor-ragpdf-parserpptxpythonxlsx
fix(logging): change info logs to debug level for model loading and native table priority
myhloli8422b1bad164
diff --git a/mineru/backend/analysis/pdf/window.py b/mineru/backend/analysis/pdf/window.pyindex 5a93c9a6..26bfca36 100644--- a/mineru/backend/analysis/pdf/window.py+++ b/mineru/backend/analysis/pdf/window.py@@ -410,7 +410,7 @@ def _prepare_pdf_window("removed_formula_blocks": native_table_summary.removed_formula_blocks,"removed_formula_layout_items": native_table_summary.removed_formula_layout_items,}- logger.bind(native_table_priority=native_table_stats).info(+ logger.bind(native_table_priority=native_table_stats).debug("Hybrid native table priority. "f"effort={native_table_stats['effort']}, total={native_table_stats['total']}, "f"accepted={native_table_stats['accepted']}, "diff --git a/mineru/model/mfr/pp_formulanet/predict_formula.py b/mineru/model/mfr/pp_formulanet/predict_formula.pyindex 0ea2ce68..41e9d031 100644--- a/mineru/model/mfr/pp_formulanet/predict_formula.py+++ b/mineru/model/mfr/pp_formulanet/predict_formula.py@@ -80,7 +80,7 @@ class FormulaRecognizer(BaseOCRV20):if model_name == "PP-FormulaNet_plus-M" and self.device.type in {"mps", "cuda", "xpu"}:# 仅为经过适配的公式解码器启用快速 attention;CPU 和其他模型保持原路径。self.net.head.set_fast_attention()- logger.info(+ logger.debug("MFR loaded: model={}, device={}, dtype={}, attention={}, cache={}",model_name, self.device, self.ocr_inference_dtype,"sdpa" if getattr(self.net.head.decoder.model.decoder.layers[0].self_attn, "use_sdpa", False) else "eager",