DuXiu Collection Document

Industrial extractive metallurgy increasingly relies on large, heterogeneous digital technical archives, yet incomplete metadata, fragmented pagination, and uncertain file integrity hinder process modelling, plant design, and reproducibility. This work aimed to quantify structural characteristics, integrity, and potential information loss in a representative metallurgical monograph archived in a large-scale digital repository, and to assess its suitability as a reliable data source for engineering analysis. A 360-page digital volume was examined using file-level hash verification, pagination mapping, and pixel-count-based page completeness analysis. Statistical descriptors were derived from total pixels (1.68×10⁹), reported uncompressed size (21.85 MB), and main-page coverage (348 of 348 expected pages, 100% completeness) to infer potential gaps in technical content. Hash analysis confirmed full consistency between header and content identifiers, with matching md5 and sha256 fingerprints and 0% detected corruption. The main technical section showed 100% page presence, while auxiliary content accounted for approximately 3.3% of total pages. The ratio of digital size to pixel count indicated an effective compression efficiency above 90%, with no evidence of lossy structural degradation. These results indicate that, for this case, digitally archived metallurgical literature can provide structurally complete datasets for downstream text and image mining, enabling quantitative extraction of process parameters and kinetic data with negligible structural loss and supporting reliable reuse in industrial flowsheet development and optimization.