Skip to contents

qviewparsR 1.1.0

Packaging

  • Bump the version to 1.1.0 for the CRAN update; the reliability changes below were previously developed as 1.0.1 and never released.
  • Quote software names (‘H2’, ‘Java’) in the DESCRIPTION and phrase the description as full sentences.
  • Align citation metadata: citation("qviewparsR") now uses the CRAN DOI, CITATION.cff uses the Zenodo concept DOI, and all version fields read 1.1.0.

Measurement integrity

  • Refuse conflicting embedded records rather than infer a current database transaction from page-copy frequency. Use a fresh producer CSV/XLSX export with read_qview_report() if a binary report contains ambiguous records.
  • Preserve sample identities and decoded quoting/encoding; reject malformed nonblank numbers, unsupported report wells and ambiguous analyte keys.
  • Retain recognized missing/censored measurements and raw values. Summaries distinguish observed, missing and censored counts; numeric plots exclude bounds and missing values. All-missing groups have NA estimates.
  • Validate template geometry, unique assignments and positive finite dilution.
  • Print imported analyte names and units literally, without evaluating brace-delimited expressions as diagnostic formatting code.

App and export reliability

  • Clear stale results when uploads or parsing options change; bound image export memory and prevent plot-error logging from causing reactive loops.
  • Stage exports before replacement; require explicit CSV overwrite and preserve unrelated files. CSV bundles include segment metadata.

These are compatibility changes for ambiguous/malformed inputs and existing CSV destinations. Version 1.0.0’s majority-vote MVCC claim below describes the old implementation and is superseded: copy frequency cannot prove transaction recency or scientific correctness. No instrument validation or CRAN acceptance of this patch is implied.

qviewparsR 1.0.0

CRAN release: 2026-07-23

  • Standardize the named v5 package logo across README, documentation pages and available application headers, with matching light and dark artwork.

First release on CRAN (doi:[10.32614/CRAN.package.qviewparsR](https://doi.org/10.32614/CRAN.package.qviewparsR)). Pure-R parser for .Q-View binary project files (chemiluminescent multiplex ELISA plate imaging and quantification). No Java runtime, no H2 database driver, no compiled code.

New features

  • read_qview_report() reads the flat report exports Q-View writes next to the binary container – the ..._auto_report and ..._auto_all-parameters_report files, as either .csv or .xlsx – and returns the same qview object read_qview() builds. Use it when only the exports were kept and the original .Q-View project file is unavailable. It differs from read_qview() in two deliberate ways:
    • it captures the plain "Reduced Concentration" point estimate (one row per sample) with statistic == "reduced"; and
    • it preserves out-of-range cells instead of dropping them: a "< 52.50" cell yields concentration = 52.50 with a new flag column set to "<" (">" for upper bound, "incalculable" for Incalculable), so limit-of-quantification information survives import. A small example-report.csv fixture ships in inst/extdata/.

Reliability

  • read_qview() now resolves the superseded MVCC page versions that the embedded H2 container retains. A version truncated at a 2048-byte page boundary (fewer analytes, or a number cut mid-digit) is no longer mistaken for the current reading: for each physical (well, replicate, analyte) the value occurring most often across the committed page copies wins, breaking ties toward the most complete source row. Per-well pixel intensities now match Q-View’s own grid exports to 1e-6, and each well appears once in plate_layout.
  • read_qview_template() auto-detects the field separator, so semicolon-delimited (European-locale) and tab-delimited templates parse the same as comma-delimited ones.
  • plot(type = "replicate_scatter") no longer errors when a well-group label maps to more than one well; duplicate readings are averaged.

Reader

  • read_qview(path, strip_prefix = FALSE): parses a .Q-View container and returns a list of class qview with project metadata, the analyte panel (units, LOD / LLOQ / ULOQ, assay-control range), well-group sample assignments, per-well replicate pixel intensities, summary statistics, optional back-calculated concentrations, curve fits, and a plate layout (all tidy tibbles).
  • read_qview_template(): parses the companion well-assignment template CSV (NxM layout with Group Name / Group Type / Dilution Factor sections).

Helpers

  • strip_qview_prefix(): reverses the producer-side naming convention (ICal N -> Cal N, GLow -> Low, HHigh -> High, NFD... / N1234... -> original sample ID).
  • well_label(): vectorised plate-coordinate helper.

Methods

  • is_qview(): predicate for the S3 class.
  • print.qview(): compact one-screen summary.
  • summary.qview(): per-analyte mean / SD / CV / min / max grouped by well type, returned as a qview_summary tibble with its own print method.
  • plot.qview(type = ...): quick-look plate map, per-analyte intensity heatmap, and replicate-1-vs-2 scatter; viridis throughout.
  • as_tibble.qview(): long-format pixel-intensity tibble.

Export

Interactive front-end

  • qview_app(): monochrome bslib Shiny app with built-in dark/light toggle, hex-sticker brand, large upload cap (default 512 MB), per-table xlsx download, and a publication-ready 2x2 Overview tab (plate layout / pixel-intensity distribution / replicate concordance / mean PI by well type) with high-DPI PNG and vector PDF export.