New demo profile inst/extdata/demo_profiles/resque_Schoenbrodt_2026-03.json in the current scheme (meta 0.3.0, pub 0.9.0; 10 publications). The profile templates fall back to it when no JSON path is given.
Fixed
Profile templates (RESQUE_profile.qmd, RESQUE_profile_section_test.qmd) rendered directly without a json_path: an empty path is now treated like a missing one, and the demo profile is used.
Changed
inst/ref_set_psy/get_new_refset.R (script to rebuild the psychology reference set) can now resume after a crash. Intermediate results are no longer written to ~/refset_temp. They go to a folder with a fixed name in the system temp directory, e.g. RESQUER_refset_C15744967_article_n100000_seed42; the name encodes the sampling parameters. The script downloads one year at a time and skips years whose refset_<year>.RDS already exists and can be read. A re-run after a crash therefore continues with the first missing year, and the sample is identical to that of a single run. options(RESQUER.refset_dir = ...) sets another location.
Removed
Removed hardcoded local file paths in comments and debugging code.
RESQUER 0.10.21 (2026-10-21)
Removed
The deprecated, unfinished Shiny dashboard (inst/dashboard/dashboard_shiny.qmd, custom.scss), its (commented-out) launch instructions in the README, and the archived publication-table code that only the dashboard used (_archive/preprocessing_archive.R).
RESQUER 0.10.20 (2026-10-21)
Fixed
Profile report, “Rare samples used in studies”: the sample size is now read from P_Methods_PowerAnalysis_SampleSize (the field of the current scheme) instead of the obsolete P_SampleSize, which made the report always show “not provided”. Line breaks in the entry are collapsed; a missing or whitespace-only entry is shown as “not provided”.
Profile report: rendering no longer aborts in the impact table when no BIP! data are available (e.g. preprocess(get_BIP = FALSE)). The unused FNPR column was formatted before being dropped, which failed on an empty table.
Profile report, “Location of Data Collection”: the filter on studies with data (P_Data == "Yes") made a inconsistent selection. Whenever some studies had no data (or a missing P_Data), location categories could silently disappear from the chart (e.g. “Naturalistic”).
render_overview(clear_cache = TRUE) now takes effect: the cached applicant_data.RData in the JSON folder is ignored, all applicants are recomputed, and the cache file is overwritten. Previously the argument was passed to the template but never used, so an existing (possibly stale) cache was always reused.
RRS_category_summary() now finds the score categories in forms$pub$config as well as in forms$config (same precedence as the scorer). Previously it stopped with an error for exports that store the categories in forms$pub$config.
preprocess(): the derived replication column (preregistered replication) is now built from the current field P_Preregistration_Replication instead of the legacy name P_PreregisteredReplication, after the “not applicable” justification checks; studies without preregistration get No. Previously the column was always NA. It is not shown in the reports, so nothing changed. As a side effect, the share of missing indicators (ind_missing) no longer counts the empty legacy placeholder column; eligibility for the rigor score is unchanged for all test cases.
RESQUER 0.10.19 (2026-10-21)
Added
Profile report: new light-grey, struck-through badge images (inst/profile_qmd/assets/*_small_grey_strike.png) for open data, open materials, open code, and preregistration.
Changed
Profile report, “Open Practices” column: a practice answered as not available (NotAvailable/No) now shows the grey struck-through badge; “not applicable” (or no answer) shows no badge. Badges are smaller (20px high).
Profile report layout: body width increased to 1100px (sidebar 250px, margin 300px).
Badge images are now the same size (71 x 78 px); the green open-code badge was cropped and rescaled so it no longer appears smaller than the others.
RESQUER 0.10.18 (2026-10-21)
Fixed
preprocess(): the “not verified” count of the independent-verification pie data (OS_pie$Repro$No) now counts both the current answer code NotAvailable and the older code No; previously NotAvailable was never counted. Expect OS_pie$Repro to differ from earlier regression baselines.
preprocess(): claims of independent verification without a link are now also set to NotAvailable for AllResultsReproducible, and when the identifier is missing (NA) or blank (whitespace only). Scores are not affected.
Scorer (inst/js/score2.js): exists() in scoring rules now treats whitespace-only strings as missing. A whitespace-only identifier no longer earns points, and a whitespace-only “not applicable” explanation (*_NAExplanation, ZK2 justification) no longer removes the item from the maximum. Scores of the test-suite JSONs are unchanged (none contains such values).
preprocess(): “not applicable” answers now follow the same justification rule as the scorer. A whitespace-only *_NAExplanation, or an explanation field that is absent from the whole file, no longer counts as a justification; the answer is then shown as NotAvailable/No with a preprocessing note. Scores are not affected.
RESQUER 0.10.17 (2026-10-21)
Changed
Removed the unused internal helpers %and%, %or%, not() and exists() (R/operators.R), leftovers of the former R scorer. The package-internal exists() masked base::exists() inside the package namespace.
RESQUER 0.10.16 (2026-10-20)
Added
The profile report publication table has a new “Open Practices” column with small badges for open data, open materials, open code, and preregistration. Badges link to the respective material when an identifier is available.
Changed
The profile report layout is wider (sidebar 250px, body 1000px, margin 300px).
RESQUER 0.10.15 (2026-10-20)
Fixed
preprocess() now ensures that Title and Year columns always exist in the indicators data, so downstream RRS code and profile reports no longer fail when they are missing.
read_RESQUE() now completes partial DOIs (e.g., 10.1177/...) to full https://doi.org/ URLs, so title links work.
RESQUER 0.10.14 (2026-10-20)
Fixed
preprocess() and profile reports now handle unavailable RRS, BIP!, and OpenAlex results without aborting rendering.
compute_RRS() and RRS_radarchart() now render the radar chart when publication years are unavailable, omitting only the year histogram.
RESQUER 0.10.13 (2026-10-20)
Fixed
RRS calculations now consistently exclude inapplicable scoring categories, preventing mismatches between category summaries, per-publication scores, and radar charts.
Changed
compute_RRS() is now documented as the authoritative, reportable RRS calculation; it restricts scoring to eligible publications with DOIs.
The minimum required version of OAmetrics is now 0.5.4.
RESQUER 0.10.12 (2026-10-14)
Changed
The minimum required version of openalexR is now 3.1.2 to avoid compatibility issues.
RESQUER 0.10.11 (2026-10-01)
Fixed
Profile and overview reports now exclude outputs without data or scoring indicators from sample and data-type summaries.
Relative Rigor Score calculations now handle outputs without scorable indicators or DOIs correctly, avoiding invalid scores and errors in profile radar charts.
RESQUER 0.10.10 (2026-08-04)
Added
New check_quarto() helper that verifies a minimum required Quarto version is installed and errors informatively if not; called at the start of render_profile() and render_overview().
DT added as a suggested package.
Changed
render_profile() now restores the working directory via on.exit(), so it is reset even if rendering fails.
RESQUER 0.10.9 (2026-07-30)
Changed
render_overview() now defaults to using the median FNCS (Field-Normalized Citation Score) instead of the mean, with a new FNCS_type argument ("median" or "mean") to switch between them.
RESQUER 0.10.8 (2026-07-07)
Added
Enable multi-study support
RESQUER 0.10.7 (2026-07-06)
Added
Scientific impact tables now include a new first column (Top 3 & Merit Statement).
Top 3 selection (P_TopPaper_Select) is displayed with a star icon.
Merit statements are now shown directly in the first column as an info popover icon (below the star when both are present).
An optional Applicant's Comments popover column is shown when any *_Comment fields contain content.
Changed
Scientific impact tables no longer use striped row styling.
Rows for selected top papers are highlighted in pastel yellow.
Impact table logic now keeps papers even when BIP returned no values.
Method-type summary in the profile now excludes *_Comment fields.
OpenAlex team-science summaries now use unique papers (DOIs) for retrieval counts and percentages.
Waffle charts in the RRS overview now use dynamic scaling with a minimum reference size (avoiding a warning)
Preregistration sector lookup now matches category labels more robustly.
Several profile labels/texts were harmonized from “papers” to “studies” where multi-study submissions are shown.
Fixed
Popover content for comments and merit statements is now HTML-escaped, preventing broken markup when comments contain quotes or special characters.
sum_MC() now excludes all *_Comment columns and sums only logical indicator columns, improving robustness across indicator sets.
RRS score joins now deduplicate per DOI before merging, avoiding many-to-many join inflation for multi-study papers.
BIP paper counters now deduplicate per DOI, so paper-level counts are no longer inflated by multi-study entries.
high_impact and low_impact tables now both explicitly suppress row names.
Known issues
In some multi-study situations, the impact tables have duplicates.