Generate the full HTML analysis report — generate

Builds an Rmd file from data and renders it to HTML.

Usage

generate_report(
  data,
  theme_set,
  correlations_df,
  insights,
  export_files,
  consolidated = NULL,
  learning_context = NULL,
  provider = NULL,
  config = NULL,
  output_file = "analysis_report.html",
  irr_result = NULL,
  comparison_result = NULL,
  self_contained = TRUE,
  coding_results = NULL,
  coding_state = NULL,
  excerpt_verification = NULL,
  theme_group_tests = NULL,
  cooccurrence_tests = NULL,
  audit_log = NULL,
  response_cache = NULL,
  coverage = NULL,
  framework_spec = NULL,
  framework_archive = NULL,
  metric_interpretation = NULL,
  methodology_articulations = NULL,
  research_coverage = NULL,
  temporal_results = NULL
)

Arguments

data: tibble with all analysis columns
theme_set: ThemeSet object
correlations_df: Correlations tibble
insights: Insights list
export_files: List of export file paths
consolidated: ConsolidatedCodes list (or NULL)
learning_context: LearningContext object (or NULL)
provider: AIProvider object (or NULL)
config: ThematicConfig object (or NULL)
output_file: Path for the HTML report
irr_result: Inter-rater reliability result list (or NULL)
comparison_result: ComparisonResult object from compare_runs() (or NULL)
self_contained: If TRUE (default), produce a self-contained HTML file with all resources embedded. Set to FALSE for faster rendering and smaller file size (external CSS/JS will be referenced).
coding_results: Legacy CodingResults list as returned by as_coding_results. Used to populate per-theme entry tables with the codes assigned to each entry. Pass NULL to omit the codes column.
coding_state: ProgressiveCodingState with saturation data (or NULL)
excerpt_verification: Optional list returned by verify_excerpts containing substring_stats and (optionally) coherence_stats. When provided, the report's data-quality appendix shows excerpt validation results.
theme_group_tests: Optional tibble returned by compare_theme_groups (Mann-Whitney U tests). When provided, the correlation section gains a 'Theme Group Comparisons' subsection.
cooccurrence_tests: Optional tibble returned by test_theme_cooccurrence (chi-square / Fisher tests). When provided, the correlation section gains a 'Theme Co-occurrence' subsection.
audit_log: Optional AuditLog object forwarded to generate_ai_synthesis so the executive-summary AI call is recorded as an ai_request audit decision.
response_cache: Optional ResponseCache object forwarded to generate_ai_synthesis so the raw API response is written to the cache and referenced from the audit log.
coverage: Optional CorpusCoverage object (T0.3) from compute_corpus_coverage. When provided, the report renders a Tier-0 corpus-coverage card asserting that every entry surviving preprocessing reached the LLM (entry-level coverage; within-entry truncation is measured and disclosed when tracked). When NULL the card renders an explicit "coverage not computed" notice rather than silently omitting – absence is itself a transparency signal per AC4.
framework_spec: Optional FrameworkSpec object (Mode 3 only). When provided AND config$methodology$mode is "framework_applied", the report renders a Framework Declaration section with the framework's name, citations, epistemic stance, anomaly handling policy, and full constructs list. NULL on Mode 1 / Mode 2 runs.
framework_archive: Optional named list returned by archive_framework_spec carrying the archived framework's path + sha256 hash. When provided alongside framework_spec, the Framework Declaration section includes the sha256 fingerprint and a link to the archived spec.
metric_interpretation: Optional MetricInterpretation (from the Methodology Assistant). Threaded to aggregate_theme_statistics so per-subtheme stats are computed via the AI's chosen primitives (per-column, matched by name) with the legacy battery as fallback. NULL -> legacy only.
methodology_articulations: Optional MethodologyArticulations bundle. Drives the report's "Methodology Setup" section (relevance criterion + per-metric interpretations + per-theme temporal panel). When NULL, the renderer falls back to reading the archived rules/methodology_articulations.json under the output dir if present; when neither is available the section is omitted. When supplied and metric_interpretation is NULL, the metric interpretation is derived from this bundle.
research_coverage: Optional ResearchCoverage object (Mode 2 only). Drives the "Research-question coverage" section – where each named focus facet landed across the themes. When NULL, the renderer falls back to the archived rules/research_coverage.json under the output dir if present; when neither is available (or no separable facets were found) the section is omitted.
temporal_results: Optional list from analyze_temporal_patterns. Drives the "Longitudinal Patterns" section (period granularity, theme prevalence over time, theme emergence). Each chart is embedded only when its PNG exists in the output directory (the prevalence chart requires more than one time period; the emergence chart at least one dated theme). NULL -> the section is omitted.

Value

Path to generated HTML report