Skip to content

CLI reference

The forensics console script is a Typer app entered at src/forensics/cli/__init__.py — version 0.1.1. Every command and subcommand below has its own page with usage, options, and subcommands (when present).

CommandDescription
forensicsAI Writing Forensics Pipeline
forensics allRun full pipeline end-to-end: scrape → extract → analyze → report.
forensics analyzeRun analysis pipeline (change-point, drift, convergence, comparison).
forensics analyze section-contrastPhase 15 J6: per-author section-contrast tests (Welch + MW + per-family BH).
forensics analyze section-profilePhase 15 J3: newsroom-wide section descriptive report and J5 gate verdict.
forensics calibrateValidate detection accuracy against synthetic ground truth.
forensics commandsDump the full command catalog (for agent discovery).
forensics dashboardFull-screen pipeline dashboard (tui extra). Incompatible with --no-progress.
forensics dedupNear-duplicate fingerprint utilities
forensics dedup recompute-fingerprintsRecompute persisted simhashes for rows not stamped at the current NFKC (v2) version.
forensics exportExport SQLite + features + analysis into a single .duckdb file.
forensics extractRun feature extraction pipeline.
forensics featuresFeature-store maintenance commands (schema migrations, etc).
forensics features migrateUpgrade every feature parquet to the current schema version.
forensics lock-preregistrationLock analysis thresholds for pre-registration (run before analyzing data).
forensics migrateApply pending SQLite migrations (Phase 15 Step 0.2).
forensics preflightRun preflight checks and report pass/warn/fail for each boundary.
forensics reportRender Quarto forensic book.
forensics scrapeCrawl and fetch articles for configured authors
forensics setupLaunch the interactive setup wizard (requires the ‘tui’ extra).
forensics surveyBlind newsroom-wide AI-adoption survey.
forensics validateValidate config.toml, run preflight, and optionally probe live endpoints.