Show HN: LIBR tracing with source ledger rows and byte-exact PDF verification

An engineering note on replaying ordered financial transactions to preserve balance dips, same-day assumptions, and non-replenishment behavior. Each ledger step replays one ordered transaction against running state. The full Exit Protocol product remains private. This public engineering note isolates one narrow technical claim: LIBR-style tracing can be represented as deterministic replay over an ordered ledger. The companion repository uses synthetic CSV fixtures and regression tests so developers, forensic accountants, and diligence reviewers can inspect the calculation model directly. The artifact is for technical feedback — not legal conclusions, expert opinions, or product completeness claims. Synthetic fixture only. Each ordered transaction advances account balance and traceable balance under explicit LIBR assumptions. Swipe horizontally if the diagram extends beyond the screen. stateDiagram-v2 [*] --> Baseline Baseline --> ClaimOpen: separate-property deposit ClaimOpen --> Replay: chronological replay Replay --> DipDetected: withdrawal below trace level DipDetected --> TraceReduced: cap at lowest intermediate balance TraceReduced --> Replay: continue replay Replay --> LaterDeposit: later deposit LaterDeposit --> HoldTrace: balance rises HoldTrace --> Replay Model State machine representation LIBR tracing is stateful ledger math. Each transaction advances account balance and traceable balance together. Reviewers can inspect the replay path rather than inferring results from narrative summaries. traceable balance cannot exceed the account balance after any step. traceable balance cannot exceed the separate-property amount being traced. later deposits do not automatically restore a traceable amount previously reduced by a dip. same inputs and same ordering mode produce the same outputs on every run. Illustrative rows only. The public repository includes a fuller synthetic CSV and regression fixtures. Swipe horizontally to view all ledger columns on tablet widths. Red dip: A red dip occurs when the account balance falls below the claimed separate-property amount, reducing the candidate traceable amount under the selected LIBR assumption. Bank statements often lack reliable intraday timestamps. When multiple rows share a date, ordering changes the replay path. These modes expose calculation assumptions for review — they are not legal conclusions. Public repo: --ordering ledger. Preserves CSV order within each date. Baseline deterministic replay. Public repo: --ordering best_case. Deposits before withdrawals on same day. Exit Protocol product maps to maximize strategy. Public repo: --ordering worst_case. Withdrawals before deposits on same day. Exit Protocol product maps to minimize strategy. Exit Protocol product uses a court-balanced neutral strategy for default replay when same-day ambiguity exists. Standalone calculator with no Django dependency. CSV commingled-account sample for public inspection. ledger, best_case, and worst_case CLI switches. Regression coverage for full account depletion. Later deposits do not restore depleted trace. Edge-case fixture for added separate-property inflows. Print material trace events from the command line. Same fixture and mode yield the same trace on rerun. The repository proves the calculation primitive. The product wraps that primitive in a review workflow: Private matter context with role boundaries. Selected financial records normalized into ledger rows. Opening, activity, and closing balances tested before tracing. Rows surfaced for attorney and expert inspection. Structured export — not legal advice or an expert report. Supporting rows linked to selected statements or exports. Fingerprint of calculation state at export time. Post-export byte check for generated workpapers. Default workpaper path scoped to one claim per account. Requires Python 3.10+. No third-party packages. git clone https://github.com/Vinaygond/libr-state-machine-demo.git cd libr-state-machine-demo python -m unittest discover -s tests python libr.py examples/minimal_dip_ledger.csv python libr.py examples/synthetic_ledger.csv --ordering ledger python libr.py examples/synthetic_ledger.csv --ordering best_case python libr.py examples/synthetic_ledger.csv --ordering worst_case python libr.py examples/minimal_dip_ledger.csv --json The CLI prints material trace events from the synthetic fixture. Compare ordering modes to see how same-day ambiguity changes replay output.

Show HN: LIBR tracing with source ledger rows and byte-exact PDF verification

Key takeaways

More in computer-science