sciforge · examples

Six modules.
Six demos.

Interactive examples for every sciforge module. Each one uses real sciforge code and shows the Python API alongside the UI.

Reproducibility Dashboard

Score your experiment codebase. Drop a .py or .ipynb file to detect missing seeds, hardcoded paths, out-of-order notebook cells, and more.

sciforge.hypotest

Hypothesis Ledger

Declare, version, test, and audit scientific hypotheses. Full lifecycle from pending → supported/rejected, with linked experiments and history.

sciforge.unitflow

Unit Pipeline Calculator

Build multi-step physical unit pipelines. See live unit propagation, catch incompatible operations, and browse the unit registry.

sciforge.crossvalbio

CV Split Explorer

Visualise patient, batch, phylo, and temporal splits side-by-side. See exactly which samples are in each fold and detect data leakage.

sciforge.papertrail

Lineage Visualiser

Explore the full data→transform→figure→claim DAG. Run a staleness check to see which figures and claims are invalidated by changed data.

sciforge.litmap

Literature Graph

Interactive citation network of landmark ML/NLP papers. Click nodes to explore dependencies, centrality, and citation chains.

Research Audit Trail

Generate a submission-ready RAT report combining repro scores, hypothesis ledger, and carbon footprint for peer reviewers.

sciforge.bundle

Reproducible Bundle

Package your entire research lifecycle into a single ZIP. Includes code, config, provenance, and the audit report.

🚀 Automated releases with sciforge

The examples/release.py script uses sciforge to validate sciforge before shipping. It runs repro.score() (must be ≥80), checks the hypotest ledger for untested hypotheses, records the release in papertrail, then builds and publishes automatically.

# Validate only (no publish) python examples/release.py --dry-run # Publish a patch release python examples/release.py --bump patch # Publish a minor release python examples/release.py --bump minor