Skip to content
stanimirivanovPublic

About

Decision and maintenance platform for automated test portfolios. Answers: Which tests should run, when, and in what order? Which tests have become stale, and can they be adapted safely? Combines deterministic Test Impact Analysis, history-based Predictive Test Selection, risk and cost optimization, and evidence-based test adaptation.

Resources

Code of conduct

Contributing

Security policy

Stars

0 stars

Watchers

0 watching

Forks

Argus

Argus selects tests for a software change, explains the evidence behind the selection, and proposes constrained, reviewable maintenance when tests become stale. The current end-to-end slice covers functional API cataloging, change impact, selection, execution evidence, and validated repair. Functional UI selection and catalog conformance are available; browser execution and repair are planned.

Start locally

Use the exact Go version in go.mod. From the repository root, start the database and apply the Argus and optional DBOS evaluation migrations:

docker compose -f compose.local.yaml up -d --wait
go run ./cmd/migrate --local
go run ./cmd/migrate --local --dbos-evaluation

To run the opt-in DBOS webhook path, supply your own GitHub read token and webhook secret, then start the control plane:

export ARGUS_GITHUB_TOKEN='<your repository-read token>'
export ARGUS_GITHUB_WEBHOOK_SECRET='<your own random webhook secret>'
export ARGUS_DBOS_EVALUATION='true'
go run ./cmd/control-plane --local

Omit the DBOS migration and flag to use the default path. --local supplies only the loopback development database; it does not supply real GitHub credentials. The server listens on 127.0.0.1:8080 and accepts signed POST /webhooks/github. Local startup and troubleshooting explains Compose, credentials, existing PostgreSQL, and the evaluation's limits. Normal builds and validation need no database or credentials.

Find the right guide

The documentation map routes each task to its canonical source. Common entry points:

Build and verify

make doctor       # report the effective toolchain
make bootstrap    # resolve pinned dependencies once
make verify       # fast build, checks, and ordinary tests
make fmt          # format before review
make validate     # complete non-mutating acceptance checks
make db-validate  # PostgreSQL integration; requires a disposable test server
make dbos-evaluate # isolated DBOS evaluation; needs Docker, no credentials

See the Makefile for focused targets and the developer quickstart for tool versions and platform requirements. Every executable supports standalone --help and --version without credentials. Direct-database commands remain administrative or transitional boundaries, not the ordinary CI integration pattern. The DBOS evaluation verdict explains the self-provisioned evaluation and its production-adoption limits.

About

Decision and maintenance platform for automated test portfolios. Answers: Which tests should run, when, and in what order? Which tests have become stale, and can they be adapted safely? Combines deterministic Test Impact Analysis, history-based Predictive Test Selection, risk and cost optimization, and evidence-based test adaptation.

Resources

Code of conduct

Contributing

Security policy

Stars

0 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages