Add WebHarbor reset and smoke verification script

[Lxr-max]

Summary

• adds scripts/check_reset_smoke.py to verify WebHarbor control-plane resets, per-site homepage smoke checks, and local runtime-vs-seed DB MD5 parity
• adds scripts/test_check_reset_smoke.py with standard-library coverage for MD5 checks, site filtering, strict mode, JSON output, and local HTTP reset/homepage behavior
• adds a short README usage section for the new reset/smoke checker

Why this helps WebHarbor review

• gives contributors and reviewers a single repo-level command to sanity-check /reset/<site> behavior
• makes it easier to spot broken homepages and missing/dirty local DB state before opening a website PR
• supports both Docker-backed runs and local dry-runs where DB files may be missing or the control server is offline

CLI usage

python scripts/check_reset_smoke.py --site amazon
python scripts/check_reset_smoke.py --control-url http://localhost:8101
python scripts/check_reset_smoke.py --base-host localhost
python scripts/check_reset_smoke.py --timeout 10
python scripts/check_reset_smoke.py --json
python scripts/check_reset_smoke.py --strict

Checks included

• site discovery from websyn_start.sh and control_server.py
• control server health check via GET /health
• per-site POST /reset/<site> checks, or optional --reset-all
• homepage smoke checks via the registered site port
• local instance/*.db vs instance_seed/*.db MD5 comparison when visible
• graceful skip/warning behavior when local DB files are not available

Test commands and results

• py -m py_compile scripts/check_reset_smoke.py ✅
• py -m py_compile scripts/test_check_reset_smoke.py ✅
• py scripts/test_check_reset_smoke.py ✅ (7 tests passed)
• py scripts/check_reset_smoke.py --help ✅
• py scripts/check_reset_smoke.py --json --site amazon --timeout 0.5 ✅ output is valid JSON; on a machine without a running control server it reports clear connection/timeout failures instead of crashing

Notes about running with/without Docker

• with Docker or a local WebHarbor control server running, the script performs live reset and homepage checks
• without Docker or without the control server, the script still runs and reports clear reset/homepage connection failures
• when local sites/<site>/instance/*.db and sites/<site>/instance_seed/*.db are not present, MD5 checks are reported as skipped warnings rather than fatal errors by default

Notes

• this is not a website contribution
• no HF assets are involved
• .assets-revision was not modified
• no existing site implementation was changed

Known limitations

• live reset checks intentionally depend on a running local control server and may fail with clear connection errors when none is available
• MD5 comparison only runs when the relevant DB files are locally visible from the current checkout

[jackjin1997]

Claiming this for a reviewer-owned continuation. I will preserve @Lxr-max's original commit and attribution, and audit scripts/check_reset_smoke.py against current main: site/port discovery against the real websyn_start.sh and control_server.py registries, control-plane /health, per-site /reset/<site> and /reset-all behavior, homepage smoke, runtime-vs-seed DB parity semantics, CLI/exit-code/JSON contract, malformed-input handling, and the bundled tests — exercised against a real built container, not only fixtures.

This is repository tooling rather than a new mirror, so site UI fidelity, task runs, and Hugging Face asset checks are not applicable. I will link a separate Draft Review: PR here before maintainer handoff, and I will not merge either PR.

[jackjin1997]

Draft reviewer PR opened: #108 — Review: WebHarbor reset and smoke verification script (#47, original by @Lxr-max). @Lxr-max's commit is preserved there; the reviewer changes sit on top.

Summary of what the review found and fixed, verified against a live control plane:

• the DB parity check hashed <repo>/sites/<site>/{instance,instance_seed}, but the control plane resets /opt/WebSyn/<site>/instance inside the deployment (COPY sites/ /opt/WebSyn/). With the environment stopped it still reported "runtime DB matches seed DB"; with the deployment verified clean, a row added to the local checkout produced a failing "differs from seed DB after reset". On a fresh clone the check was skipped for all 24 sites. Now there is --docker-container / --db-root, every result reports the source it hashed, and no parity verdict is reported without a reset to attribute it to.
• registry drift, a missing websyn_start.sh, unparseable SITES and a BASE_PORT mismatch escaped as tracebacks; they are structured findings now.
• --strict failed a correct environment because absent local instance/ warned; the README's own --strict example therefore failed on the documented Docker workflow.
• a failed --reset-all was counted once per registered site.

The registry discovery, HTTP reset/health/homepage checks and the structured handling of unknown sites and ambiguous DB pairs were already correct and are unchanged. #108 stays in draft until an independent blind review of the frozen scenario runs comes back; I will not merge either PR.