redis-enterprise 0.10.0

Redis Enterprise REST API client library
Documentation
# API Inventory

Redis Enterprise does not publish an OpenAPI specification for the Enterprise
REST API, so this repo maintains a generated inventory seed based on the
official Redis request-reference pages.

The generated CSV lives at
[api-inventory.csv](./api-inventory.csv).

## Snapshot Metadata

- Refreshed: `2026-07-31`
- Source selector: Redis Software REST request reference under
  `https://redis.io/docs/latest/operate/rs/references/rest-api/requests/`
- Documentation version currently selected by `latest`: Redis Software `7.22`

## What The Inventory Contains

Each row records:

- the request-reference page where the endpoint was found
- the endpoint method
- the documented path
- the Redis docs page title and description
- a best-effort local SDK module guess

This is a discovery seed, not a final statement of completeness.

## Caveats

- The official docs are the best primary source currently available, but they
  are not an OpenAPI spec.
- The docs' `latest` Redis Software reference currently resolves to the 7.22
  documentation tree, while the local validation target used during this pass
  was Redis Enterprise Software `8.0.10-81`.
- Live clusters can expose behavior or fields that do not appear in the docs.
- Some SDK modules combine multiple documented request families, and some
  documented request families map to nested methods rather than standalone
  modules.

## Regenerate

Run:

```bash
python3 scripts/export_api_inventory.py
```

Or write to a custom path:

```bash
python3 scripts/export_api_inventory.py --output docs/api-inventory.csv
```

The weekly [Enterprise API contract workflow](./contract-automation.md) writes
the crawl to a temporary artifact and compares normalized method/path sets. It
never overwrites this reviewed inventory. Fetch, parse, local-output, CSV, and
semantic-drift failures are classified separately so an external outage cannot
masquerade as API removal.

## Audit SDK And Test Evidence

The generated coverage audit deduplicates the inventory by normalized HTTP
method and path, then records distinct static evidence for handler declarations,
Wiremock method/path matchers, request body and query matchers, response
fixtures, typed fixture deserialization, and explicitly annotated live tests.

Run its regression tests and regenerate both audit artifacts with:

```bash
python3 -m unittest discover -s scripts/tests -p 'test_*.py'
python3 scripts/audit_api_coverage.py
git diff --exit-code -- docs/api-coverage-audit.csv docs/api-coverage-audit.md
```

The audit deliberately does not count ordinary path literals or comments as
behavioral coverage. A Wiremock method/path matcher is reported separately from
an explicit call-count expectation, and response fixtures are not presented as
proof of field completeness. See
[api-coverage-audit.md](./api-coverage-audit.md) for all interpretation limits.

## Recommended Workflow

1. Regenerate the CSV from the official docs.
2. Run the method-aware static audit and review changed evidence dimensions.
3. Run the live smoke or compliance suite against a real Redis Enterprise
   instance.
4. Compare live-tested endpoints to the generated inventory.
5. File follow-up issues for undocumented, missing, or mismatched paths.