docs(caching): a declared primary_key is a row constraint, not the cache key The caching refresh-mode page said a declared `acceleration.primary_key` replaces request metadata as the cache key, and its worked example passed a YAML list. Neither is true. A cache lookup is built from the request alone, so a key over response fields could not serve one even in principle; `primary_key` is a uniqueness constraint on the rows the accelerator stores. And `primary_key` deserializes as a string in every release, so the list form never parsed — the whole Spicepod is rejected at load. The example now projects the response fields it keys on with a `columns:` block and passes the key as a parenthesized string. That projection is only available from v2.0.0, so the 1.10.x and 1.11.x copies drop the example instead and state what a key can name there.

Check in versioned docs and remove generate-versions script (#1386) * Check in versioned docs and remove generate-versions script Transition from per-branch versioning to all versions checked in. Generated versioned docs from release branches 1.5 through 1.11. Removed generate-versions.sh and updated .gitignore. * Remove generate-versions script from build workflow Versioned docs are now checked in directly. Remove the script invocation from package.json build, drop the fetch-depth: 0 from CI (no longer needed), and update README to reflect the new process.

docs(retention): retention_sql alone satisfies an enabled policy, and an incomplete one is now reported (#2174) Sources: spiceai/spiceai#13857 Co-authored-by: claudespice <claudespice@users.noreply.github.com>

docs(hash-index): the row-count threshold never gated index creation; partitioned_arrow is supported (#2095) * docs(hash-index): the row-count threshold is never applied; partitioned_arrow is supported; two metrics are histograms index_threshold()/try_build() are dead outside benchmarks -- both IndexedMemTable construction paths call .build() unconditionally, so the documented 256 x CPU_cores minimum has never gated index creation in any release. is_hash_index_enabled() matches Arrow | PartitionedArrow since v2.0.0. hash_index_entries and hash_index_memory_bytes are registered as u64_histogram, not gauges. * docs(arrow): the hash index stores no key bytes, so index memory is per-entry not per-key ShardTable holds only Vec<Slot> (8-byte hash + packed 8-byte RowLocation); keys are never stored, which is why get_by_key documents a hash-collision risk. 'plus the key size' overstated the footprint for wide string keys. --------- Co-authored-by: claudespice <claudespice@users.noreply.github.com>

fix: Add missing acceleration engine modes and Turso snapshot support The data accelerator tables were missing file_create and file_update modes for DuckDB, SQLite, and Turso. The SQLite engine selection table was also missing memory mode. The snapshot mode page omitted Turso from the list of snapshot-capable engines. Verified against spiceai/spiceai trunk: - file_create: introduced in v1.10.0 (commit 9b07efea3) - file_update: introduced in vNext only (commit d26f07647) - Turso snapshot reload: supports_snapshot_reload() returns true

Check in versioned docs and remove generate-versions script (#1386) * Check in versioned docs and remove generate-versions script Transition from per-branch versioning to all versions checked in. Generated versioned docs from release branches 1.5 through 1.11. Removed generate-versions.sh and updated .gitignore. * Remove generate-versions script from build workflow Versioned docs are now checked in directly. Remove the script invocation from package.json build, drop the fetch-depth: 0 from CI (no longer needed), and update README to reflect the new process.

Improve documentation formatting for clarity and consistency (#1389) * Refine documentation across various sections for clarity and consistency - Updated descriptions in datasets, models, runtime, and views documentation to enhance readability and precision. - Improved explanations regarding retry mechanisms and refresh attempts in acceleration settings. - Clarified the purpose and functionality of the `WITH` clause in SQL documentation. - Enhanced troubleshooting guidance with detailed common issues and solutions for dataset loading, query performance, and AI chat results. - Revised use case descriptions for agentic apps, edge AI, federated MCP server, object-store AI engine, and more to improve clarity and impact. - Streamlined language throughout the documentation to ensure consistency and better user understanding. * Improve documentation formatting for clarity and consistency across multiple sections * Apply suggestion from @lukekim * Update documentation links for consistency and clarity across multiple sections

docs(snapshots): snapshot files carry the engine extension, not .db Since v1.11.1 (spiceai/spiceai#9318) the snapshot filename ends in the acceleration engine's extension — .duckdb, .sqlite, .cayenne or .turso — via AccelerationEngine::snapshot_extension(). The examples still showed the pre-v1.11.1 ".db" suffix. Scoped to v1.11.x and later; 1.8.x-1.10.x correctly document ".db", which is what snapshot_filename() hardcoded in those releases.