Release on Project Indexly

Release v2.1.3

Fri, 29 May 2026 00:00:00 +0000

Release v2.1.3 (2026-05-29)

Indexly v2.1.3 is a production hardening release for structured-data inference workflows, backend routing, and test/runtime reliability.

Highlights

Introduces analytical dataset routing with optional DuckDB backend support for inference and analysis pipelines.
Expands infer-csv backend controls with freshness checks, safer lifecycle handling, and stronger merge diagnostics.
Hardens boxplot backend routing and static rendering behavior for CSV analysis outputs.
Improves optional dependency loading so analysis and backup workflows fail later and more clearly when extras are unavailable.
Refreshes CI/runtime dependency coverage and deep regression tests around routing and inference edge cases.

User-Facing Fixes and Features

Added backend-aware routing for analytical datasets and infer-csv commands.
Added columnar inference freshness checks to avoid stale inference artifacts.
Corrected CSV and boxplot routing to prefer catalog artifacts and preserve intended aggregation semantics.
Hardened datetime predictor coercion in regression-oriented inference paths.
Improved optional cryptography handling for backup workflows in constrained environments.

Dependency, CI, and Tooling Updates

Added visualization dependency coverage to CI where inference and plotting paths require it.
Kept pytest runtime artifacts inside repo-local ignored paths for cleaner local and CI runs.
Added a Docker test harness to improve reproducibility for dependency-heavy test paths.

Breaking Changes

None expected.

Release v2.1.2

Sat, 23 May 2026 00:00:00 +0000

Release v2.1.2 (2026-05-23)

Indexly v2.1.2 is a production-hardening release focused on structured data analysis reliability, dependency coverage for CI and package installs, and index hygiene around temporary Office lock files.

Highlights

CSV analysis preserves numeric statistics as numeric values while rendering terminal tables safely.
JSON and NDJSON loading now has stricter malformed-line handling, bounded record materialization, gzip support, and safer identifier string preservation.
SQLite analysis has safer identifier quoting, sampled profiling controls, multi-format exports, and clearer warnings.
Indexing now ignores Office lock files before they enter the scan pipeline.
Documentation and docs build dependencies were refreshed for the current analysis and AutoDoctor guidance.

Reliability Fixes

Hardened CSV date parsing, all-missing numeric handling, delimiter detection, and pipeline metadata propagation.
Hardened JSON, NDJSON, Socrata-style JSON, and persisted Indexly JSON summary behavior.
Hardened SQLite table profiling, export parsing, Mermaid schema rendering, and summary persistence.
Added CI/runtime coverage for table rendering and statistical analysis dependencies used by the CSV pipeline.
Resolved docs npm audit advisories and preserved markdown links in docs alert cards.

Release Tooling

Added a Homebrew tap update helper for release operations.
Refreshed release metadata, changelog data, and documentation hub highlights for v2.1.2.

Breaking Changes

None expected.

Release v2.1.1

Sat, 16 May 2026 00:00:00 +0000

Release v2.1.1 (2026-05-16)

Indexly v2.1.1 is a production-hardening release focused on backup restore safety, observer reliability, compare workflows, organizer/lister behavior, rename collision handling, and image metadata coverage.

Highlights

Backup restore now includes local verification and dry-run restore workflows.
Incremental backup restore selects the correct base more reliably.
Restore safety checks allow temp-root registries while continuing to block unsafe external temp archives.
Semantic observers are more stable, with persisted snapshots and graceful fallback when home paths are unwritable.
Compare, organizer, lister, and rename workflows have tighter handling around profiles, caches, counters, and database synchronization.
BMP image metadata extraction is handled consistently with the broader image metadata pipeline.

Reliability Fixes

Hardened backup restore validation, registry handling, and cryptography dependency coverage.
Stabilized lister root handling, cache fallback behavior, organizer execution, and profile-specific flags.
Reserved implicit rename counters for collision cases and avoided reapplying existing date prefixes.
Hardened rename database synchronization when handing renamed files to organizer workflows.
Improved compare workflow resilience for file and folder comparisons.

Documentation

Updated organizer documentation for lister cache and fallback behavior.
Refreshed release metadata, changelog data, and documentation hub highlights for v2.1.1.

Breaking Changes

None expected.

Release v2.1.0

Sat, 09 May 2026 00:00:00 +0000

Release v2.1.0 (2026-05-09)

Indexly v2.1.0 introduces safe search-index deletion with indexly clear-search. The command removes rows from fts_index.db without deleting source files and includes dry-run previews, confirmations, cache handling, and audit logging.

Highlights

clear-search --path PATH removes exact, prefix, or basename-matched search index rows.
clear-search --tag TAG [TAG ...] removes files matching any listed tag.
clear-search --all clears the full search index before a rebuild.
--dry-run previews all modes without database changes.
--yes supports safe automation for path, tag, and full-index cleanup.
Operation IDs connect the deletion plan, NDJSON logs, and final summary.

Search Improvements

Lowercase words such as and, or, not, and near are treated as literal search text unless uppercase FTS operators are used.
Search results can be sorted by relevance, newest, oldest, or path.
Tag statistics and search maintenance feedback are clearer.

Fixes And Reliability

Cache invalidation failures during clear-search are warnings rather than confusing post-delete hard failures.
Database errors during search-index cleanup now include recovery hints such as indexly doctor and indexly update-db.
Cleaned-data deletion handles basename, source path, and normalized full-path variants more robustly.
Debug table inspection uses the expected database connection behavior.

Documentation

Added Clear Search Results Safely.
Updated the usage guide, search guide, developer guide, FAQ data, database design notes, and changelog.

Breaking Changes

None expected.