v0.2.10.dev1

What's Changed

• Feat/fix by @rejojer in #3
• Restructure and update CLI entry point to run_pageindex.py by @rejojer in #4
• Fix prompt by @rejojer in #6
• Working by @rejojer in #10
• Fix: handle TOC items exceeding document length by @clarenceluo78 in #21
• Feat/markdown tree by @zmtomorrow in #32
• Feat/markdown tree by @zmtomorrow in #33
• Fix typo in header for the step: Extract JSON results by @mooncos in #118
• Fix TOC prompt variable typo (tob -> toc) by @matiasinsaurralde in #109
• Add GitHub Actions automation for issue deduplication and auto-close by @BukeLy with @Copilot in #128
• Fix backfill-dedupe pagination: replace gh issue list with gh api by @BukeLy in #132
• Allow github-actions bot to trigger claude-code-action by @BukeLy in #133
• Allow all users to trigger issue dedup by @BukeLy in #142
• fix: make ChatGPT_API_with_finish_reason return consistent tuple by @luojiyin1987 in #63
• fix: prevent infinite loop in extract_toc_content by @luojiyin1987 in #65
• Fix list_index variable shadowing in fix_incorrect_toc by @BukeLy in #167
• Integrate litellm for multi-provider LLM support by @KylinMountain in #168
• Add PageIndexClient with agent-based retrieval via OpenAI Agents SDK by @KylinMountain in #125
• Simplify root directory by @rejojer in #184
• Restructure examples directory and improve document storage by @rejojer in #189
• Simplify agentic vectorless RAG demo by @rejojer in #191
• Disable agent tracing and auto-add litellm prefix for retrieve_model by @rejojer in #194
• Polish demo docstring and migrate to pathlib by @rejojer in #197
• Bump litellm from 1.82.0 to 1.83.0 in the pip group across 1 directory by @dependabot[bot] in #214
• Add Dependabot config for GitHub Actions updates by @BukeLy in #241
• Bump the pip group across 1 directory with 2 updates by @dependabot[bot] in #247
• Add security CI workflows by @BukeLy in #248
• Update README with PageIndex File System announcement by @rejojer in #253
• Trim Cloud Service note in README by @rejojer in #255
• Fix Agentic RAG entry formatting in Updates by @rejojer in #256
• update README by @rejojer in #259
• update README by @rejojer in #261
• update README by @rejojer in #262
• Update README by @rejojer in #271
• Update README by @rejojer in #305
• Update README by @rejojer in #307
• Update README: Connect with Us buttons and header tagline by @rejojer in #311
• Tighten FinanceBench sentence in README by @rejojer in #312
• update readme by @rejojer in #313
• update readme by @rejojer in #318
• update readme by @rejojer in #319
• update readme by @rejojer in #322
• update readme by @rejojer in #332
• edit readme by @rejojer in #335
• edit readme by @rejojer in #336
• edit readme by @rejojer in #337
• Adds missing re import by @Mouli04 in #281
• fix: bump litellm to 1.84.0 to resolve python-dotenv install conflict by @KylinMountain in #342
• Bump actions/checkout from 4 to 7 by @dependabot[bot] in #338
• fix: prevent KeyError crash and context exhaustion in TOC processing by @Shreyansh1729 in #188
• Recognize whole-line bold as level-1 heading in markdown parser by @BukeLy in #250
• Add PageIndex Flash preview to README by @rejojer in #361
• Add PageIndex Flash section to README by @rejojer in #362
• Update README by @rejojer in #363
• Lazy-load litellm by @rejojer in #364
• Trim comments by @rejojer in #365
• Add page-level thinning to PDF pipeline by @rejojer in #366
• Add summary_model config by @rejojer in #367
• Add PageIndex Flash by @rejojer in #369
• Update Flash README links by @rejojer in #370
• Add --flash CLI flag and update README by @rejojer in #371
• Use OpenAI SDK directly for OpenAI models by @rejojer in #372
• Add Flash dependencies to requirements by @rejojer in #374
• Tree optimization for Flash (merge & expand) by @rejojer in #373
• Add benchmark to Flash README by @rejojer in #376
• Refine the Flash tree with merge by @rejojer in #375
• Keep merge inside --optimize by @rejojer in #377
• Trim the Flash README by @rejojer in #379
• Trim the Flash README by @rejojer in #380
• Fail fast on rejected keys and unknown models by @rejojer in #381
• Use embedded PDF bookmarks in Flash trees by @rejojer in #384
• Update README by @rejojer in #385
• Add a Flash summary flag by @rejojer in #386
• Exclude common words from roman numeral conversion by @rejojer in #387
• ci: run tests on every push and PR by @rejojer in #394
• feat: local mode for the PageIndex SDK (v0.2.9) by @rejojer in #389
• feat: agent tools and local chat for the PageIndex SDK (v0.2.10) by @rejojer in #396
• fix: post-merge review fixes for v0.2.10 by @rejojer in #402
• feat: Flash with full optimization becomes the default local indexing mode by @rejojer in #404

New Contributors

• @clarenceluo78 made their first contribution in #21
• @zmtomorrow made their first contribution in #32
• @mooncos made their first contribution in #118
• @matiasinsaurralde made their first contribution in #109
• @BukeLy with @Copilot made their first contribution in #128
• @BukeLy made their first contribution in #132
• @luojiyin1987 made their first contribution in #63
• @dependabot[bot] made their first contribution in #214
• @Mouli04 made their first contribution in #281
• @Shreyansh1729 made their first contribution in #188

Full Changelog: https://github.com/VectifyAI/PageIndex/commits/v0.2.10.dev1