v0.7.3

What's Changed 🚀

✨ Features

• feat(observability): Support exporting Flotilla metrics @srilman (#6122)
• feat: Nightly installations under nightly.daft.ai @desmondcheongzx (#6175)
• feat: Support additional OTEL_* configuration envs @srilman (#6148)
• feat(lance): Support lance namespace read and write @shaofengshi (#5980)
• feat: add unity oauth m2m access token support @cckellogg (#5839)
• feat: Support snapshot properties for Iceberg writes @desmondcheongzx (#6139)
• feat: delimiter for agg_concat @aaron-ang (#6099)
• feat: add uuid function @everySympathy (#5983)
• feat: add custom date and timestamp formatting for CSV writes @madvart (#6073)
• feat(observability): Export metrics out as a table with the result @srilman (#6055)
• feat: comparison ops for list and struct types @aaron-ang (#6104)
• feat(frontend): enhance dashboard UI and fix Ray runner state reporting @Jay-ju (#6063)
• feat(dashboard): backend implementation @Jay-ju (#6062)
• feat: string casing functions @aaron-ang (#6096)
• feat: list_contains expression @aaron-ang (#6095)
• feat: specify text embedding dim @aaron-ang (#6097)
• feat: Expr.var with ddof @aaron-ang (#6105)
• feat(lance): add nearest vector search support @huleilei (#6025)
• feat: distance and similarity functions @aaron-ang (#6098)
• feat: add gravitino connector optional dependency @shaofengshi (#6083)
• feat: Add support for ignoring null fields when writing json @gpathak128 (#6049)

🐛 Bug Fixes

• fix(docs): Fix broken links in modalities documentation @everettVT (#6197)
• fix: Render Map columns as python dicts instead of list[struct] in terminal prints @srilman (#6198)
• fix: Fix merge conflict with OTEL configuration PR @srilman (#6187)
• fix(observability): Truncate progress bar names by characters, not bytes @desmondcheongzx (#6180)
• fix: fix concat.rs @colin-ho (#6174)
• fix(test): bump actor UDF timeout from 10s to 60s to reduce flakiness @ykdojo (#6163)
• fix(optimizer): Fix bug with pushing filters through anti-joinsI s @desmondcheongzx (#6150)
• fix: Allow is_in to accept sets, tuples, and other iterables @desmondcheongzx (#6115)
• fix(ci): restore Rust code coverage by pinning cargo-llvm-cov @desmondcheongzx (#6146)
• fix: make df.into_partitions() work when input num == num_partitions @everySympathy (#6061)
• fix(udf): ensure per-call kwargs in udf v2 are uniquely bound per call site @huleilei (#6079)
• fix(rustfmt): ignore parquet directories @aaron-ang (#6101)
• fix: No need extra 1 for OffsetBufferBuilder @colin-ho (#6057)
• fix: use unique bucket names for running tests in parallel @rchowell (#6052)

🚀 Performance

• perf: zero copy on from_vec @universalmind303 (#6172)
• perf: Only Serialize Required Cols in Actor UDFs @plotor (#5884)

♻️ Refactor

• refactor(arrow2): migrate time.rs temporal methods to arrow-rs @rohitkulshreshtha (#6160)
• refactor(arrow2): migrate null checks to arrow-rs @rohitkulshreshtha (#6152)
• refactor(arrow2): migrates sparse-tensor to arrow-rs @cckellogg (#6179)
• refactor(swordfish): Separate concat and make streaming sink single input @colin-ho (#6059)
• refactor(arrow2): migrates hll-sketch to arrow-rs @cckellogg (#6169)
• refactor(arrow2): use arrow-rs for python conversions @universalmind303 (#6130)
• refactor: Migrate hll_merge.rs from arrow2 to arrow-rs @rohitkulshreshtha (#6158)
• refactor(arrow2): remove arrow2 from minhash @cckellogg (#6149)
• refactor: Migrate product.rs from arrow2 to arrow-rs @rohitkulshreshtha (#6159)
• refactor(arrow2): migrates hashing kernel from arrow2 to arrow-rs @rchowell (#6056)
• refactor(arrow2): migrate the extra easy kernels to arrow-rs @universalmind303 (#6145)
• refactor(arrow2): migrate BinaryArray iceberg_truncate to arrow-rs @rohitkulshreshtha (#6157)
• refactor(arrow2): migrate DataArray full_null/empty to arrow-rs @rohitkulshreshtha (#6151)
• refactor(arrow2): remove a bunch of from impls that used arrow2 @universalmind303 (#6137)
• refactor(scalarudf): followup to uuid pr @universalmind303 (#6129)
• refactor(arrow-rs): Remove arrow2 from the search_sorted kernel @desmondcheongzx (#6034)
• refactor(arrow-rs): Migrate approx count distinct @desmondcheongzx (#6038)
• refactor(arrow-rs): migrate is_in and get_lit ops to Arrow-rs arrays @huleilei (#6085)
• refactor: Migrate utf8 left function to arrow-rs @huleilei (#6004)
• refactor(arrow2): migrate daft-functions-utf8/split to arrow-rs @colin-ho (#6046)
• refactor(arrow-rs): Migrate sketch_percentile kernel @desmondcheongzx (#6044)
• refactor(swordfish): Separate joins @colin-ho (#6042)
• refactor(arrow2): remove some arrow2 based from impls @universalmind303 (#6054)
• refactor(arrow2): migrate sum, min, max agg kernels @kevinzwang (#6045)
• refactor(arrow2): utf8 comparison kernels @kevinzwang (#6032)
• refactor(flotilla): Use stream::iter instead of channel in flotilla source nodes @colin-ho (#6043)
• refactor(arrow2): remove arrow2 *_array() methods from ImageArray @universalmind303 (#6050)
• refactor(arrow-rs): Migrate binary from_iter methods @desmondcheongzx (#6040)
• refactor(observability): Refactor StatSnapshot to be predefined structs @srilman (#6033)

📖 Documentation

• docs: Add daft.File usage throughout modalities @everettVT (#6074)
• docs: Governance proposed changes @jaychia (#6117)
• docs: Clarify imagenet benchmark setup @desmondcheongzx (#6147)
• docs: add TosConfig documentation reference in config.md @huleilei (#6068)
• docs: Update daft.File API and add docstrings, improve error handling in tests @everettVT (#5877)
• docs: add mm structured outputs tutorial @everettVT (#5816)
• docs: Add end-to-end image pipeline example and regression test @huleilei (#6006)

✅ Tests

• test: Add missing result parameter to on_query_end in Google AI test @desmondcheongzx (#6184)

👷 CI

• ci: Exclude common-arrow-ffi from rust tests @desmondcheongzx (#6173)
• ci: Accept pandas StringDtype in schema override tests @desmondcheongzx (#6165)
• ci: Add pyarrow to wheel build test dependencies @desmondcheongzx (#6109)
• ci: Add pytz and numpy to wheel build test dependencies @desmondcheongzx (#6108)
• ci: Fix nightly workflow permissions @desmondcheongzx (#6037)

🔧 Maintenance

• chore: Intermediate op single input @colin-ho (#6189)
• chore: Revert "refactor(arrow2): migrates hashing kernel from arrow2 to arrow-rs" @universalmind303 (#6164)
• chore: Remove ci/ folder @srilman (#6154)
• chore: agg_concat kernel nits @colin-ho (#6136)
• chore: Remove vendored parquet-format-safe, use upstream patched version @desmondcheongzx (#6118)
• chore(deps): bump the all group with 5 updates @dependabot[bot] (#5908)
• chore: Allows make clean to skip cleaning the python virtual environment @plotor (#6103)
• chore: correct comment of enable_scan_task_split_and_merge in set_execution_config @everySympathy (#6077)
• chore: Update lru dependency @desmondcheongzx (#6041)

⬆️ Dependencies

• chore(deps): bump the all group with 5 updates @dependabot[bot] (#5908)

Full Changelog: v0.7.2...v0.7.3