Daft/daft/execution · xuanwu/Daft

GGitHubfeat: Support configuring Conda Env for Class UDFs in Flotilla (#5117 )

文件	最后提交记录	最后更新时间
__init__.py	Cython based Pyarrow hash functions (#79) * Implements Hash functions with XXHash * Leverages cython and pyarrow c++ api to do function calls * Currently hash functions for integers types and string are implemented * Implements Null handling for chunked_arrays and arrays * Updates poetry build to inject cython builder into setup.py * builds extensions in CI	4 年前
native_executor.py	fix: Add outer try-finally block in executor generator (#5633) ## Changes Made fixes the `no_running_event_loop errors` tests/dataframe/test_show_limit.py::test_show_limit_n_plain_counts_rows /home/runner/work/Daft/Daft/.venv/lib/python3.13/site-packages/_pytest/unraisableexception.py:78: PytestUnraisableExceptionWarning: Exception ignored in: <async_generator object NativeExecutor.run.<locals>.run_executor at 0x7f29ca809fc0> Traceback (most recent call last): File "/home/runner/work/Daft/Daft/daft/execution/native_executor.py", line 58, in run_executor yield batch GeneratorExit During handling of the above exception, another exception occurred: Traceback (most recent call last): File "/home/runner/work/Daft/Daft/daft/execution/native_executor.py", line 60, in run_executor _ = await result_handle.finish() ~~~~~~~~~~~~~~~~~~~~^^ RuntimeError: no running event loop warnings.warn(pytest.PytestUnraisableExceptionWarning(msg)) ## Related Issues Closes https://github.com/Eventual-Inc/Daft/issues/5636 ## Checklist - [ ] Documented in API Docs (if applicable) - [ ] Documented in User Guide (if applicable) - [ ] If adding a new documentation page, doc is added to `docs/mkdocs.yml` navigation - [ ] Documentation builds and is formatted properly	9 个月前
ray_actor_pool_udf.py	feat: Support configuring Conda Env for Class UDFs in Flotilla (#5117) ## Changes Made <!-- Describe what changes were made and why. Include implementation details if necessary. --> ## Related Issues <!-- Link to related GitHub issues, e.g., "Closes #123" --> #5104 ## Checklist - [ ] Documented in API Docs (if applicable) - [ ] Documented in User Guide (if applicable) - [ ] If adding a new documentation page, doc is added to `docs/mkdocs.yml` navigation - [ ] Documentation builds and is formatted properly (tag @/ccmao1130 for docs review) Signed-off-by: plotor <zhenchao.wang@hotmail.com>	8 个月前
udf.py	chore: bump `mypy` and `ruff` in pre-commit (#5836) ## Changes Made Bump `mypy` to 1.19.1 and `ruff` to 0.14.10 and resolve pre-commit errors. Rationale: my local `ruff` version is much higher than that in the project `pre-commit`, so I am getting a lot of linter warnings in my IDE. It is also about time to upgrade lints to utilize newer Python language features. ## Related Issues <!-- Link to related GitHub issues, e.g., "Closes #123" -->	8 个月前
udf_worker.py	feat: UDF metrics (#5507) ## Changes Made Adds ability to create and export custom metrics from udfs, particularly counter and gauge https://prometheus.io/docs/concepts/metric_types. Users define unique metrics by name, and can increment / set them in the udfs. These updates are stored in a static `MetricStore`, which is keyed by the metric name and type. The udf operator in swordfish can access this `MetricStore` to update it's own `RuntimeStats`, which is then subsequently used for dashboard + progress bar. <img width="1249" height="831" alt="Screenshot 2025-11-07 at 10 45 14 PM" src="https://github.com/user-attachments/assets/36cf9d2b-65d2-4b35-88d0-26c2fbd3aa88" /> ## Related Issues <!-- Link to related GitHub issues, e.g., "Closes #123" --> ## Checklist - [ ] Documented in API Docs (if applicable) - [ ] Documented in User Guide (if applicable) - [ ] If adding a new documentation page, doc is added to `docs/mkdocs.yml` navigation - [ ] Documentation builds and is formatted properly	10 个月前
vllm.py	feat: experimental vllm provider (#5443) ## Changes Made Adds the experimental `VLLMPrefixCachedProvider` for `daft.functions.ai.prompt`. Does async batching and prefix routing. When using the `VLLMPrefixCachedProvider`, `prompt` will create a `VLLMExpr` instead of a UDF, which Daft will turn into a custom VLLM operator. This operator is implemented as a streaming sink, and I had to make some minor changes to our streaming sink APIs to make the async batching mechanism work. ## Related Issues <!-- Link to related GitHub issues, e.g., "Closes #123" --> ## Checklist - [ ] Documented in API Docs (if applicable) - [ ] Documented in User Guide (if applicable) - [ ] If adding a new documentation page, doc is added to `docs/mkdocs.yml` navigation - [ ] Documentation builds and is formatted properly	10 个月前