-
Notifications
You must be signed in to change notification settings - Fork 335
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
SDK DataLoader
s 3: barebones C and C++ support
#5330
Conversation
Size changes
|
6439524
to
48498ce
Compare
b997a7f
to
66144c4
Compare
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
haven't looked through the entire pr yet but to pre-empt the question: using std::string_view
to pass around non-string bytes is super confusing and strange, let's not do that. I wish the obvious next thing to say was "just use array view" but that's std::span
and it's C++20. So std::byte* data, size_t data_size
is the way to go I fear..
bcd19c1
to
bcf4aa7
Compare
4e2bd3d
to
c5ecc9c
Compare
bcf4aa7
to
9e28fc9
Compare
c5ecc9c
to
cc8f15d
Compare
Exposes raw `DataLoader`s to the Rust SDK through 2 new methods: `RecordingStream::log_file_from_path` & `RecordingStream::log_file_from_contents`. Everything streams asynchronously, as expected. There's no way to configure the behavior of the `DataLoader` at all, yet. That comes next. ```rust rec.log_file_from_path(filepath)?; ``` ![image](https://github.com/rerun-io/rerun/assets/2910679/fac1514a-a00b-40c3-8950-df076080e1fc) This highlighted some brittleness on the shutdown path, see: - #5335 --- Part of series of PR to expose configurable `DataLoader`s to our SDKs: - #5327 - #5328 - #5330 - #5337
cc8f15d
to
fc4d655
Compare
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
feeling fairly strongly about the api for raw bytes not using a string_view, otherwise (minus typos) lgtm
Exposes raw `DataLoader`s to the Python SDK through 2 new methods: `RecordingStream.log_file_from_path` & `RecordingStream.log_file_from_contents`. Everything streams asynchronously, as expected. There's no way to configure the behavior of the `DataLoader` at all, yet. That comes next. ```python rr.log_file_from_path(filepath) ``` ![image](https://github.com/rerun-io/rerun/assets/2910679/0fe2d39c-f069-44a6-b836-e31001b3adaa) --- Part of series of PR to expose configurable `DataLoader`s to our SDKs: - #5327 - #5328 - #5330 - #5337 - #5351 - #5355
Co-authored-by: Andreas Reich <[email protected]>
Co-authored-by: Andreas Reich <[email protected]>
Introduces `RecordingStream::clone_weak`, so `DataLoader` threads started from the SDK don't prevent the recording from flushing once `main` goes out of scope, all the while making sure that `DataLoader`s run to completion. ![image](https://github.com/rerun-io/rerun/assets/2910679/94bf7b16-cf9b-40ce-a190-d255328af3f8) - Mitigates #5335 --- Part of series of PR to expose configurable `DataLoader`s to our SDKs: - #5327 - #5328 - #5330 - #5337 - #5351 - #5355
Adds new `RecommendedLoadSettings` that gets passed to all `DataLoader`s -- builtin and external -- in order to customize their behaviors. This includes: - A recommended recording ID - The ID of the currently opened recording in the viewer (not implemented) - Related: #5350 - A recommended entity path prefix - A recommended timepoint ```bash cargo r -p rerun-loader-rust-file -- run_wasm/src/main.rs --recording-id this-one --entity-path-prefix a/b/c --time sim_time=1000 --time wall_time=1709204046 --sequence sim_frame=42 | rerun - ``` ![image](https://github.com/rerun-io/rerun/assets/2910679/631d9798-c198-4e86-b6f8-32d0b849e7b2) Checks: - [x] external loader ran manually (`loader | rerun`) - [x] external loader via rerun (`rerun xxx.rs`) - [x] log_file with external loader (`log_file xxx.rs`) --- Part of series of PR to expose configurable `DataLoader`s to our SDKs: - #5327 - #5328 - #5330 - #5337 - #5351 - #5355
Introduces the new `DataLoaderSettings` business to Python and update examples accordingly (`external_data_loader` & `log_file`). ```bash python examples/python/external_data_loader/main.py --recording-id this-one --entity-path-prefix a/b/c --time sim_time=1000 --time wall_time=1709204046 --sequence sim_frame=42 examples/python/dna/main.py | rerun - ``` ![image](https://github.com/rerun-io/rerun/assets/2910679/bfda567d-3d16-42cd-be8e-8b1a0767a784) Checks: - [x] external loader ran manually (`python loader | rerun`) - [x] external loader via rerun (`rerun xxx.py`) - [x] log_file with external loader (`log_file xxx.py`) --- Part of series of PR to expose configurable `DataLoader`s to our SDKs: - #5327 - #5328 - #5330 - #5337 - #5351 - #5355 - #5361
This makes the `log_file` APIs behave more like the standard `log` APIs; i.e. they inherit the state of their associated `RecordingStream` (app_id, rec_id, timepoint, etc...). Also inherit the application ID while we're at it. Makes the API much nicer to use _and_ much more consistent with the rest. Checks: - [x] external loader ran manually (`python loader | rerun`) - [x] external loader via rerun (`rerun xxx.py`) - [x] log_file with external loader (`log_file xxx.py`) - [x] external loader ran manually (`loader | rerun`) - [x] external loader via rerun (`rerun xxx.rs`) - [x] log_file with external loader (`log_file xxx.rs`) --- Part of series of PR to expose configurable `DataLoader`s to our SDKs: - #5327 - #5328 - #5330 - #5337 - #5351 - #5355 - #5379 - #5361 - #5388 --------- Co-authored-by: Andreas Reich <[email protected]>
Introduces the new `DataLoaderSettings` business to C++ and update examples accordingly (`external_data_loader` & `log_file`). ```bash ./build/debug/examples/cpp/log_file/example_log_file --recording-id this-one --entity-path-prefix a/b/c --time sim_time=1000 --time wall_time=1709204046 --sequence sim_frame=42 rerun_cpp/tests/main.cpp | rerun - ``` ![image](https://github.com/rerun-io/rerun/assets/2910679/b979a24c-29b6-473b-91b1-de3832bea436) Checks: - [x] external loader ran manually (`loader.exe | rerun`) - [x] external loader via rerun (`rerun xxx.cpp`) - [x] log_file with external loader (`log_file xxx.cpp`) --- Part of series of PR to expose configurable `DataLoader`s to our SDKs: - #5327 - #5328 - #5330 - #5337 - #5351 - #5355 - #5379 - #5361 - #5388
I guess that's good enough 🤷. I don't know, my brain has been completely friend by C++ non-sense all day. This includes a fix to make sure that a viewer that was spawned from the python SDK is still allowed to spawn dataloaders implemented in python (`RERUN_APP_ONLY` shenaniganeries). - Fixes #4526 --- Part of series of PR to expose configurable `DataLoader`s to our SDKs: - #5327 - #5328 - #5330 - #5337 - #5351 - #5355 - #5379 - #5361 - #5388
Exposes raw
DataLoader
s to the C++ SDK through 2 new methods:RecordingStream::log_file_from_path
&RecordingStream::log_file_from_contents
.Everything streams asynchronously, as expected.
There's no way to configure the behavior of the
DataLoader
at all, yet. That comes next.string_view
(seerr_bytes
) and related, and the internet has been unhelpful to tell e whether I should or shouldn't.Part of series of PR to expose configurable
DataLoader
s to our SDKs:DataLoader
s 1: barebones Rust support #5327DataLoader
s 2: barebones Python support #5328DataLoader
s 3: barebones C and C++ support #5330DataLoader
s 4: working around shutdown brittleness #5337DataLoader
s 5: customizable (external) loaders for Rust #5351DataLoader
s 6: customizable (external) loaders for Python #5355Checklist
main
build: app.rerun.ionightly
build: app.rerun.io