neuropsyol / senda Goto Github PK

The current solution of including liblsl has multiple issues:
We use an ancient version of liblsl which can not be compiled w/ current versions of the NDK. This leads to several problems: The ancient version of the NDK we use is not available easily on GH actions using pre-packaged workflows. I tried to update the liblsl and abandoned the effort as too painful. Solution: Just use liblsl-Android. liblsl updates come more or less for free, we can declutter our code by a lot, etc...

This makes the sensors unavailable for other apps/other instances of SENDA until a power cycle.

I already did some work on this, but I am not nearly done. The logic containing the data flow is all over the place, intermixed w/ UI code, mis-using the sensor interface, etc...
Solution: Separate the UI code and the data flow logic.

At the moment always 60Hz

.. and whether we can stream their data using our app in case we do not succeed in creating an LSL stream from XSense.

This issue will be assigned to Sören once he has a Github account.

Implement a status indicator for outgoing streams that show whether the streams are receiving data from the built-in sensors when opening the LSL stream.
The indicator can be a simple green bubble, or anything intuitively understandable by the user.

This will be assigned to Sören once he has a Github account.

The current solution still depends on manual copies of files and edits. Find a way to make the migration possible by e.g. changing a single line in a config file.

Googling SENDA the original, unmaintained copy on Ali's repository shows up. Make people use this version instead.

Use consistent names, decide on a naming scheme in the first place, etc...

Deprecated Gradle features were used in this build, making it incompatible with Gradle 8.0.
See https://docs.gradle.org/7.5/userguide/command_line_interface.html#sec:command_line_warnings

A minimum viable feature is included now, but the location update rate is rather low und irregular. Find ways to make it more steady and give location updates while the app is in the background.

The paper version of SENDA presented the user with a nice and regular, fixed sampling rate, stream from the sensors, but this is essentially a lie: Sensor sampling rates may change all the time. Ask the users if they want the lie or an irregular but honest stream of if they want it configurable. Do not just use the last known sensor value in the "I want it regular" case but maybe consider resampling.

The audio classifier has some useful parameters like whitelist, blacklist, overlap, max results, score threshold. Expose them to the user!

We need user and developer documentation on

• how to set up the project in order to add features/fix bugs
• how to navigate the code
• how to use the app
• where to find releases

It would be neat if the documentation can be generated (partially) automatically from the code using Doxygen or something similar. I will investigate, whether that is feasible and also how we can add this step to the Ci/CD pipeline.

Measurement mode, Sampling rate, etc...

In some use cases the BT connection is sketchy. Recording on the sensor and streaming at the same time improves reliability.

@pmaanen Yes, let's use Github Issues!