[WIP] Implement new Level of Detail algorithm and tests

[Copilot]

Thanks for asking me to work on this. I will get started on it and keep this PR's description up to date as I form a plan and make progress.

Original prompt

The goal is to add a new Level of Detail (LOD) algorithm to this project, replace the current one's invocation in the GUI, and add tests for the new module.

1. New LOD Algorithm Implementation

A new LOD algorithm needs to be implemented with the following specifications:

• Input Data: The algorithm will process a series of mouse movement events. Each event is a struct: {dx: i16, dy: i16, tsc: u32, utsc: u32}. The tsc and utsc fields are timestamps. This data is used to plot dx vs. time and -dy vs. time.

• Data Segmentation:

  • The algorithm must analyze the events to identify and segment them into groups of "good events" and "discrete events".
  • Discrete/Not Good Events: An event is considered "not good" or "discrete" if it meets any of the following criteria:
    1. Both dx and dy are zero.
    2. It shows poor time consistency. The relationship between the event index and its timestamp should be linear (for a constant report rate) or fit a cubic polynomial well. Deviations indicate inconsistency.
    3. It is an outlier when fitting dx vs. time or dy vs. time to a cubic polynomial.
  • Segmentation Logic: The final output of this stage should be a List<(List<good events> or discrete event)>. The goal is to maximize the length of the List<good events> while ensuring a high coefficient of determination (R-squared, close to 1) for the polynomial fit. This may involve an iterative approach, starting with an initial number of events for regression and expanding the window (e.g., by a factor of 1.5 or 2) to find the optimal balance between segment length and fit quality.

• Result Caching: The results of the segmentation and regression analysis should be cached to improve performance on subsequent rendering calls.

2. LOD Plotting Function

Create a function that the plotting module will call to get the list of events to render.

• Function Signature & Parameters:

  • events: The source event data.
  • resolution: The rendering resolution of the plot.
  • x_range, y_range: The visible ranges of the plot.
  • tolerance: A threshold for the number of events that can map to a single screen coordinate before simplification is triggered.
  • zoom_factor: A factor (>1) to anticipate user zooming. The simplification should be calculated for a higher resolution (resolution * zoom_factor) to improve drag performance.

• Simplification Logic:

  • The function should leverage the cached segmentation results.
  • If a cached result can be used for the given zoom_factor and ranges, it should be reused.
  • It must calculate which events fall onto the same physical screen coordinates based on the effective resolution and plot ranges.
  • If the number of events at a coordinate exceeds the tolerance, some events must be hidden.
  • Constraint: The first and last events of any "good event list" segment must not be hidden to ensure continuity in the plot lines.

• Return Value: The function must return a list of indices corresponding to the original events array that are not hidden and should be drawn.

3. Integration

• In the GUI code, find the call site for the current LOD/plotting function.
• Comment out the call to the old algorithm.
• Add a new call to the function created in step 2.

4. Testing

• Create a new test module for the LOD algorithm.
• Write tests to verify its correctness.
• Use the test datasets located in the /examples/test/ directory as input for these tests.

This pull request was created as a result of the following prompt from Copilot chat.

The goal is to add a new Level of Detail (LOD) algorithm to this project, replace the current one's invocation in the GUI, and add tests for the new module.

1. New LOD Algorithm Implementation

A new LOD algorithm needs to be implemented with the following specifications:

• Input Data: The algorithm will process a series of mouse movement events. Each event is a struct: {dx: i16, dy: i16, tsc: u32, utsc: u32}. The tsc and utsc fields are timestamps. This data is used to plot dx vs. time and -dy vs. time.

• Data Segmentation:

  • The algorithm must analyze the events to identify and segment them into groups of "good events" and "discrete events".
  • Discrete/Not Good Events: An event is considered "not good" or "discrete" if it meets any of the following criteria:
    1. Both dx and dy are zero.
    2. It shows poor time consistency. The relationship between the event index and its timestamp should be linear (for a constant report rate) or fit a cubic polynomial well. Deviations indicate inconsistency.
    3. It is an outlier when fitting dx vs. time or dy vs. time to a cubic polynomial.
  • Segmentation Logic: The final output of this stage should be a List<(List<good events> or discrete event)>. The goal is to maximize the length of the List<good events> while ensuring a high coefficient of determination (R-squared, close to 1) for the polynomial fit. This may involve an iterative approach, starting with an initial number of events for regression and expanding the window (e.g., by a factor of 1.5 or 2) to find the optimal balance between segment length and fit quality.

• Result Caching: The results of the segmentation and regression analysis should be cached to improve performance on subsequent rendering calls.

2. LOD Plotting Function

Create a function that the plotting module will call to get the list of events to render.

• Function Signature & Parameters:

  • events: The source event data.
  • resolution: The rendering resolution of the plot.
  • x_range, y_range: The visible ranges of the plot.
  • tolerance: A threshold for the number of events that can map to a single screen coordinate before simplification is triggered.
  • zoom_factor: A factor (>1) to anticipate user zooming. The simplification should be calculated for a higher resolution (resolution * zoom_factor) to improve drag performance.

• Simplification Logic:

  • The function should leverage the cached segmentation results.
  • If a cached result can be used for the given zoom_factor and ranges, it should be reused.
  • It must calculate which events fall onto the same physical screen coordinates based on the effective resolution and plot ranges.
  • If the number of events at a coordinate exceeds the tolerance, some events must be hidden.
  • Constraint: The first and last events of any "good event list" segment must not be hidden to ensure continuity in the plot lines.

• Return Value: The function must return a list of indices corresponding to the original events array that are not hidden and should be drawn.

3. Integration

• In the GUI code, find the call site for the current LOD/plotting function.
• Comment out the call to the old algorithm.
• Add a new call to the function created in step 2.

4. Testing

• Create a new test module for the LOD algorithm.
• Write tests to verify its correctness.
• Use the test datasets located in the /examples/test/ directory as input for these tests.

---

✨ Let Copilot coding agent set things up for you — coding agent works faster and does higher quality work when set up for your repo.