Add control-subtracted LASSO behavioral prediction

[colehanan1]

Adds control-subtraction support in LASSO fits, CLI for ΔPER runs, tests, and doc update.

Summary by CodeRabbit

• New Features

  • Added control-subtracted (ΔPER) mode for LASSO behavioral predictions, enabling analysis of control-corrected behavioral responses with flexible missing-data handling policies
  • New command-line script for executing behavioral predictions with control-subtraction support

• Documentation

  • Updated with control-subtracted workflow examples and command-line usage patterns

• Tests

  • Added comprehensive test coverage for control-subtraction functionality

_{✏️ Tip: You can customize this high-level summary in your review settings.}

[coderabbitai]

Caution

Review failed

The pull request is closed.

📝 Walkthrough

Walkthrough

This PR introduces control-subtraction (ΔPER) functionality to LASSO-based behavioral prediction. It adds a new CLI script, extends the core behavioral prediction library with control-condition matching and missing-control policies, updates documentation with usage examples, and includes comprehensive tests for the new workflow.

Changes

Cohort / File(s)	Summary
Documentation docs/BEHAVIORAL_PREDICTION_ANALYSIS.md	Added new subsection "5.6 Control-Subtracted (ΔPER) Runs" with CLI examples for LASSO fitting on control-subtracted targets, including --subtract_control, --missing_control_policy, and optional control-condition overrides.
Core Library src/door_toolkit/pathways/behavioral_prediction.py	Extended LassoBehavioralPredictor.fit_behavior() with control-subtraction parameters (subtract_control, control_condition, missing_control_policy). Added BehaviorModelResults fields for tracking control metadata. Introduced helper methods for control-condition resolution and dataset normalization.
CLI Script scripts/run_lasso_behavioral_prediction.py	New script orchestrating LASSO behavioral prediction with control subtraction support. Handles condition parsing, missing-control fallback logic, per-condition output generation, and optional summary CSV export.
Tests tests/test_lasso_behavioral_prediction.py	Added control-subtraction test fixtures and new TestLassoBehavioralPredictorControlSubtraction suite validating control alignment, NaN handling, and missing-control error policies.

Sequence Diagram

sequenceDiagram
    participant User
    participant CLI as run_lasso_behavioral_prediction.py
    participant Parser as Argument Parser
    participant Predictor as LassoBehavioralPredictor
    participant DataLoader as Data Alignment
    participant LASSO as LASSO Fitting
    participant FileOutput as Output Generation

    User->>CLI: Execute with --subtract_control flags
    CLI->>Parser: Parse conditions, lambda range, policies
    Parser->>CLI: Return parsed arguments
    loop Per Condition
        CLI->>Predictor: fit_behavior(condition, subtract_control=True)
        Predictor->>Predictor: _infer_control_condition()
        Predictor->>DataLoader: Load opto & control behavior data
        DataLoader->>DataLoader: Align pairs by missing_control_policy
        DataLoader->>Predictor: Return ΔPER (opto − control)
        Predictor->>LASSO: Fit model on ΔPER targets
        LASSO->>Predictor: Return BehaviorModelResults
        Predictor->>CLI: Return results with metadata
        CLI->>FileOutput: Save plots, CSV, JSON (with mode prefix)
    end
    FileOutput->>User: Generated outputs per condition

Loading

Estimated code review effort

🎯 4 (Complex) | ⏱️ ~50 minutes

Possibly related PRs

• colehanan1/door-python-toolkit#1: Introduces new subtract_control, control_condition, and missing_control_policy parameters to LassoBehavioralPredictor.fit_behavior() that the CLI script directly depends on.

Poem

🐰 Control-subtracted deltas hop and bound,
ΔPER whispers calculations profound,
When opto meets control in paired embrace,
The neural predictions find their rightful place! ✨

✨ Finishing touches

• 📝 Generate docstrings

---

📜 Recent review details

Configuration used: defaults

Review profile: CHILL

Plan: Pro

📥 Commits

Reviewing files that changed from the base of the PR and between 3225982 and 64c79ce.

📒 Files selected for processing (4)

• docs/BEHAVIORAL_PREDICTION_ANALYSIS.md
• scripts/run_lasso_behavioral_prediction.py
• src/door_toolkit/pathways/behavioral_prediction.py
• tests/test_lasso_behavioral_prediction.py

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}