opto analysis

Opto analysis is a package for analyzing free-moving behavioral data during optogenetics experiments. It is currently optimized for and tested on Windows terminals with debugging in VS Code, but it can be extended to other systems upon request. It uses behavioral videos, additional data streams and a trained DeepLabCut tracking network and it outputs visualizations and statistics of navigational trajectories.

Installation

• Clone the repository

• Get the sample data at "...Dropbox (UCL)\DAQ\upstairs_rig\21MAR16_9718_block evs"; place this folder in the sample_data folder

• Navigate to the opto-analysis folder and run the following commands:

  • Users: Create a new Python 3.8 environment (default name "opto-user") with conda env create -f environment_user.yml

  • Devs: Create a new Python 3.8 environment (default name "opto") with conda env create -f environment_dev.yml

  • Both: Install the package with the command pip install -e .

Usage

To process data, verify data synchronization, and register videos, fill in the the file ./settings/settings_process.py with your desired settings, and run the terminal command process

---

To track data (DeepLabCut), fill in the the file ./settings/settings_track.py with your desired settings, and run the command track

---

To visualize trials, fill in the the file ./settings/settings_visualize.py with your desired settings, and run the command visualize. This requires a trained DLC network.

---

To analyze and plot data, fill in the ./settings/analyses with your desired analysis program options and then fill in ./settings/settings_analyze.py with global analysis settings (including the program you'd like to run), and run the command analyze

Testing (devs)

• In VS code: run test discovery and then run or debug the desired tests in the Test Explorer. Run each test in ./tests/test_run.py to debug an entire workflow (processing, tracking, visualization, analysis)

• Alternatively, run in the terminal: pytest or pytest tests/test_file.py

• To examine the test coverage, run: pytest --cov-report term-missing --cov=opto_analysis tests/

• To instead generate a coverage report, run: pytest --cov-report xml --cov=opto_analysis tests/

• ...and then run in bash shell: bash <(curl -s https://codecov.io/bash) -t TOKEN-NUMBER to upload to codecov

Dependency management

• To remake the environment file, run: conda env export > dev_environment.yml