### Install all requirements
**Recommended**: start a new conda environment

- Install dependencies in requirement.txt
	- At the root level of the repository, use `pip install -r requirements.txt`
- `pip install -e .` to install the roboviz package
- You will also need to install lerobot. Use the installation guide in the [lerobot repo](https://github.com/huggingface/lerobot)


## Running the scripts on recorded data

#### Recording.py configuration
You will find the file `Recording.py` in `roboviz/omar/Recoding.py`
You will need to change the configuration in `Recording.py`.
The configuration looks something like this
```Python
input_dir   = "path/to/your/raw_data"             # contains demo_0000, demo_0001, ...

output_dir  = "path/to/output/LeRobot_dataset"

script_dir = "path/to/script/dir"

DATASET_NAME = "Dataset Name"

S3_ENDPOINT = "https://s3.kopah.uw.edu"

fps         = 20.0                                # robot’s recording rate

chunk_size  = 1000                                # episodes per chunk folder

cameras     = detect_cameras(input_dir)

DEPTH_MAX   = 1.0                                 # depth clip at 1 m
```
Replace `input_dir` to the directory of your recorded data. This directory should contain demo_000, demo_0001, ...

Replace `output_dir` with the directory you wish to place the generated Lerobot format dataset in.

Replace `script_dir` with the location of the roboviz scripts. This is usually found in `roboviz/app/algos`

Replace `DATASET_NAME` with your dataset name

If you wish to upload the dataset to your own bucket in S3, you will need to change `S3_BUCKET` and `S3_PREFIX`

Replace `KOPAH_CREDS` with the path to your json kopah credentials. An example credential file looks like the following
```json
{
  "access_key": "YOUR ACCESS KEY",
  "secret_key": "YOUR SECRET KEY"
}
```
### Generating LeRobot dataset and running scripts
You can run the `Recoding.py` with two flags
- `--run` will run the analysis script after the data is converted to the LeRobot format
- `--upload` will upload the converted LeRobot dataset and analysis data (if any) to the specified S3 bucket

## Running the visualizer
Once you have generated the visualization by (by passing in `--run` flag when running Recording.py), you can now use our visualizer to analyze your data

You can self host the visualizer by running by calling `python roboviz/app/app.py` located in 

This will start a flask web server at `localhost port 5000`

You can use the web visualizer to explore the plots generated by our scripts
You can click on the different buttons on the left panel of the visualizer to view different visualizations