Wrong paths

[sparisi]

Habitat-Lab and Habitat-Sim versions

Habitat-Lab: hab_suite

Habitat-Sim: master

I am trying to run the pick tasks, but the YAML file points to a missing file. The file is in the DATA_PATH JSON

data/datasets/rearrange_pick/replica_cad/v0/rearrange_pick_replica_cad_v0/pick.json.gz

When I launch my experiments, the dataset is automatically downloaded by habitat, and the only JSON there is

data/datasets/rearrange_pick/replica_cad/v0/train/train_counter_L_analysis_5000_500.json.gz

However, this JSON has some fields that are not recognized by the Episode class

  File "/private/home/sparisi/habitat-baselines/habitat-lab/habitat/datasets/rearrange/rearrange_dataset.py", line 70, in from_json
    rearrangement_episode = RearrangeEpisode(**episode)
TypeError: __init__() got an unexpected keyword argument 'scene_config_path'

Not sure if I am doing something wrong, or the branch is still in dev and so the paths are broken.

[ASzot]

The most recent version of Habitat Sim should download the correct files. Can you install the most recent version of the Habitat Sim main branch? And then run the following to force download the assets:

python -m habitat_sim.utils.datasets_download --uids rearrange_task_assets

I just installed from scratch and it worked fine for me.

[sparisi]

I installed the latest with conda install -c aihabitat-nightly habitat-sim and ran your code. It downloaded files that were missing. However, I still get

  File "/private/home/sparisi/habitat-baselines/habitat-lab/habitat_baselines/utils/env_utils.py", line 28, in make_env_fn
    dataset = make_dataset(
  File "/private/home/sparisi/habitat-baselines/habitat-lab/habitat/datasets/registration.py", line 21, in make_dataset
    return _dataset(**kwargs)  # type: ignore
  File "/private/home/sparisi/habitat-baselines/habitat-lab/habitat/datasets/rearrange/rearrange_dataset.py", line 62, in __init__
    super().__init__(config)
  File "/private/home/sparisi/habitat-baselines/habitat-lab/habitat/datasets/pointnav/pointnav_dataset.py", line 93, in __init__
    with gzip.open(datasetfile_path, "rt") as f:
  File "/private/home/sparisi/.conda/envs/hab/lib/python3.8/gzip.py", line 58, in open
    binary_file = GzipFile(filename, gz_mode, compresslevel)
  File "/private/home/sparisi/.conda/envs/hab/lib/python3.8/gzip.py", line 173, in __init__
    fileobj = self.myfileobj = builtins.open(filename, mode or 'rb')
FileNotFoundError: [Errno 2] No such file or directory: 'data/datasets/rearrange_pick/replica_cad/v0/rearrange_pick_replica_cad_v0/pick.json.gz'

I would rather not to reinstall it from scratch, since I am currently using Habitat for other experiments and they are running fine (and in the past I had trouble setting them up, so I do not want to risk breaking my working code).

The task I am trying to run uses

config_file = 'habitat_baselines/config/rearrange/ddppo_pick.yaml'
base_config_file = 'configs/tasks/rearrange/pick.yaml'

[lyl1140292193]

Maybe you could look at this and add the keys in RearrangeEpisode. https://github.com/3dlg-hcvc/multion-challenge/blob/0b771eef0c9fb638b488e89d1729a78f5b8beca2/habitat/datasets/rearrange/rearrange_dataset.py#L30