mirror of
https://github.com/huggingface/lerobot.git
synced 2026-05-24 04:59:47 +00:00
removed unused functions from tests/fixtures
This commit is contained in:
Vendored
-52
@@ -17,8 +17,6 @@ from pathlib import Path
|
|||||||
import datasets
|
import datasets
|
||||||
import numpy as np
|
import numpy as np
|
||||||
import pandas as pd
|
import pandas as pd
|
||||||
import pyarrow.compute as pc
|
|
||||||
import pyarrow.parquet as pq
|
|
||||||
import pytest
|
import pytest
|
||||||
from datasets import Dataset
|
from datasets import Dataset
|
||||||
|
|
||||||
@@ -186,53 +184,3 @@ def create_hf_dataset(hf_dataset_factory):
|
|||||||
write_hf_dataset(hf_dataset, dir, data_file_size_in_mb, chunk_size)
|
write_hf_dataset(hf_dataset, dir, data_file_size_in_mb, chunk_size)
|
||||||
|
|
||||||
return _create_hf_dataset
|
return _create_hf_dataset
|
||||||
|
|
||||||
|
|
||||||
@pytest.fixture(scope="session")
|
|
||||||
def single_episode_parquet_path(hf_dataset_factory, info_factory):
|
|
||||||
def _create_single_episode_parquet(
|
|
||||||
dir: Path, ep_idx: int = 0, hf_dataset: datasets.Dataset | None = None, info: dict | None = None
|
|
||||||
) -> Path:
|
|
||||||
raise NotImplementedError()
|
|
||||||
if info is None:
|
|
||||||
info = info_factory()
|
|
||||||
if hf_dataset is None:
|
|
||||||
hf_dataset = hf_dataset_factory()
|
|
||||||
|
|
||||||
data_path = info["data_path"]
|
|
||||||
chunks_size = info["chunks_size"]
|
|
||||||
ep_chunk = ep_idx // chunks_size
|
|
||||||
fpath = dir / data_path.format(episode_chunk=ep_chunk, episode_index=ep_idx)
|
|
||||||
fpath.parent.mkdir(parents=True, exist_ok=True)
|
|
||||||
table = hf_dataset.data.table
|
|
||||||
ep_table = table.filter(pc.equal(table["episode_index"], ep_idx))
|
|
||||||
pq.write_table(ep_table, fpath)
|
|
||||||
return fpath
|
|
||||||
|
|
||||||
return _create_single_episode_parquet
|
|
||||||
|
|
||||||
|
|
||||||
@pytest.fixture(scope="session")
|
|
||||||
def multi_episode_parquet_path(hf_dataset_factory, info_factory):
|
|
||||||
def _create_multi_episode_parquet(
|
|
||||||
dir: Path, hf_dataset: datasets.Dataset | None = None, info: dict | None = None
|
|
||||||
) -> Path:
|
|
||||||
raise NotImplementedError()
|
|
||||||
if info is None:
|
|
||||||
info = info_factory()
|
|
||||||
if hf_dataset is None:
|
|
||||||
hf_dataset = hf_dataset_factory()
|
|
||||||
|
|
||||||
data_path = info["data_path"]
|
|
||||||
chunks_size = info["chunks_size"]
|
|
||||||
total_episodes = info["total_episodes"]
|
|
||||||
for ep_idx in range(total_episodes):
|
|
||||||
ep_chunk = ep_idx // chunks_size
|
|
||||||
fpath = dir / data_path.format(episode_chunk=ep_chunk, episode_index=ep_idx)
|
|
||||||
fpath.parent.mkdir(parents=True, exist_ok=True)
|
|
||||||
table = hf_dataset.data.table
|
|
||||||
ep_table = table.filter(pc.equal(table["episode_index"], ep_idx))
|
|
||||||
pq.write_table(ep_table, fpath)
|
|
||||||
return dir / "data"
|
|
||||||
|
|
||||||
return _create_multi_episode_parquet
|
|
||||||
|
|||||||
Reference in New Issue
Block a user