Dataset Module (stx.dataset)
Unified access to public scientific datasets from major neuroscience and biomedical repositories. Search, browse, and fetch datasets without learning each repository’s API.
Supported Sources
Source |
Description |
|---|---|
Neurophysiology data (NWB format) – electrophysiology, calcium imaging |
|
Brain imaging data (BIDS format) – fMRI, EEG, MEG, PET |
|
Physiological signal data – ECG, EEG, clinical waveforms |
Quick Start
import scitex as stx
# Search across all sources
results = stx.dataset.search("epilepsy EEG")
# Fetch a specific dataset
data = stx.dataset.fetch("dandi", dandiset_id="000003")
# List available sources
sources = stx.dataset.list_sources()
# ['dandi', 'openneuro', 'physionet']
Key Functions
search(query, source=None)
Search for datasets across one or all supported repositories.
# Search all sources
results = stx.dataset.search("motor imagery BCI")
# Search a specific source
results = stx.dataset.search("sleep staging", source="physionet")
for r in results:
print(f"{r.source}: {r.title} ({r.id})")
fetch(source, **kwargs)
Download or stream a dataset from a specific source.
# Fetch from DANDI
data = stx.dataset.fetch("dandi", dandiset_id="000003")
# Fetch from OpenNeuro
data = stx.dataset.fetch("openneuro", dataset_id="ds000117")
# Fetch from PhysioNet
data = stx.dataset.fetch("physionet", database="mimic-iv", version="2.0")
list_sources()
List all available dataset sources and their status.
sources = stx.dataset.list_sources()
# ['dandi', 'openneuro', 'physionet']
Local Database
Build a local search index for faster repeated queries:
# Build/update the local database
stx.dataset.db_build(sources=["dandi", "openneuro"])
# Search the local index (fast, offline)
results = stx.dataset.db_search("resting state fMRI")
# Get database statistics
stx.dataset.db_stats()
CLI Access
# Search datasets
scitex dataset search "epilepsy EEG"
# Fetch a dataset
scitex dataset fetch dandi --dandiset-id 000003
# List sources
scitex dataset list-sources
# Build local database
scitex dataset db-build