# install cell skipped during CI (deps preinstalled into system Python)

# Suppress hdmf warnings about cached namespaces
import warnings
warnings.filterwarnings("ignore", category=UserWarning, module="hdmf.utils")

from helpers.stream_nwbfile import stream_nwbfile
import dandi.dandiarchive as da
from IPython.display import display, Markdown


DANDISET_ID = "000458"

parsed_url = da.parse_dandi_url(f"https://dandiarchive.org/dandiset/{DANDISET_ID}")

with parsed_url.navigate() as (client, dandiset, assets):
    if dandiset is None:
        raise Exception(f"Dandiset {DANDISET_ID} not found.")

    num_consecutive_not_nwb = 0
    asset_index = 0
    for asset_obj in dandiset.get_assets('path'):
        if not asset_obj.path.endswith(".nwb"):
            num_consecutive_not_nwb += 1
            if num_consecutive_not_nwb >= 20:
                # For example, this is important for 000026 because there are so many non-nwb assets
                raise Exception("Too many consecutive non-NWB files.")
            continue
        else:
            num_consecutive_not_nwb = 0
        asset = {
            "identifier": asset_obj.identifier,
            "path": asset_obj.path,
            "size": asset_obj.size,
            "download_url": asset_obj.download_url,
            "dandiset_id": DANDISET_ID,
        }
        nwbfile, io = stream_nwbfile(DANDISET_ID, asset["path"])
        md = ''
        ns_url = f'https://neurosift.app/?p=/nwb&url={asset["download_url"]}&dandisetId={DANDISET_ID}'
        md += f'**Asset {asset_index + 1}: [{asset["path"]}]({ns_url})**\n\n'
        experiment_description = nwbfile.experiment_description  # type: ignore
        session_description = nwbfile.session_description  # type: ignore
        stimulus_notes = nwbfile.stimulus_notes  # type: ignore
        md += experiment_description + '\n\n'
        md += session_description + '\n\n'
        md += stimulus_notes + '\n\n'
        trials = nwbfile.intervals['trials'].to_dataframe()  # type: ignore
        num_trials = len(trials)
        md += f'* Size: {asset["size"] / 1e6:.2f} MB\n'
        md += f"* {num_trials} trials\n"
        behavioral_epochs = trials['behavioral_epoch']
        is_valid = trials['is_valid']
        is_running = trials['is_running']
        unique_behavioral_epochs = behavioral_epochs.unique()
        md += f"* BEHAVIORAL EPOCHS\n"
        for be in unique_behavioral_epochs:
            num = len(behavioral_epochs[behavioral_epochs == be])
            num_valid = len(is_valid[(behavioral_epochs == be) & is_valid])
            num_valid_and_not_running = len(is_valid[(behavioral_epochs == be) & is_valid & ~is_running])
            md += f"  - {be}: {num_valid_and_not_running} valid and not running\n"
        stimulus_types = trials['stimulus_type']
        unique_stimulus_types = stimulus_types.unique()
        md += f"* STIMULUS TYPES\n"
        for st in unique_stimulus_types:
            num = len(stimulus_types[stimulus_types == st])
            num_valid = len(is_valid[(stimulus_types == st) & is_valid])
            num_valid_and_not_running = len(is_valid[(stimulus_types == st) & is_valid & ~is_running])
            md += f"  - {st}: {num_valid_and_not_running} valid and not running\n"
        stimulus_descriptions = trials['stimulus_description']
        unique_stimulus_descriptions = stimulus_descriptions.unique()
        md += f"* STIMULUS DESCRIPTIONS\n"
        for sd in unique_stimulus_descriptions:
            num = len(stimulus_descriptions[stimulus_descriptions == sd])
            num_valid = len(is_valid[(stimulus_descriptions == sd) & is_valid])
            num_valid_and_not_running = len(is_valid[(stimulus_descriptions == sd) & is_valid & ~is_running])
            md += f"  - {sd}: {num_valid_and_not_running} valid and not running\n"
        estim_target_regions = trials['estim_target_region']
        unique_estim_target_regions = estim_target_regions.unique()
        md += f"* ESTIM TARGET REGIONS\n"
        for etr in unique_estim_target_regions:
            num = len(estim_target_regions[estim_target_regions == etr])
            num_valid = len(is_valid[(estim_target_regions == etr) & is_valid])
            num_valid_and_not_running = len(is_valid[(estim_target_regions == etr) & is_valid & ~is_running])
            md += f"  - {etr}: {num_valid_and_not_running} valid and not running\n"
        display(Markdown(md))
        asset_index += 1

Installing requirements¶

000458 - summarize contents¶