data_base ❭ db_initializers ❭ prepare_ANN_batches

prepare_ANN_batches¶

Parse a simrun-initialized database for ANN batch preparation.

This module provides various methods to bin synapse actications, morphologies, and voltage traces. This binned data can be used as input data for training an artificial neural network (ANN).

Functions¶

`get_binsize`(length, binsize_goal)	Calculate the bin size and number of bins closest to the desired bin size goal.
`get_bin`(value, bin_min, bin_max, n_bins, tolerance)	Find the bin index of a given value.
`get_neuron_param_file`(db)	Get the path to the Cell parameters file from a simrun-initialized database.
`get_section_distances_df`(neuron_param_file, silent)	Bin the morphology of a neuron from a Cell parameters file.
`get_spatial_bin_names`(section_distances_df)	Get the bin names from a dataframe describing the distance to soma of all sections.
`get_bin_soma_distances_in_section`(section_id, section_distances_df)	Get the distance to the soma for all bins in this section.
`get_bin_adjacency_map_in_section`(cell, section_id, section_distances_df)	Create an adjacency map with bin-specific resolution for a given section.
`get_neighboring_spatial_bins`(cell, bin_id, section_distances_df)	Get all the neighboring bins from a `bin_id`.
`augment_synapse_activation_df_with_branch_bin`(sa_, section_distances_df, synaptic_weight_dict, excitatory_celltypes)	Augment a Synapse activation dataframe with bin information.
`spike_times_to_onehot`(spike_times, min_time, max_time, time_step)	One-hot encode spike times to a binned time vector.
`compute_ISI_from_st_list`(st, min_time, max_time, time_step)	Construct a time-binned list of inter-spike intervals from a list of spike times.
`compute_ISI_from_st`(st, timepoint, fillna)	Calculate the time since the last spike in ms for each element in a spike time pd.Series or pd.DataFrame.
`compute_ISI_array`(st, min_time, max_time, fillna, step)	Calculate the time since the last spike in ms for each element in a spike time Array.
`compute_AP_array`(st, min_time, max_time, fillna, step)	One-hot encode spike times to a binned time vector.
`load_syn_weights`(db, client, excitatory_celltypes)	Load synapse weights from a simrun-initialized database
`temporal_binning_augmented_sa`(sa_augmented, min_time, max_time, bin_size, use_weights)	Bin synapse activation times into temporal bins.
`get_synapse_activation_array_weighted`(sa_, selected_stis, spatial_bin_names, min_time, max_time, bin_size, use_weights)	Create a 4D array of synapse activation times.
`save_SA_batch`(sa_, selected_stis, batch_id, outdir, section_distances_df, spatial_bin_names, min_time, max_time, bin_size, synaptic_weight_dict)	Save a batch of synapse activation times to a file.
`get_time_groups`(vt_pandas)	Get the time groups for binning voltage traces.
`get_max_per_ms_on_pandas_dataframe`(vt_pandas)	Subsample a voltage trace pandas dataframe to 1ms bins based on the maximum depolarization.
`get_max_depolarization_per_ms`(vt_dask)	Subsample a voltage trace dask dataframe to 1ms bins based on the maximum depolarization.
`run_delayeds_incrementally`(client, delayeds)	Convenience method to run a list of dask delayed objects incrementally.