Skip to main content
Version: 0.93

snorkelflow.ingest

Functionality to ingest data of various formats into Snorkel Flow.

Reference

Functions

conversation_json_to_parquet(...)

Generates SnorkelFlow ingestible PARQUET file from JSON.

dirtree_to_parquet(dir_root[, native, ...])

Generate a partitioned parquet file from a directory of PDF documents.

import_utterance_ground_truth(node, ...[, ...])

Uploads ground truth for utterance classification applications.

time_series_csv_to_parquet(...[, label_col, ...])

Generate SnorkelFlow-ingestible PARQUET file from CSV.