IO Foundations

Overview

Interchange helpers in quantsmind.io: JSON documents, CSV tables, and streaming readers on the standard library, plus SDK-object boundaries for anything exposing to_dict / from_dict.

Purpose

Let users persist and reload SDK data deterministically with explicit, typed errors instead of bare I/O exceptions.

Concept

read_json/write_json for documents, read_csv/write_csv for header-checked tables, iter_lines/iter_chunks for bounded streams; dump_object/load_object bridge SDK objects; JsonError/CsvError wrap every failure mode.

API

read_json(), write_json(), dump_object(), load_object(), read_csv(), write_csv(), iter_lines(), iter_chunks(), JsonError, CsvError.

Input / Processing / Output

Input: paths and JSON-safe data. Processing: UTF-8 encode/decode with validation. Output: restored documents, tables, streams.

Example

python examples/io/roundtrip.py round-trips all three formats through a temporary directory.

Limitations

Local filesystem only; no network transports, no Parquet/Arrow, no schema evolution — boundaries are explicit, not magical.