Releases

First stable release of csv-to-json, a dependency-free CSV-to-JSON converter for developer workflows.

What it does

Reads CSV from a file path, a raw string, or a text stream (including stdin) and emits JSON — either as a string, to stdout, or written to a file.

Zero dependencies — Python 3.10+ standard library only.
Type inference — empty cells become null; true/yes/false/no become booleans; numeric strings become int or float; everything else stays a string.
Nested-key expansion — dot-notation headers like user.address.city expand into nested JSON objects (opt-in via --expand-keys).
Streaming — stream_csv_to_json() yields one record at a time so large files never load fully into memory.
Flexible delimiters — comma by default, override with -d (e.g. -d $'\t' for TSV).
CLI + library — usable from the shell or importable as from csv_converter import csv_to_json.

python src/csv_converter.py data.csv -o data.json --expand-keys
cat data.tsv | python src/csv_converter.py - -d $'\t' --compact

from csv_converter import csv_to_json, stream_csv_to_json
json_str = csv_to_json("data.csv", expand_keys=True)