tap-jsonfile

Singer tap that reads JSON files from local filesystem or S3-compatible storage (including MinIO).

Built with the Meltano Singer SDK.

Features

• Glob patterns to match files across directories (data/**/*.json)
• S3 and MinIO support via fsspec / s3fs
• Automatic schema inference by sampling the first N files
• JSON, JSON arrays, and JSONL formats detected automatically
• Incremental sync: tracks file content hashes in Singer state, skips unchanged files on subsequent runs
• Adds _sdc_source_file to every record for lineage

Configuration

S3 / MinIO credentials

Set these environment variables (standard AWS naming, also accepts S3_* prefix):

Usage

Standalone

# Local files
tap-jsonfile --config '{"paths": ["data/**/*.json"]}' > output.jsonl

# S3 / MinIO (credentials via env vars)
export AWS_ACCESS_KEY_ID=...
export AWS_SECRET_ACCESS_KEY=...
export AWS_ENDPOINT_URL=https://minio.example.com
tap-jsonfile --config '{"paths": ["s3://bucket/prefix/**/*.json"]}'

Incremental sync

Pass state from a previous run to skip unchanged files:

tap-jsonfile --config config.json > output.jsonl 2>/dev/null

# Extract state for next run
grep '"type":"STATE"' output.jsonl | tail -1 | python3 -c \
  "import sys,json; print(json.dumps(json.loads(sys.stdin.read())['value']))" > state.json

# Next run skips files whose content hash hasn't changed
tap-jsonfile --config config.json --state state.json

With Meltano

uv tool install meltano
meltano install
meltano run tap-jsonfile target-jsonl

See meltano.yml for the default configuration (paths: ["data/**/*.json"]).

Development

Prerequisites: Python 3.10+, uv

uv sync
uv run pytest          # 41 tests
uv run tap-jsonfile --about