v0.2.1

Patch release.

• Requires pointblank>=1.1.0. The two issues we reported upstream are fixed there: unique: true is now enforced on coherent presets such as name, email and city (posit-dev/pointblank#441), and get_dataframe_report() works with multi-column steps (posit-dev/pointblank#442).
• The "known limit" about unique on presets is removed from the fauxdata agent guide.
• Fewer dependencies: pointblank 1.1 no longer pulls in anthropic[bedrock], so about 20 transitive packages (anthropic, boto3, pydantic, ...) are no longer installed.
• The source distribution now ships only src, tests, schemas, README, LICENSE and pyproject.toml.
• README: links to the bundled example schemas.

v0.2.0

Highlights

• Agent-friendly CLI: new fauxdata agent command prints a Markdown guide for AI agents (workflow, schema format, types, presets, rules, JSON output, exit codes, example). generate, validate, infer and preview accept --json: stdout carries a single JSON object, errors come back as {"error": "..."} with exit code 1.
• pointblank 1.0.1: minimum version raised from 0.25 (also brings Polars 2.0 compatibility).
• New validation rule col_vals_str_len (column, min and/or max). infer emits it for free-text string columns.
• Native float precision: rounding never pushes values outside min/max (bounds are moved inward to the precision grid) and unique columns stay unique after rounding. Clear error when no value fits the range.

Fixes

• infer --categorical-threshold from the CLI was read as a fraction of rows, so every non-preset string column became categorical and its real values were copied into values:. Whole numbers are now counts.
• Value rules (col_vals_between, gt/lt/ge/le, regex, in_set) no longer count nulls as failures; use col_vals_not_null to forbid nulls.
• values on int/float columns is now honored by the generator (previously ignored, breaking infer round-trips).
• rows_distinct no longer treats null keys as duplicates of each other.
• A validation step on a column missing from the data is now reported as not evaluated ("evaluated": false, NOT RUN in the table) instead of passing with 0 rows.

Behavior changes

• Value rules ignore nulls.
• rows_distinct totals count only rows with non-null keys.
• --categorical-threshold N with a whole number is a count of distinct values (1 means one value, not 100%).
• Float columns with precision and bounds that have more decimals than precision generate slightly narrower ranges than 0.1.6.
• pointblank now requires anthropic[bedrock], so installing fauxdata pulls in about 20 extra packages that fauxdata does not use.

Known limitations (upstream)

• unique: true is not enforced on coherent presets such as email, name, city (posit-dev/pointblank#441).

Closes #1.

v0.1.6

Fix

• float precision applied at generation time (#2): float columns with a precision set are now rounded to the requested number of decimals (pointblank has no native parameter → rounding is done right after generation, preserving min/max/unique/nullable handling).

Notes

• Upstream feature request for native support: posit-dev/pointblank#410
• PyPI: https://pypi.org/project/fauxdata-cli/0.1.6/

v0.1.5

What's new

• fauxdata infer DATASET: build a YAML schema from a real table (csv/parquet/json/jsonl) via pointblank schema_from_tbl — inferred ranges, categorical value sets, string presets (email/url/…), uniqueness, null rates, and string lengths, plus matching validation: rules. Feed the result to fauxdata generate for synthetic data that mirrors the real shape without shipping real values. Flags: --out/-o, --name/-n, --rows/-r, --format/-f, --categorical-threshold, --detect-presets/--no-detect-presets, --sample-size.
• String length: new min_length/max_length schema fields (faithful round-trip), passed through to generation.
• infer on CSV uses try_parse_dates so date/datetime columns are typed, not frozen as categorical strings.
• generate raises a clear error on the unique + narrow-range conflict (amplifying a unique integer column beyond its range) instead of pointblank's cryptic message.

Dependencies

• Requires pointblank>=0.25.

Known limits

• Low-cardinality columns are frozen to their exact source values (values:) — raise --categorical-threshold for sensitive columns.
• A column inferred as unique over a narrow integer range cannot be amplified far beyond the source size.

Breaking changes

• None.

Installation

uv tool install fauxdata-cli==0.1.5

Full changelog: https://github.com/aborruso/fauxdata/commits/v0.1.5

v0.1.4

What's new

• init non-interactive: --description, --rows, --format, --yes flags; questionary prompts are now fallback when flags are omitted
• generate --dry-run: preview what would be generated (schema, rows, seed, output path) without writing any file
• Structured output on success: generate prints output_path, format, rows as key: value; init prints schema_path
• Examples in --help: every subcommand has an epilog with 2–4 concrete examples

Breaking changes

None.

Installation

```bash
pip install fauxdata-cli==0.1.4
```

Full changelog: https://github.com/aborruso/fauxdata/commits/v0.1.4

v0.1.3 — Python version badge fix

What's new

Bug fix

• The Python Versions badge (shields.io/pypi/pyversions) was blank because pyproject.toml lacked classifiers. Added:
  • Programming Language :: Python :: 3.11
  • Programming Language :: Python :: 3.12
  • Programming Language :: Python :: 3.13
  • License :: OSI Approved :: MIT License

Internal

• Version tests now read __version__ dynamically — no more hardcoded version strings in the test suite

Installation

pip install fauxdata-cli==0.1.3

Full changelog: https://github.com/aborruso/fauxdata/commits/v0.1.3

v0.1.2 — Tests, pattern generation, null_probability, --version

What's new

Test suite

• 79 tests, 84% coverage, threshold enforced at 80% via pytest-cov
• Covers schema parsing, generation, validation, output formats, and CLI end-to-end
• Run with uv run pytest

New schema fields

• pattern (string columns): generate values matching a regex via pointblank

  code:
    type: string
    pattern: "[A-Z]{2}[0-9]{4}"

• null_probability: fractional null control, replaces the binary nullable: true

  score:
    type: float
    null_probability: 0.15   # ~15% of values will be null

  Accepts values in [0.0, 1.0] — validated at schema parse time.

CLI

• fauxdata --version / fauxdata -V now works

Internals

• Removed unused faker dependency (pointblank handles everything)
• Added docs/deployment.md with the release workflow

Installation

pip install fauxdata-cli==0.1.2
# or
uv add fauxdata-cli==0.1.2

Full changelog: https://github.com/aborruso/fauxdata/commits/v0.1.2