dataprof 0.12.0

High-performance data profiler with ISO 8000/25012 quality metrics for CSV, JSON/JSONL, and Parquet files
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33

# Line endings
#
# Every text file is stored and checked out as LF, whatever a contributor's
# `core.autocrlf` happens to be. Git on Windows defaults that setting to
# `true`, which stores LF but checks CRLF out; the working tree then disagrees
# with the repository and the damage is quiet rather than loud:
#
#   * `cargo fmt --check` and `ruff format --check` flag files nobody touched
#   * scripts that split on "\n" stop matching, for no visible reason
#   * generated files (docs/schema/) show as modified with an empty diff
#
# Every text file here is already stored as LF, so this governs how files are
# checked out, not what is committed. Adding it produces no content change.
* text=auto eol=lf

# Binary files. `text=auto` detects these on content alone, but naming them
# keeps a misdetection from corrupting an asset or a test fixture — a Parquet
# file with its bytes "converted" would still open and profile wrongly.
*.png binary
*.jpg binary
*.webp binary
*.parquet binary

# Build artefacts, should any ever be committed as a fixture.
*.whl binary
*.so binary
*.pyd binary
*.dll binary

# Fuzz seeds are exact byte sequences, including malformed UTF-8 and CRLF.
fuzz/seeds/** -text
fuzz/seeds/**/*.bin binary
fuzz/seeds/csv/quoted-record.csv whitespace=cr-at-eol