fix: auto-detect CSV delimiter via Sniffer in import_set and detect()

[terminalchai]

Fixes #622

Problem

tablib.Dataset().load(data, format='csv') always uses , as the delimiter, even when the file uses ;, :, |, or other separators. This is because import_set calls kwargs.setdefault('delimiter', cls.DEFAULT_DELIMITER) without ever sniffing the actual content.

A secondary issue: CSVFormat.detect() passed delimiters=delimiter or cls.DEFAULT_DELIMITER to csv.Sniffer().sniff(), which restricted sniffing to comma-only. This meant non-comma CSV files were never recognised during format auto-detection.

Fix

import_set: when no delimiter kwarg is provided, read up to 2048 bytes as a sample, call csv.Sniffer().sniff(sample) with no delimiter restriction, and use the detected delimiter. The stream is then seeked back to 0 before the actual read. A guard rejects alphabetic/digit detections (the Sniffer occasionally misidentifies a letter on very short or ambiguous samples). Falls back to DEFAULT_DELIMITER on any csv.Error.

detect(): when no explicit delimiter is given, uses a candidate string of common non-tab separators (',;:|') for CSVFormat, and the format's DEFAULT_DELIMITER for subclasses (e.g. TSVFormat uses tab). This keeps tab-delimited files out of CSV auto-detection while still recognising ;, :, and | separated files.

Explicit delimiter kwargs are fully respected and take precedence.

Tests

• Updated test_csv_formatter_support_kwargs: old assertion documented the broken behaviour (1 header instead of 3); updated to assert correct parsing.
• Added test_csv_import_set_auto_detect_delimiter: colon- and pipe-delimited CSV auto-detection.
• Added test_csv_detect_non_comma_delimiter: detect() correctly recognises :, |, and ; separated files.

All 155 tests pass.

[codecov]

Codecov Report

❌ Patch coverage is 96.96970% with 1 line in your changes missing coverage. Please review.
✅ Project coverage is 93.18%. Comparing base (564619d) to head (e8c3477).

Files with missing lines	Patch %	Lines
src/tablib/formats/_csv.py	93.75%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #639      +/-   ##
==========================================
+ Coverage   93.14%   93.18%   +0.03%     
==========================================
  Files          29       29              
  Lines        3226     3256      +30     
==========================================
+ Hits         3005     3034      +29     
- Misses        221      222       +1     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.