Make adversarial generator seedable and random-looking

[Copilot]

The adversarial generator always produced identical, hardcoded pairs (fixed module names, parameter names, types), making outputs recognizable and non-reproducible-on-demand. Now every pair is randomized but fully reproducible via an optional seed parameter.

Changes

adversarial_generator.py

• Added name pools: _MODULES, _FUNC_NAMES, _PARAM_NAMES, _SCALAR_TYPES, _MISLEADING_KWONLY_NAMES, _DECORATOR_NAMES, _VERSION_SUFFIXES
• Added _rand_fqname(rng) and _rand_params(rng, n) helpers
• All 10 _strategy_* functions now accept a random.Random instance — module names, function names, parameter names, types, and decorators are all drawn from pools via the RNG
• generate() and generate_all() accept a seed keyword argument; a random.Random(seed) is created and threaded through all strategy calls

tests/test_adversarial_generator.py

• Added TestSeeding class covering: same seed → identical output, different seeds → different output, string seeds, unseeded invocation, and fqname variance across 20 seeds

Usage

# Reproducible — same seed always yields the same pair
p1 = generate("type_narrowing_disguised_as_cleanup", seed=42)
p2 = generate("type_narrowing_disguised_as_cleanup", seed=42)
assert p1.old_signatures == p2.old_signatures  # True

# Random-looking — different seeds yield different names
p3 = generate("type_narrowing_disguised_as_cleanup", seed=99)
# p3 will have different fqname, param names, and types than p1

# Unseeded — draws OS entropy, non-reproducible
p4 = generate("type_narrowing_disguised_as_cleanup")

# Seed the full batch
pairs = generate_all(seed=7)

[sourcery-ai]

Sorry @Copilot, you have reached your weekly rate limit of 500000 diff characters.

Please try again later or upgrade to continue using Sourcery

[codacy-production]

Not up to standards ⛔

🔴 Issues 12 high · 1 medium · 4 minor

Alerts:
⚠ 17 issues (≤ 0 issues of at least minor severity)

Results:
17 new issues

Category	Results
Documentation	4 minor
Security	1 medium 12 high

View in Codacy

🟢 Metrics 20 complexity · 0 duplication

Metric	Results
Complexity	20
Duplication	0

View in Codacy

_{NEW Get contextual insights on your PRs based on Codacy's metrics, along with PR and Jira context, without leaving GitHub. Enable AI reviewer
TIP This summary will be updated as you push new changes.}