fix(complexity): calibrate Halstead thresholds to file level, add p90 outlier category

[aspala]

Die Complexity-Schwellen stammen aus der Halstead-Literatur, wo sie für einzelne Funktionen gelten. codeqa berechnet Halstead aber pro Datei, wodurch sie zwei Größenordnungen zu niedrig lagen: eine gewöhnliche 300-LOC-Datei misst effort ~867k gegen einen A-Schwellwert von 5000, sodass jedes reale Repo D- bekam — codeqa-action auf sich selbst eingeschlossen (30/D-, jetzt 85/A-).

Der Mittelwert allein verbarg zudem genau die Ausreißer, die man sehen will, hinter der Masse kleiner Dateien; deshalb kommt p90 als eigene Kategorie dazu, die die Verteilung statt des Durchschnitts bewertet.

[github-actions]

Score: B → B- | Δ -2 pts | 0 blocks flagged across 5 files | 5 modified, 0 added

🟡 Code Health: B- (71/100)

196 files · codeqa-action · 2026-09-01

Combined metric scores use cosine similarity: +1 = metric profile perfectly matches healthy pattern for this behavior, 0 = no signal, −1 = anti-pattern detected. Mapped to 0–100 using breakpoints (approx: ≥0.5→A, ≥0.2→B, ≥0.0→C, ≥−0.3→D, <−0.3→F); actual letter grades use the full 15-step scale.

Metric Changes

Category	Base	Head	Δ
Readability	88.64	97.89	+9.25
Complexity	30.63	41.57	+10.94
Duplication	0.57	0.59	+0.02
Structure	6.23	9.20	+2.97

%%{init: {'theme': 'neutral'}}%%
xychart-beta
    title "Code Health Scores"
    x-axis ["Readability", "Complexity", "Complexity Outliers", "Structure", "Duplication", "Naming", "Magic Numbers", "Combined Metrics"]
    y-axis "Score" 0 --> 100
    bar [94, 85, 75, 87, 49, 96, 100, 65]

Loading

Readability          ███████████████████░   94  🟢 A
Complexity           █████████████████░░░   85  🟢 A-
Complexity Outliers  ███████████████░░░░░   75  🟡 B
Structure            █████████████████░░░   87  🟢 A-
Duplication          ██████████░░░░░░░░░░   49  🟠 C-
Naming               ███████████████████░   96  🟢 A
Magic Numbers        ████████████████████  100  🟢 A
Combined Metrics     █████████████░░░░░░░   65  🔴 D

[github-actions]

🔍 Top Likely Issues (cosine similarity)

Most negative cosine = file's metric profile best matches this anti-pattern.

Behavior	Cosine	Score
dependencies.low_coupling	-0.56	-12.98
file_structure.single_responsibility	-0.52	-12.68
file_structure.line_count_under_300	-0.45	-9.59
code_smells.no_dead_code_after_return	-0.41	-23.23
scope_and_assignment.shadowed_by_inner_scope	-0.31	-5.13
file_structure.line_length_under_120	-0.30	-8.41
variable_naming.name_contains_type_suffix	-0.23	-2.28
variable_naming.loop_var_is_single_letter	-0.23	3.48
variable_naming.name_contains_and	-0.21	-36.63
consistency.same_concept_same_name	-0.21	-2.34

🟢 Readability — A (94/100)

Codebase averages: flesch_adapted=97.89, fog_adapted=4.82, avg_tokens_per_line=9.57, avg_line_length=35.81

Metric	Value	Score
readability.flesch_adapted	97.89	100
readability.fog_adapted	4.82	100
readability.avg_tokens_per_line	9.57	72
readability.avg_line_length	35.81	100

🟢 Complexity — A- (85/100)

Codebase averages: difficulty=41.57, effort=243432.62, volume=4137.53, estimated_bugs=1.38

Metric	Value	Score
halstead.difficulty	41.57	81
halstead.effort	243432.62	83
halstead.volume	4137.53	89
halstead.estimated_bugs	1.38	90

🟡 Complexity Outliers — B (75/100)

Codebase averages: p90_volume=9542.41, p90_difficulty=67.33, p90_effort=659285.40

Metric	Value	Score
halstead.p90_volume	9542.41	88
halstead.p90_difficulty	67.33	55
halstead.p90_effort	659285.40	81

🟢 Structure — A- (87/100)

Codebase averages: branching_density=0.14, mean_depth=3.85, avg_function_lines=8.37, max_depth=9.18, max_function_lines=20.13, variance=6.77, avg_param_count=1.16, max_param_count=2.08

Metric	Value	Score
branching.branching_density	0.14	76
indentation.mean_depth	3.85	88
function_metrics.avg_function_lines	8.37	89
indentation.max_depth	9.18	87
function_metrics.max_function_lines	20.13	90
indentation.variance	6.77	100
function_metrics.avg_param_count	1.16	100
function_metrics.max_param_count	2.08	100

🟠 Duplication — C- (49/100)

Codebase averages: redundancy=0.59, bigram_repetition_rate=0.54, trigram_repetition_rate=0.37

Metric	Value	Score
compression.redundancy	0.59	58
ngram.bigram_repetition_rate	0.54	38
ngram.trigram_repetition_rate	0.37	41

🟢 Naming — A (96/100)

Codebase averages: entropy=0.89, mean=6.64, variance=18.73, avg_sub_words_per_id=1.17

Metric	Value	Score
casing_entropy.entropy	0.89	100
identifier_length_variance.mean	6.64	100
identifier_length_variance.variance	18.73	85
readability.avg_sub_words_per_id	1.17	100

🟢 Magic Numbers — A (100/100)

Codebase averages: density=0.00

Metric	Value	Score
magic_number_density.density	0.00	100

🔴 Combined Metrics — D (65/100)

Category	Score	Grade
Code Smells	25	🔴 D-
Consistency	82	🟡 B+
Dependencies	19	🔴 E+
Documentation	85	🟢 A-
Error Handling	91	🟢 A-
File Structure	48	🟠 C-
Function Design	82	🟡 B+
Naming Conventions	90	🟢 A-
Scope And Assignment	29	🔴 D-
Testing	83	🟡 B+
Type And Value	89	🟢 A-
Variable Naming	74	🟡 B

🔴 Code Smells — D- (25/100)

Cosine similarity scores for 1 behaviors.

Behavior	Cosine	Score	Grade
no_dead_code_after_return	-0.41	25	D-

🟡 Consistency — B+ (82/100)

Cosine similarity scores for 1 behaviors.

Behavior	Cosine	Score	Grade
consistent_function_style	0.39	82	B+

🔴 Dependencies — E+ (19/100)

Cosine similarity scores for 1 behaviors.

Behavior	Cosine	Score	Grade
low_coupling	-0.56	19	E+

🟢 Documentation — A- (85/100)

Cosine similarity scores for 3 behaviors.

Behavior	Cosine	Score	Grade
file_has_module_docstring	0.32	78	B+
docstring_is_nonempty	0.46	88	A-
function_has_docstring	0.47	88	A-

🟢 Error Handling — A- (91/100)

Cosine similarity scores for 3 behaviors.

Behavior	Cosine	Score	Grade
error_message_is_descriptive	0.46	87	A-
does_not_swallow_errors	0.62	92	A-
returns_typed_error	0.66	93	A

🟠 File Structure — C- (48/100)

Cosine similarity scores for 5 behaviors.

Behavior	Cosine	Score	Grade
single_responsibility	-0.52	20	E+
line_count_under_300	-0.45	24	E+
line_length_under_120	-0.30	30	D-
has_consistent_indentation	0.28	75	B
no_magic_numbers	0.58	92	A-

🟡 Function Design — B+ (82/100)

Cosine similarity scores for 3 behaviors.

Behavior	Cosine	Score	Grade
is_less_than_20_lines	0.33	79	B+
has_verb_in_name	0.40	83	B+
no_magic_numbers	0.40	83	B+

🟢 Naming Conventions — A- (90/100)

Cosine similarity scores for 1 behaviors.

Behavior	Cosine	Score	Grade
function_name_is_not_single_word	0.50	90	A-

🔴 Scope And Assignment — D- (29/100)

Cosine similarity scores for 1 behaviors.

Behavior	Cosine	Score	Grade
shadowed_by_inner_scope	-0.31	29	D-

🟡 Testing — B+ (83/100)

Cosine similarity scores for 2 behaviors.

Behavior	Cosine	Score	Grade
test_single_concept	0.28	75	B
test_name_describes_behavior	0.54	91	A-

🟢 Type And Value — A- (89/100)

Cosine similarity scores for 1 behaviors.

Behavior	Cosine	Score	Grade
hardcoded_url_or_path	0.48	89	A-

🟡 Variable Naming — B (74/100)

Cosine similarity scores for 1 behaviors.

Behavior	Cosine	Score	Grade
name_is_generic	0.26	74	B

[github-actions]

---

kind: refactoring-tasks
path: /home/runner/work/codeqa-action/codeqa-action
timestamp: 2026-09-01T17:53:01.452397Z
overall_grade: B-
overall_score: 71
task_count: 0
critical: 0
high: 0
instructions: >-
Address the tasks below in order of severity (critical first).
After each fix, run the project's test suite and confirm it passes
before moving on.

No critical or high-severity blocks need attention. ✅