skylos_to_gitlab

Convert Skylos dead code reports to GitLab Code Quality format

Output: JSON

Arg	Description
`min-confidence`	Minimum confidence threshold (0-100). Items below this value are excluded.

This mold converts JSON reports from duriantaco/skylos (a Python static analysis tool) into the GitLab Code Quality (Code Climate) format. This allows Skylos results to be visualized directly in GitLab Merge Requests.

Usage

fimod s -i skylos_report.json -m @skylos_to_gitlab -o code_quality.json

# Filter out low-confidence results (potential false positives)
fimod s -i skylos_report.json -m @skylos_to_gitlab --arg min-confidence=80 -o code_quality.json

Arguments

Argument	Default	Description
`min-confidence`	`0` (no filter)	Minimum confidence threshold (0-100). Items below this value are excluded.

In a GitLab CI pipeline

code_quality:
  script:
    - skylos --json > skylos_report.json
    - fimod s -i skylos_report.json -m @skylos_to_gitlab -o gl-code-quality-report.json
  artifacts:
    reports:
      codequality: gl-code-quality-report.json

Transformation Logic

The converter maps Skylos report sections to Code Climate issue categories:

Skylos Section	Code Climate `check_name`
`unused_functions`	`unused-function`
`unused_imports`	`unused-import`
`unused_variables`	`unused-variable`
`unused_classes`	`unused-class`
`unused_parameters`	`unused-parameter`
`unused_files`	`dead-file`

Confidence & Severity

Skylos reports a confidence score (0-100) for each item. The mold uses it in two ways:

Filtering: Pass --arg min-confidence=N to exclude items with confidence below the threshold.
Severity mapping: Items with confidence = 100 get severity info. Items with lower confidence get severity minor.

Fingerprinting

A deterministic fingerprint is generated for each issue using the most qualified name available (full_name → name), the check name, file path, and line number. This allows GitLab to track issues across commits.

Output Format

The output is a JSON array of issue objects:

[
  {
    "description": "Unused function: my_func",
    "check_name": "unused-function",
    "fingerprint": "a1b2c3d4",
    "severity": "info",
    "location": {
      "path": "src/main.py",
      "lines": {
        "begin": 42
      }
    }
  }
]