Configure analyzers

After activating a repository, DeepSource auto-detects languages and enables the matching analyzers. This guide walks you through fine-tuning that setup — enabling secrets detection, adjusting analyzer settings, excluding files, and turning on code formatters.

All of the settings below are available from the Settings tab on any repository page. For a full reference of every option, see Repository Settings.

Enable or disable analyzers

Open Settings > Code Review > Analyzers to see every analyzer available for your repository. Toggle an analyzer on or off, then expand it to configure language-specific options.

Common per-analyzer settings include:

• Cyclomatic complexity threshold — The risk tolerance for complex functions (low, medium, high, very-high, critical)
• Skip doc coverage — Artifact types to exclude from documentation coverage checks (e.g., magic methods, init, test)
• Runtime or language version — The language version your project targets (e.g., Python 3.x.x, Java 21, Rust stable)

Some analyzers expose additional options — framework selection for JavaScript, import root for Go, SQL formatting rules, and more. See the Core Analyzers reference for the full list.

Enable secrets detection

Under Settings > Code Review, toggle Enable secrets detection. When enabled, DeepSource scans every pull request for accidentally exposed API keys, passwords, and tokens — and fails the check if a secret is found.

For details on detection engines, see Secrets Detection.

Set test and excluded file patterns

Under Settings > Code Review > Codebase Context, define glob patterns so DeepSource understands your project layout:

• Test file patterns — e.g., **/*.test.js, **/*_test.go. Helps reduce false positives by distinguishing test code from production code.
• Excluded file patterns — e.g., **/.github/**, **/*migration*.*. Files matching these patterns are skipped during analysis.

To auto-format code on every pull request, see Enable code formatters.