Module: selfHealing

Self-Healing Utility for Playwright test execution.

Features

  • Multi-strategy element finding with retry logic
  • Healing history: records which strategy index succeeded per element so future runs try the winning strategy first (adaptive self-healing)
  • Comprehensive ARIA role coverage in assertion transforms
  • Code transform engine that rewrites raw Playwright calls into self-healing helpers

Exports

  • recordHealing — Record a successful healing result.
  • recordHealingFailure — Record a failed healing attempt.
  • getHealingHint — Get the previously-successful strategy index.
  • getHealingHistoryForTest — Serialise healing history for runtime injection.
  • getSelfHealingHelperCode — Generate the runtime helper code string.
  • applyHealingTransforms — Rewrite Playwright code to use self-healing helpers.
  • CORE_RULES — Native Playwright rules for local models (~200 tokens).
  • SELF_HEALING_PROMPT_RULES — Full self-healing helper rules for cloud models.
  • getPromptRules — Tier-aware getter: returns CORE_RULES or full rules.
Source:

Methods

(static) applyHealingTransforms(code) → {string}

Rewrite raw Playwright code to use self-healing helpers (safeClick, safeFill, safeExpect). Only transforms human-readable text selectors — CSS/XPath selectors are left untouched.

Parameters:
Name Type Description
code string

Raw Playwright test code.
Source:
Returns:

Transformed code with self-healing helper calls.

Type
string

(static) getHealingHint(testId, action, label) → {number}

Get the previously-successful strategy index for an action+label, or -1.

Parameters:
Name Type Description
testId string

Test ID.
action string

Action type.
label string

Element label/text.
Source:
Returns:

Strategy index (0-based), or -1 if no history.

Type
number

(static) getHealingHistoryForTest(testId) → {Object.<string, number>}

Serialise healing history for a test so it can be injected into runtime code.

Parameters:
Name Type Description
testId string

Test ID.
Source:
Returns:

Map of "action::label" → winning strategy index.

Type
Object.<string, number>

(static) getPromptRules(tier) → {string}

Return the appropriate prompt rules for the given tier.

  • "cloud" — full exhaustive rules (SELF_HEALING_PROMPT_RULES)
  • "local" — compact core rules only (CORE_RULES)
Parameters:
Name Type Description
tier "cloud" | "local"

The prompt tier.
Source:
Returns:

Prompt rules string.

Type
string

(static) getSelfHealingHelperCode(healingHintsopt) → {string}

Generate the self-healing runtime helper code as a string for injection into Playwright test execution context. Includes findElement, safeClick, safeFill, safeExpect, and retry logic.

Parameters:
Name Type Attributes Description
healingHints Object.<string, number> <optional>

Map of "action::label" → strategy index from previous runs.
Source:
Returns:

JavaScript code string to be prepended to test execution.

Type
string

(static) recordHealing(testId, action, label, strategyIndex)

Record a successful healing result.

Parameters:
Name Type Description
testId string

Test ID (e.g. "TC-1").
action string

Action type ("click", "fill", "expect").
label string

Element label/text used in the action.
strategyIndex number

Index of the winning strategy in the waterfall.
Source:

(static) recordHealingFailure(testId, action, label)

Record a failed healing attempt (all strategies exhausted).

Parameters:
Name Type Description
testId string

Test ID.
action string

Action type.
label string

Element label/text.
Source: