Response Editing - MQUAD Platform

Edit Rules

Survey responses can be edited after submission, subject to configurable rules:

class EditRules(BaseModel):
    enabled: bool                        # Allow editing at all
    time_limit_hours: Optional[int]      # Edit window (e.g., 24h)
    max_edits: Optional[int]             # Maximum revisions
    require_reason: bool                 # Require change reason
    require_approval: bool               # Re-approval after edit

Version Tracking

Every edit creates a new version:

{
  "id": "response-uuid",
  "survey_id": "survey-uuid",
  "current_version": 3,
  "versions": [
    {
      "version": 1,
      "data": { "q1": "initial answer" },
      "edited_by": "user-uuid",
      "edited_at": "2026-05-13T10:00:00Z",
      "reason": "Initial submission"
    },
    {
      "version": 2,
      "data": { "q1": "corrected answer" },
      "edited_by": "user-uuid",
      "edited_at": "2026-05-13T11:30:00Z",
      "reason": "Data entry correction"
    }
  ],
  "audit_trail": [...]
}

Edit Flow

User submits edit
       │
       ▼
Check edit rules
  ├── Time limit expired? → Reject
  ├── Max edits reached? → Reject
  │
  ▼
Create new version
  ├── Store delta (only changed fields)
  ├── Increment version number
  │
  ▼
Generate audit event (response.edited)
       │
       ▼
Require re-approval? → Notify approvers

Conflict Detection

When editing a response that was also edited elsewhere:

async def edit_response(response_id, new_data, expected_version):
    response = await get_response(response_id)
    if response.current_version != expected_version:
        raise ConflictError(
            "Response has been modified since you loaded it. "
            f"Expected version {expected_version}, got {response.current_version}."
        )

API

PUT /responses/{id}
{
  "data": { "q1": "updated answer" },
  "reason": "Corrected typo in patient name",
  "expected_version": 2
}

# Response
{
  "id": "response-uuid",
  "current_version": 3,
  "status": "pending_approval"
}

Audit

Every edit is recorded in the audit log:

await log_audit(
    event_type="response.edited",
    resource_type="response",
    resource_id=response_id,
    metadata={
        "previous_version": 2,
        "new_version": 3,
        "changed_fields": ["q1"],
        "reason": "Corrected typo"
    }
)