implement-verified-lightdash-api-client

Description

This skill implements a new Lightdash API operation in the internal/lightdash/api directory, along with necessary models in internal/lightdash/models. It relies on the research-lightdash-api skill for researching and verifying the API schema before implementation.

Input

The user should provide:

1. HTTP Method: (GET, POST, PUT, DELETE, PATCH)
2. API Endpoint: (e.g., /api/v1/projects/:projectUuid/spaces)
3. Function Name: (e.g., GetSpaceV1, CreateDashboardV1)
4. Documentation URL: (Optional) Link to the official Lightdash API docs.

Workflow

1. Research & Verification

• Invoke Research: Use the research-lightdash-api skill to research the endpoint and verify its schema against the live API (if LIGHTDASH_API_KEY is available).
• Outcome: Ensure you have a verified JSON response or documentation-based schema before proceeding to model generation.

2. Model Analysis & Generation

• Check existing models: Look in internal/lightdash/models for existing structs that match the resource.
• Generate/Update models: If no suitable model exists, create/update the file internal/lightdash/models/<resource>.go.
  • Use strict typing based on the verified JSON.
  • Add json tags.
  • Handle nullable fields with pointers where appropriate.
  • Follow Go naming conventions (PascalCase for exported fields).

3. API Client Implementation

• Create File: Create a new file internal/lightdash/api/<verb>_<resource>_v1.go (snake_case).
• Implement Method: Add the method to the Client struct: func (c *Client) <FunctionName>(...) (*<ReturnType>, error).
• Request Construction:
  • Use http.NewRequest.
  • Construct the path using fmt.Sprintf and c.HostUrl.
  • Validate input parameters (check for empty strings for UUIDs).
• Execution:
  • Call c.doRequest(req).
  • Unmarshal the response body into the typed model.
  • Return the Results field if the API wraps the response in a Results envelope.

4. Verification

• Compilation Check: Ensure the code compiles.
• Unit Tests: Create/Update internal/lightdash/api/<verb>_<resource>_v1_test.go to test unmarshaling logic with sample JSON.

Example Pattern

Ref: internal/lightdash/api/get_space_v1.go

package api

import (
 "encoding/json"
 "fmt"
 "net/http"
 "strings"

 "github.com/ubie-oss/terraform-provider-lightdash/internal/lightdash/models"
)

type GetResourceV1Response struct {
 Results models.Resource `json:"results"`
 Status  string          `json:"status"`
}

func (c *Client) GetResourceV1(id string) (*models.Resource, error) {
    if len(strings.TrimSpace(id)) == 0 {
        return nil, fmt.Errorf("id is empty")
    }

    path := fmt.Sprintf("%s/api/v1/resource/%s", c.HostUrl, id)
    req, err := http.NewRequest("GET", path, nil)
    if err != nil {
        return nil, fmt.Errorf("error creating request: %w", err)
    }

    body, err := c.doRequest(req)
    if err != nil {
        return nil, fmt.Errorf("error performing request: %w", err)
    }

    var response GetResourceV1Response
    if err := json.Unmarshal(body, &response); err != nil {
        return nil, fmt.Errorf("error unmarshaling response: %w", err)
    }

    return &response.Results, nil
}

Reference

• Lightdash API Documentation
• Lightdash Client Rules
• API Client Structure
• API Response Envelope

Assets

• API Operation Boilerplate
• Model Boilerplate
• Unit Test Boilerplate
• Schema Verification Script