API Testing Portfolio

This project was created to demonstrate my API testing skills using a real public API. The collection was designed by first understanding the API from a user perspective, then examining actual response data to build accurate schemas, rather than relying solely on documentation.

Attribution

This product uses TMDB and the TMDB APIs but is not endorsed, certified, or otherwise approved by TMDB.

Files

• tmdb-api-tests.json - Postman collection (v2.1)

Collection structure

• Authentication - validates successful and unsuccessful authentication.
• Popular lists - validates movie and TV popular list endpoints.
• Page detail - validates movie and TV detail endpoints, chained from list results.
• Search - validates move and TV search endpoints, including a no-results case.

The API

The Movie Database (TMDB) is a community-built movie and TV database. This collection tests the following endpoints:

• https://api.themoviedb.org/3/authentication - validates the current access token.
• https://api.themoviedb.org/3/movie/popular - returns a paginated list of popular movies.
• https://api.themoviedb.org/3/tv/popular - returns a paginated list of popular TV shows.
• https://api.themoviedb.org/3/movie/{movie_id} - returns the details of a movie.
• https://api.themoviedb.org/3/tv/{series_id} - returns the details of a TV show.
• https://api.themoviedb.org/3/search/movie - searches movies by query string.
• https://api.themoviedb.org/3/search/tv - searches TV shows by query string.

What the collection tests

Authentication

• Schema validation with AJV on the authentication response.
• Success state validation for valid and invalid credentials.
• Status code validation for both authorized and unauthorized responses.
• Content-Type header and response time validation.

Popular lists

• Schema validation with AJV on movie and TV list results.
• Range validation on numeric fields.
• Array item type validation.
• Null handling on fields with known undocumented behavior.
• Date format validation (ISO 8601).
• ID integrity checks — type and range validation via schema and duplicate detection.
• Pagination metadata validation.
• Status code, Content-Type header, and response time validation.

Page detail

• Schema validation with AJV on movie and TV detail responses.
• Nested object and array validation.
• Nullable object validation using oneOf.
• ISO 3166-1 and ISO 639-1 format validation using regex patterns.
• Date format validation (ISO 8601), including nested episode and season dates.
• Data integrity checks (season count consistency, episode/season ratio, ended shows without next episodes).
• Response ID verification confirming chained requests return the correct record.
• Status code, Content-Type header, and response time validation.

Search

• Schema validation with AJV on movie and TV search results.
• Range validation on numeric fields.
• Array item type validation.
• Null handling on fields with known undocumented behavior.
• Date format validation (ISO 8601).
• ID integrity checks — type and range validation via schema and duplicate detection.
• Pagination metadata validation.
• Empty results validation for no-results queries, including explicit assertions on results array, total results, and total pages.
• Status code, Content-Type header, and response time validation.

Design decisions

• The schemas were designed from actual response data, not just documentation.
• Null handling was added based on observed API behavior vs documented behavior.
• List sort order validation was investigated but not implemented. TMDB's popularity scores update in real time, meaning the sort order of a given page can shift between requests. Due to this inconsistency, a test written against this behavior would pass or fail at different times based on the data state. I decided to document this limitation rather than include a test that yields no reliable signal.
• Detail page request IDs are captured dynamically from list responses using collection variables, requiring no manual configuration beyond the access token.
• poster_path and backdrop_path were not marked as required fields despite being displayed on the UI, as they are not guaranteed for all entries on this community-contributed platform.
• number_of_seasons counts aired official seasons and excludes specials, which TMDB lists as season 0 in the seasons array. The number of seasons test accounts for this by filtering out any specials before comparing against number_of_seasons.
• Search result sort order validation was investigated but not implemented. Results returned by the search endpoints show no consistent ordering pattern across any field. This limitation is documented rather than tested.
• Missing and empty search query parameter behavior was investigated. Both cases return a 200 with an empty results array identical to a no-results search, rather than an error response. These cases are covered by the no-results request rather than treated as separate error conditions.
• Response time threshold set at 500ms based on observed response times consistently under 100ms.

Findings during development

These findings were identified during schema design by comparing official documentation against actual API responses.

Authentication

• status_code and status_message are documented as returned fields for both 200 and 401 responses, but are absent from the 401 response. Only success is returned.

TV popular list

• The adult field is present in responses but missing from the official documentation.
• The documentation field order doesn't match the actual response order.

Movie details

• belongs_to_collection is documented as a string but returns either null or an object containing id, name, poster_path, and backdrop_path.

TV details

• next_episode_to_air is documented as a string but returns either null or a full episode object.
• vote_average in season objects is documented as an integer, but returns a number with decimal values.
• episode_type appears in episode objects in observed responses, but is not documented.

Search

• Search results total is capped at 10,000, regardless of actual matches. This appears to be an intentional API limitation rather than a bug.

Test results

How to run

1. Register for a free TMDB API key by creating an account at https://www.themoviedb.org/signup and then going to https://www.themoviedb.org/settings/api.
2. Import the collection into Postman.
3. Create an environment with the following variables:
   • accessToken - your TMDB Read Access Token, used as a Bearer token.
4. Run the collection.