Add multi-archive search support and identifiable search results

[HelloWorld-25]

Issue : #230 As highlighted by #229, Searcher's getResults() only yields result's entry path. while convenient for single-archive search, it prevents implementing multi-ZIM search as results would only be path strings from multiple ZIMs.

We should then implement multiple ZIM search properly by

Binding addArche to Searcher (ref impl in #229)
Change Searcher API so that results can be identified

---

Our Changes :

Allowing multiple archives to be bound to a Searcher

1. Added a _archives list to track all registered archives.

 2. Introduced Searcher.addArchive(archive: Archive) method to register additional archives.

 3. Searcher.search(query) now searches across all bound archives.

Returning identifiable search results

    1. Introduced a new class: SearchResult containing:

              class SearchResult:
                    archive: Archive
                    path: str

    2. Updated SearchResultSet.__iter__() to yield SearchResult objects instead of strings.

   3. Results are now unambiguous and include both archive and entry path.

Updated type hints in search.pyi

   1. SearchResultSet.__iter__() now returns Iterator[SearchResult]

   2. Searcher.addArchive() is added to type hints

   3. Python API fully matches the new Cython implementation.

---

Benefits :

1.Enables multi-ZIM search

2.Ensures search results are uniquely identifiable

3.Clean, maintainable API, consistent with libzim C++ internals

4.Future-proof for features like ranking, filtering, and deduplication across multiple archives.

---

Backward Compatibility :

1.The API change from Iterator[str] → Iterator[SearchResult] is intentional to support multi-ZIM search.

2.Users can still access the path via result.path to simplify migration.

---