[lahirumaramba]

Newer @types/node releases declare a global Web Storage interface, which collides with the Storage class exported by firebase-admin/storage. To avoid the global name clash, API Extractor renames our class to Storage_2 in the generated .api.json doc model, causing api-documenter to publish a Storage_2 class page and drop inline {@link Storage} links in getStorage().

This change normalizes Storage_2 back to Storage during .api.json post-processing before api-documenter runs, restoring the clean Storage class name, page filename, and cross-links without altering the API Extractor .api.md reports.

Technical Details

• Renames stripHiddenDocsFromApiJson in generate-reports.js to postProcessApiJson and its inner walker to visitNode so it covers both @excludeFromDocs member filtering and symbol normalization.
• Rewrites Storage_2 node names and !Storage_2 canonical references across members and excerptTokens in generated temp/*.api.json files.
• Moves @microsoft/api-extractor loading into generateReportForEntryPoint so it is only required when generating reports.

[gemini-code-assist]

Code Review

This pull request updates generate-reports.js to post-process the generated API JSON report by renaming stripHiddenDocsFromApiJson to postProcessApiJson and adding logic to normalize global name collisions (specifically renaming Storage_2 back to Storage in node.name and node.canonicalReference). The feedback suggests also normalizing node.text when it is 'Storage_2' to ensure that the rendered TypeScript signatures in the documentation (which rely on excerptTokens' text property) do not display Storage_2.