feat(page): enrich page view JSON with metadata via official API

[coseto6125]

Summary

page view --json today omits page metadata (timestamps, authors, parent, icon, archived state) because the MCP notion-fetch tool doesn't expose those fields — the output.Page struct has CreatedTime/LastEditedTime but they're left as zero values.

This PR adds an opt-in enrichment path: when an official API token is configured (via auth api setup or NOTION_API_TOKEN), page view --json makes a supplementary GET /pages/<id> call to the Notion REST API and merges the result into the JSON output.

When no token is configured, the command behaves exactly as before — no network overhead, no behavior change.

This continues the MCP-first + REST-fallback pattern established in #17 and the groundwork laid by #28.

Changes

internal/api/client.go

• Add PageMetadata, PartialUser, PageIcon, PageParent types mapping the Notion REST API response
• Add GetPageMetadata(ctx, pageID) method calling GET /pages/<id>

internal/output/types.go

• Add CreatedBy and LastEditedBy string fields to Page (timestamps were already defined)

cmd/page.go

• renderFetchedPageView calls a new loadPageMetadata helper when JSON output is requested, and merges metadata via applyPageMetadata before calling printViewedPageFn
• loadPageMetadata is swappable (test seam) and swallows errors from missing token / API failure so the primary output is never blocked
• Text and --raw modes skip the API call entirely

cmd/page_view_test.go

• 4 new tests: enrichment happy path, token-missing fallback, text-mode skip, parent/icon variant coverage

internal/api/client_test.go

• 2 new tests: full response parsing, blank page ID rejection

README.md

• Document the enrichment behavior under the Pages section

Behavior matrix

Mode	API token	Metadata enrichment	Extra API call
--json	configured	Yes	1 × GET /pages/
--json	missing/invalid	No (fields empty)	No (short-circuit)
default text	either	No	No
--raw	either	No	No

Test plan

• go build ./... clean
• go test ./... passes (existing + new tests)
• go vet ./... clean
• gofmt -l . empty
• Manual smoke: NOTION_API_TOKEN=ntn_xxx notion-cli page view <url> --json | jq '{CreatedBy, LastEditedBy, Archived}' returns populated fields
• Manual smoke without token: output matches previous v0.6.0 behavior

Notes

• Intentionally silent on API errors: if the REST call fails, the MCP-sourced fields (ID, Title, URL, Content, Comments) still render, preserving the fundamental page view UX
• Only JSON output triggers the API call — reasoning: the text renderer doesn't display these fields today, so there's no benefit to paying the extra latency
• Archived combines archived || in_trash since the Notion API returns both flags

Refs #17

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: e1f7804f38

ℹ️ About Codex in GitHub

Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".

[chatgpt-codex-connector]

Handle data_source parent IDs in metadata mapping

This switch only maps database_id, but with newer Notion API versions pages in databases commonly come back as parent.type = "data_source_id"; in that case parentFromMetadata falls through and returns an empty ParentID. The new page view --json enrichment therefore drops parent IDs for data-source-backed pages, which breaks the advertised ParentType/ParentID metadata for a large class of pages.

Useful? React with 👍 / 👎.