feat(contacts): add dedupe preview

[steipete]

Summary

• extract the useful contacts dedupe idea from feat: drive reporting/sync + contacts dedupe #116 as a preview-only command
• add gog contacts dedupe with email/phone matching by default and opt-in name matching
• emit JSON/table merge plans without applying updates or deletes
• add generated command docs, README/spec coverage, changelog credit, and regression tests

Testing

• go test ./internal/cmd -run 'Test(ContactsDedupe|ParseContactsDedupe|BuildContactsDedupe)'\n- make build\n- make docs-site\n- ./bin/gog --json --account clawdbot@gmail.com contacts dedupe --max 50\n- make lint\n- make test\n- make ci\n\nSupersedes the contacts dedupe slice of feat: drive reporting/sync + contacts dedupe #116; Drive reporting already landed in feat(drive): add read-only reporting commands #554 and Drive sync remains intentionally separate.

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 0967f53931

ℹ️ About Codex in GitHub

Your team has set up Codex to 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 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

Count match keys per contact before building matched_on

keyCounts is incremented for every occurrence in a contact, so repeated values inside a single person (for example the same email stored twice) are treated as if the key matched multiple contacts. When that person is grouped with others via a different key, the output can incorrectly list that repeated key in matched_on, which makes the merge plan claim evidence that does not actually exist between members. De-duplicating keys per person before incrementing keyCounts avoids these false-positive match reasons.

Useful? React with 👍 / 👎.