Add Common Crawl and multi-source archive providers - #9

Summary

Add a provider layer for archive URL discovery while keeping Wayback as the default, backward-compatible source.

Sources

• --source wayback remains the default
• --source commoncrawl queries the Common Crawl CDXJ index
• --source wayback,commoncrawl (or reverse order) queries both and deduplicates normalized URLs across providers in source priority order
• JSONL provenance uses the provider that discovered a normalized URL first

Common Crawl

• dynamically resolves latest from collinfo.json
• supports explicit/repeatable or comma-separated --cc-index crawl IDs
• walks ZipNum pages using showNumPages + page
• adds --cc-page-size, --cc-rps, and --cc-filter
• serializes Common Crawl requests per run and defaults to 1 request/sec
• uses a descriptive wayparam User-Agent unless explicitly overridden
• treats Common Crawl 404/no-capture responses as empty results
• continues to later providers if one archive source fails

Compatibility

• existing commands still use Wayback only
• the normal normalization/filter/output pipeline is shared by every provider
• historical intelligence remains Wayback-only for now; incompatible source combinations are rejected explicitly
• optional web UI exposes Wayback, Common Crawl, or combined collection

Tests/docs

• Common Crawl collection discovery, pagination, metadata parsing, 404 behavior and UA tests
• cross-provider deduplication and partial-provider-failure tests
• CLI/GUI source configuration tests
• HTTP empty-status tests
• README, MkDocs source guide, architecture/API/CLI docs and manpage updates

Common Crawl integration queries only the public archive index; it does not fetch WARC bodies or contact live targets.