feat(cli): digest command for windowed per-channel summary

[mvanhorn]

Summary

Adds discrawl digest, a per-channel activity summary over a time window. discrawl already has report for repo-wide README dumps and messages / search for retrieval; digest answers "what happened in this guild over the last N days, per channel."

Ports vincentkoc/slacrawl#9 (merged 2026-04-22). Same SQL recipe, adapted to discrawl's Discord schema (guild_id, members, mention_events) and the existing stdlib-flag CLI dispatch.

Why this matters

• Issue Using discrawl as memory augmentation for AI agents #15 (@codexGW's writeup on using discrawl as memory augmentation for AI agents) describes building a --mode raw|distilled|auto reader on top of discrawl's FTS5 tables. A built-in digest gives that workflow a first-class entry point.
• The schema, indexes (idx_messages_guild_created_id, idx_mentions_guild_event), and internal/report/ package are already exactly what digest needs. One SQL recipe away.
• Keeps the focused-commands shape. Per-channel windowed activity is one of the most-asked questions of an org-memory archive.

Output

$ discrawl digest --since 7d

general (text)
  messages=42 threads=3 authors=8
  top posters  alice (12), bob (9), carol (7)
  top mentions @oncall (4), @sre (2)

incidents (text)
  messages=11 threads=2 authors=4
  top posters  alice (5), dave (3)
  top mentions @oncall-bot (3)

releases (text)
  messages=4 threads=1 authors=2
  top posters  carol (3)
  top mentions -

Window: 2026-04-18 to 2026-04-25 (7d)
Totals: messages=57 threads=6 channels=3 authors=11

Flags

• --since lookback window. Accepts Go durations (72h) plus Nd shorthand (7d, 30d). Default 7d.
• --guild filter to one guild. When omitted, falls back to default_guild_id from config when set.
• --channel drill into one channel by id or name.
• --top-n number of top posters / mention targets per channel. Default 3.
• Inherits --json and --plain from the root CLI.

Scope

• No schema changes. Uses existing messages / channels / members / mention_events tables.
• No new dependencies in go.mod.
• New: internal/report/digest.go, internal/report/digest_test.go, internal/cli/digest.go, internal/cli/digest_test.go.
• Touched: internal/cli/cli.go (one switch case), internal/cli/output.go (text and plain renderers + usage), internal/report/report.go (added snake_case JSON tags to RankedCount), README.md, SPEC.md.
• Tests cover: happy path, window filter, channel filter (id and name), guild filter, default top-n, top-mentions ranking, negative-since normalization, JSON output schema, CLI dispatch.

Test plan

• gofmt -l . clean
• go vet ./... clean
• go build ./cmd/discrawl
• go test ./... passes (11 packages, all green)

Self-reviewed via codex review before pushing. Found one snake_case JSON tag inconsistency on RankedCount, fixed in a follow-up commit on this branch.

Open questions

1. Default --top-n: went with 3 to match what report already does internally for top channels and authors. Happy to change to 1 or 5.
2. Naming: digest matched the natural framing vs summary or folding into report --digest. The slacrawl PR went the same direction; happy to rename.
3. Helper sharing: kept digest's SQL local to internal/report/digest.go. If you'd rather factor topPosters / topMentions style helpers into report.go so both commands share them later, easy follow-up.
4. JSON tag change on RankedCount: added json:"name" and json:"count" so digest's top_posters and top_mentions serialize as snake_case (consistent with the rest of the digest schema). The existing report command renders Markdown, so this change is observable only via JSON consumers; no current caller uses JSON for RankedCount.

Related work

This is the first half of a two-PR sequence. The follow-up is discrawl analytics (subcommand group with quiet and trends), porting vincentkoc/slacrawl#13. Stacked on top of this PR so analytics digest can shim to runDigest and share the implementation.

This contribution was developed with AI assistance.

[chatgpt-codex-connector]

To use Codex here, create a Codex account and connect to github.