fix(feishu): route tables and code blocks to CardKit 2.0 with post/text fallback by lqhl · Pull Request #23861 · NousResearch/hermes-agent

lqhl · 2026-05-11T15:25:25Z

Summary

Routes GFM tables and multi-line fenced code blocks to CardKit 2.0 interactive messages in Feishu, with automatic fallback to plain text when card delivery fails.

Problem

Feishu's post/md tag has two rendering limitations:

Tables — GFM table syntax (| col1 | col2 |) is silently dropped
Code blocks — Fenced code blocks longer than ~6 lines are truncated

Solution

Three new regex detectors + CardKit 2.0 builder + failure fallback:

_TABLE_MARKDOWN_RE — detects GFM tables (pipe-separated rows + alignment separator)
_CODE_BLOCK_RE — detects fenced code blocks with 2+ content lines (short blocks stay in post/md)
_build_card_payload() — wraps content in a CardKit 2.0 card with tag: markdown element
_build_outbound_payload() — routes detected content to interactive msg_type

Fallback mechanism (not present in original PR #19038)

If the interactive card send fails (bot lacks card permission, API rejection, etc.), the message automatically falls back to plain text. Three layers of protection:

Exception handler in send() — catches send failures, retries as plain text
Response check in send() and edit_message() — catches API rejection without exception
_feishu_send_with_retry() — interactive failures raise immediately (no retry loop)

Testing

11 new unit tests covering table detection, code block detection, CardKit JSON format, regex boundary cases, and mixed content
209/209 existing Feishu tests pass (no regressions)
Live-tested on a real Feishu bot (both tables and code blocks render correctly as CardKit 2.0 cards)

Changes

gateway/platforms/feishu.py: +62/-3 lines
tests/gateway/test_feishu.py: +114/-0 lines

Cherry-pick of PR 19038 plus fallback mechanism: - _TABLE_MARKDOWN_RE: detect GFM tables - _CODE_BLOCK_RE: detect fenced code blocks with 2plus content lines - _build_card_payload: CardKit 2.0 interactive card with tag:markdown - _build_outbound_payload: route tables and long code blocks to interactive - send() and edit_message(): fallback to plain text on interactive failure - _feishu_send_with_retry(): fail fast on interactive errors Fixes 19035 and 9549 (table and code block rendering in Feishu)

lqhl · 2026-05-11T15:29:18Z

I've tested this PR with my feishu bot:

liuhao1024 · 2026-05-11T15:48:58Z

Minor: Comment/regex mismatch in `_CODE_BLOCK_RE`

The inline comment says "route 3+ content lines to CardKit 2.0" but the regex uses {2,} which matches 2+ content lines:

# Multi-line fenced code block (opening fence + 2+ content lines + closing fence)
# Empty or 1-line blocks render fine in post/md; route 3+ content lines to CardKit 2.0.
_CODE_BLOCK_RE = re.compile(r"^```[^\n]*\n(.*\n){2,}```", re.MULTILINE | re.DOTALL)

The test at line 217 confirms a 2-line block ("```\na\nb\n```") does match, so the regex behavior is {2,}. Either:

Change {2,} → {3,} if the intent is truly 3+ content lines, or
Update the comment to say "2+ content lines"

Additionally, re.DOTALL is unnecessary here since \n is already explicit in the pattern. It makes .* match across newlines which adds backtracking complexity without changing the outcome on well-formed input. Consider dropping it.

…drop redundant DOTALL _CODE_BLOCK_RE uses {2,} to match 2+ content lines, but the comment claimed '3+ content lines'. Fix the comment to match the actual regex. Also remove re.DOTALL — the pattern already uses explicit \n, so DOTALL only adds unnecessary backtracking complexity.

lqhl · 2026-05-12T07:55:27Z

@liuhao1024 Thanks for catching this! Fixed in 8a60abe:

Comment updated: "3+ content lines" → "2+ content lines" to match the {2,} regex
Removed re.DOTALL — as you noted, \n is already explicit so DOTALL only adds backtracking overhead

Both changes are comment/cosmetic — the regex behavior is unchanged, 209/209 feishu tests still pass.

lqhl · 2026-05-18T15:16:49Z

@teknium1 hi, can you take a look?

…ards Closes NousResearch#9549, NousResearch#19035. Supersedes NousResearch#23861. ## Problem Feishu's post-type 'md' element does not render GFM tables and truncates multi-line fenced code blocks. The current behaviour (force-text fallback, PR NousResearch#20275) avoids the blank-message symptom but leaks raw markdown source to the user — pipes, separators and code fences are visible as plain text. ## Approach Route content containing GFM tables or multi-line fenced code blocks to CardKit 2.0 interactive messages (schema: 2.0, tag: markdown), which render both natively. Plain markdown without tables/code stays in post/md as before. Falls back to plain text when the interactive card is rejected (bot lacks card permission, malformed payload, etc.) so failures degrade gracefully. ## ErrCode 11310: per-card table cap CardKit 2.0 caps the total number of GFM tables across an entire card at 5. Exceeding this triggers ErrCode 11310 ('card table number over limit') and the API rejects the whole card. The cap is per-CARD, not per-element — verified empirically against the live Feishu API on 2026-05-27. A single element with 6 tables fails. So do two elements with 4+2 tables, three elements with 2+2+2, and a layout with one table per element. Five tables in any layout pass. When content has more than 5 tables, _build_outbound_messages() splits it into multiple cards and send() iterates the resulting (msg_type, payload) list. Splits cut at section boundaries — the first paragraph break after the previous table block ends — so each table's heading and lead-in prose travel into the same card as the table itself, rather than being orphaned in the previous chunk. ## Tests tests/gateway/test_feishu.py::TestCardTableLimitSplitting - single card for ≤5 tables - two cards for 6 tables (5 + 1) - three cards for 12 tables (5 + 5 + 2) - non-table content keeps post/text routing untouched - prose between tables stays with the following table at split points ## Migration notes - Bots without card-send permission will silently fall back to plain text (existing fallback path, extended for the interactive case). - Edit_message remains on the single-payload path; streaming edits do not currently produce > 5 tables in a single chunk in practice.

alt-glitch added type/bug Something isn't working comp/gateway Gateway runner, session dispatch, delivery platform/feishu Feishu / Lark adapter P2 Medium — degraded but workaround exists labels May 11, 2026

github-actions Bot mentioned this pull request May 12, 2026

🦞 OpenClaw 生态日报 2026-05-12 ivanweng2077/big_model_radar#31

Open

This was referenced May 12, 2026

feat(feishu): render markdown tables as interactive cards via column_set layouts #24249

Open

feat(feishu): support markdown tables via CardKit table element #27695

Open

alt-glitch mentioned this pull request May 26, 2026

feat(feishu): Schema 2.0 markdown card for GFM tables, code blocks, and inline formatting #32488

Open

huangyoje mentioned this pull request May 27, 2026

fix(feishu): render markdown tables and code blocks via CardKit 2.0 cards #33310

Open

alt-glitch mentioned this pull request May 31, 2026

feat(feishu): send GFM tables via Schema 2.0 interactive card #35781

Open

banditburai mentioned this pull request Jun 1, 2026

fix(feishu): render markdown tables via post instead of force-text downgrade #36877

Open

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

fix(feishu): route tables and code blocks to CardKit 2.0 with post/text fallback#23861

fix(feishu): route tables and code blocks to CardKit 2.0 with post/text fallback#23861
lqhl wants to merge 2 commits into
NousResearch:mainfrom
lqhl:fix/feishu-card-tables-codeblocks-v2

lqhl commented May 11, 2026

Uh oh!

lqhl commented May 11, 2026

Uh oh!

liuhao1024 commented May 11, 2026

Uh oh!

lqhl commented May 12, 2026 •

edited

Loading

Uh oh!

lqhl commented May 18, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

Conversation

lqhl commented May 11, 2026

Summary

Problem

Solution

Fallback mechanism (not present in original PR #19038)

Testing

Changes

Related

Uh oh!

lqhl commented May 11, 2026

Uh oh!

liuhao1024 commented May 11, 2026

Minor: Comment/regex mismatch in _CODE_BLOCK_RE

Uh oh!

lqhl commented May 12, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

lqhl commented May 18, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

Minor: Comment/regex mismatch in `_CODE_BLOCK_RE`

lqhl commented May 12, 2026 •

edited

Loading