feat(drive): add +search shortcut with flat filter flags#658
Merged
fangshuyu-768 merged 1 commit intomainfrom Apr 25, 2026
Merged
feat(drive): add +search shortcut with flat filter flags#658fangshuyu-768 merged 1 commit intomainfrom
fangshuyu-768 merged 1 commit intomainfrom
Conversation
|
Note Reviews pausedIt looks like this branch is under active development. To avoid overwhelming you with review comments due to an influx of new commits, CodeRabbit has automatically paused this review. You can configure this behavior by changing the Use the following commands to manage reviews:
Use the checkboxes below for quick actions:
📝 WalkthroughWalkthroughAdds a new Changes
Sequence DiagramsequenceDiagram
participant CLI as CLI Client
participant Parser as Flag Parser
participant Validator as Validator/Clamper
participant Builder as Request Builder
participant API as Search API
participant Normalizer as Time Normalizer
participant Output as Output Renderer
CLI->>Parser: Parse flags (--query, --mine, --opened-*, --doc-types, etc.)
Parser->>Validator: Validate IDs, doc-types, page-size, mutual exclusivity
alt open_time span > 90 days
Validator->>Validator: Clamp into 90-day slices\nemit stderr continuation notices
end
Validator->>Builder: Produce request spec (filters, slices, page-token)
loop for each slice
Builder->>API: POST /open-apis/search/v2/doc_wiki/search (body)
API-->>Builder: Response (results, res_units, has_more)
end
Builder->>Normalizer: Add recursive *_time_iso fields
Normalizer->>Output: Format table, strip highlights, show "more available" hint
Output->>CLI: Print results and notices
Estimated code review effort🎯 4 (Complex) | ⏱️ ~45 minutes Possibly related PRs
Suggested reviewers
Poem
🚥 Pre-merge checks | ✅ 4 | ❌ 1❌ Failed checks (1 warning)
✅ Passed checks (4 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
github-actions
Bot
added
domain/ccm
liujinkun2025
force-pushed
the
feat/drive-search
branch
from
c568dcc to
fb8c104
Compare
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 2
🧹 Nitpick comments (2)
shortcuts/drive/drive_search.go (2)
505-511: Consider including a timezone indicator in hour-snap notices.The snap notice renders the before/after timestamps with
"2006-01-02 15:04:05"in the process-local timezone and no offset. Agents consuming stderr across machines/timezones can't tell whether16:00:00is UTC or local. Usingtime.RFC3339(or appending the zone) keeps the notice unambiguous and matches the RFC3339 used in the--opened-*slice notice (line 426). Deferring is reasonable if you prefer consistency with other shortcut notice styles in this repo.♻️ Suggested tweak
func formatHourSnapNotice(key, side string, before, after int64) string { return fmt.Sprintf("notice: %s has hour-level granularity server-side; %s %s → %s", key, side, - time.Unix(before, 0).Format("2006-01-02 15:04:05"), - time.Unix(after, 0).Format("2006-01-02 15:04:05"), + time.Unix(before, 0).Format(time.RFC3339), + time.Unix(after, 0).Format(time.RFC3339), ) }🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@shortcuts/drive/drive_search.go` around lines 505 - 511, The hour-snap notice produced by formatHourSnapNotice currently formats before/after with "2006-01-02 15:04:05" which omits timezone info; update formatHourSnapNotice to render the timestamps with an explicit timezone (e.g. use time.RFC3339 or append the zone/offset) so the notice is unambiguous and consistent with the RFC3339 style used by the opened-* slice notices.
326-341: Silent fallback for non-positive--page-sizeis inconsistent with the non-numeric path.
--page-size abcrightly returns a validation error, but--page-size 0or--page-size -5silently becomes 15. Either both should error, or both should clamp, to avoid masking user typos (e.g.,--page-size -20probably means--page-size 20). Given the PR explicitly lists "non-positive auto-fallback to 15" as intended behavior, this is optional — but worth considering anstderrnotice so the user sees what got substituted.🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@shortcuts/drive/drive_search.go` around lines 326 - 341, The function parseDriveSearchPageSize currently errors on non-numeric input but silently falls back to 15 for n <= 0; make behavior consistent by returning a validation error for non-positive values instead of silently substituting 15. Update parseDriveSearchPageSize to use output.ErrValidation when n <= 0 (e.g., "--page-size must be a positive number, got %q") and remove the silent fallback branch; keep the existing max clamp (n > 20 -> n = 20). This touches parseDriveSearchPageSize and reuses output.ErrValidation to surface the error to users.
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Inline comments:
In `@internal/core/types.go`:
- Around line 43-46: The default Open API host was pointed at a pre-prod
endpoint; change the Open constant back to the production URL
("https://open.feishu.cn") so unrecognized brands (via ParseBrand() falling back
to BrandFeishu) use production, remove the TEMP-PPE override comment, and add a
unit test that asserts ParseBrand() (or the default-brand resolution logic) maps
the default/unknown brand to BrandFeishu and that its Open host equals the
production URL to prevent regression.
In `@shortcuts/drive/drive_search.go`:
- Around line 610-625: The regex htmlTagRe in renderDriveSearchTable only strips
<h> tags but the server can also return <hb> / </hb>; update the pattern used to
strip highlights so it matches both single-letter and two-letter highlight tags
(e.g., change htmlTagRe from `</?h>` to a pattern that matches `<h>`, `</h>`,
`<hb>`, `</hb>`), and ensure you apply the same stripping to both
title_highlighted and summary_highlighted processing in renderDriveSearchTable
so no <hb> tags leak into the rendered table.
---
Nitpick comments:
In `@shortcuts/drive/drive_search.go`:
- Around line 505-511: The hour-snap notice produced by formatHourSnapNotice
currently formats before/after with "2006-01-02 15:04:05" which omits timezone
info; update formatHourSnapNotice to render the timestamps with an explicit
timezone (e.g. use time.RFC3339 or append the zone/offset) so the notice is
unambiguous and consistent with the RFC3339 style used by the opened-* slice
notices.
- Around line 326-341: The function parseDriveSearchPageSize currently errors on
non-numeric input but silently falls back to 15 for n <= 0; make behavior
consistent by returning a validation error for non-positive values instead of
silently substituting 15. Update parseDriveSearchPageSize to use
output.ErrValidation when n <= 0 (e.g., "--page-size must be a positive number,
got %q") and remove the silent fallback branch; keep the existing max clamp (n >
20 -> n = 20). This touches parseDriveSearchPageSize and reuses
output.ErrValidation to surface the error to users.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: defaults
Review profile: CHILL
Plan: Pro
Run ID: 9881e466-8f46-4bd2-90d9-07a8354768e6
📒 Files selected for processing (9)
internal/core/types.goshortcuts/drive/drive_search.goshortcuts/drive/drive_search_test.goshortcuts/drive/shortcuts.goshortcuts/drive/shortcuts_test.goskills/lark-doc/SKILL.mdskills/lark-doc/references/lark-doc-search.mdskills/lark-drive/SKILL.mdskills/lark-drive/references/lark-drive-search.md
🚀 PR Preview Install Guide🧰 CLI updatenpm i -g https://pkg.pr.new/larksuite/cli/@larksuite/cli@d9209a1429e3aeb27ace4750990b2ba95cd21147🧩 Skill updatenpx skills add larksuite/cli#feat/drive-search -y -g |
There was a problem hiding this comment.
Actionable comments posted: 1
🧹 Nitpick comments (4)
shortcuts/drive/drive_search.go (3)
326-341:parseDriveSearchPageSizesilently maps0/ negative to the default.
--page-size 0or--page-size -1are very likely typos / bugs rather than "please use default", yet they're silently coerced to 15 with no feedback. Help text documents the range as "1–20". Consider rejecting non-positive values with a validation error so a misconfigured agent gets a clear signal instead of an unrelated page size.♻️ Proposed refactor
n, err := strconv.Atoi(raw) if err != nil { return 0, output.ErrValidation("--page-size must be a number, got %q", raw) } - if n <= 0 { - return 15, nil - } + if n <= 0 { + return 0, output.ErrValidation("--page-size must be positive, got %d", n) + } if n > 20 { n = 20 }🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@shortcuts/drive/drive_search.go` around lines 326 - 341, The function parseDriveSearchPageSize currently treats n <= 0 as a silent default; change it to return a validation error for non-positive values instead. In parseDriveSearchPageSize, when strconv.Atoi succeeds but n <= 0, return a descriptive validation error (e.g., "--page-size must be between 1 and 20, got %d") rather than defaulting to 15; keep the existing clamp for n > 20 (set to 20) and unchanged behavior for empty string defaulting to 15.
580-598: OverwritingDetail.Hintmay drop a server-supplied hint.When
Detail.Code == 99992351, the server-suppliedHint(if any) is replaced outright by the CLI-specific message. If the server ever starts including actionable detail (e.g., the specific offending open_id), that context will be lost. Consider appending the CLI hint (e.g., prefix withserver: …; cli: …) or at least preserving the original when non-empty.🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@shortcuts/drive/drive_search.go` around lines 580 - 598, In enrichDriveSearchError, instead of overwriting exitErr.Detail.Hint when Detail.Code == driveSearchErrUserNotVisible, preserve any existing server hint by appending/prepending the CLI hint (e.g., "cli: …") or combine them (e.g., "server: <original>; cli: <our message>"); update the construction of detail.Hint so it checks exitErr.Detail.Hint and builds a combined string before returning the new *output.ExitError (refer to exitErr, exitErr.Detail.Hint, and driveSearchErrUserNotVisible).
524-570: Verify1m = 30 days/1y = 365 daysis the intended contract.
driveSearchRelativeRetreatsmas 30-day units andyas 365-day units. This matches the help text "7d, 1m, 1y" and comment, but readers/agents may interpret1mas "1 calendar month" or3mas "last quarter". The values pass through to boundary logic (e.g.,3m = 90dis exactly the slice boundary and therefore never triggers clamping, while4m = 120ddoes), which is a sharp cliff worth documenting explicitly in the flag description or tips so agents don't craft edge-case queries unintentionally.🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@shortcuts/drive/drive_search.go` around lines 524 - 570, The code currently treats "m" as 30 days and "y" as 365 days (see driveSearchRelativeRe and parseTimeValue), so explicitly state that contract and the resulting cliff behavior in the code and user-facing help: update the comment above driveSearchRelativeRe/parseTimeValue to say "1m = 30 days, 1y = 365 days (e.g. 3m = 90d)" and update the flag/usage/help text where the search time filter is documented to the same wording and note the boundary edge-case (e.g., 3m == 90d exactly). If instead you want calendar-month or year semantics, change parseTimeValue to compute months/years via time.AddDate and update tests accordingly (referencing parseTimeValue and driveSearchRelativeRe).skills/lark-doc/references/lark-doc-search.md (1)
22-218: Consider adding a pointer todrive +searchflag equivalents in this file too.Since this reference is now in maintenance mode, readers landing here from old links would benefit from a compact
--filterJSON → flat-flag translation table (e.g.,creator_ids→--creator-ids,open_time→--opened-since/--opened-until,folder_tokens→--folder-tokens). This avoids the "deprecated but no migration guide" pitfall during the transition period. Non-blocking.🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@skills/lark-doc/references/lark-doc-search.md` around lines 22 - 218, Add a brief pointer and compact mapping table in the lark-doc-search reference that translates --filter JSON fields (e.g., creator_ids, doc_types, chat_ids, sharer_ids, folder_tokens, space_ids, only_title, only_comment, open_time, create_time) to equivalent flat CLI flags (suggested symbols: --creator-ids, --doc-types, --chat-ids, --sharer-ids, --folder-tokens, --space-ids, --only-title, --only-comment, --opened-since/--opened-until, --created-since/--created-until, plus --page-size/--page-token/--format), and add a short note near the top (around the --filter / 参数 section) pointing readers to the drive +search flag equivalents and the migration guidance when folder_tokens vs space_ids are used; keep it concise and non-blocking for maintenance-mode docs.
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Inline comments:
In `@skills/lark-doc/references/lark-doc-search.md`:
- Around line 4-8: The new standalone blockquote introduced before the existing
blockquote causes MD028; merge them into one blockquote by changing the lone
non-'>' line (the blank or plain line between the two quote blocks) into a
'>'-prefixed line so lines containing the new notice and the existing "前置条件："
paragraph form a single contiguous blockquote; locate the two adjacent quote
fragments in lark-doc-search.md around the inserted notice and make the middle
line start with '>' to remove the MD028 violation.
---
Nitpick comments:
In `@shortcuts/drive/drive_search.go`:
- Around line 326-341: The function parseDriveSearchPageSize currently treats n
<= 0 as a silent default; change it to return a validation error for
non-positive values instead. In parseDriveSearchPageSize, when strconv.Atoi
succeeds but n <= 0, return a descriptive validation error (e.g., "--page-size
must be between 1 and 20, got %d") rather than defaulting to 15; keep the
existing clamp for n > 20 (set to 20) and unchanged behavior for empty string
defaulting to 15.
- Around line 580-598: In enrichDriveSearchError, instead of overwriting
exitErr.Detail.Hint when Detail.Code == driveSearchErrUserNotVisible, preserve
any existing server hint by appending/prepending the CLI hint (e.g., "cli: …")
or combine them (e.g., "server: <original>; cli: <our message>"); update the
construction of detail.Hint so it checks exitErr.Detail.Hint and builds a
combined string before returning the new *output.ExitError (refer to exitErr,
exitErr.Detail.Hint, and driveSearchErrUserNotVisible).
- Around line 524-570: The code currently treats "m" as 30 days and "y" as 365
days (see driveSearchRelativeRe and parseTimeValue), so explicitly state that
contract and the resulting cliff behavior in the code and user-facing help:
update the comment above driveSearchRelativeRe/parseTimeValue to say "1m = 30
days, 1y = 365 days (e.g. 3m = 90d)" and update the flag/usage/help text where
the search time filter is documented to the same wording and note the boundary
edge-case (e.g., 3m == 90d exactly). If instead you want calendar-month or year
semantics, change parseTimeValue to compute months/years via time.AddDate and
update tests accordingly (referencing parseTimeValue and driveSearchRelativeRe).
In `@skills/lark-doc/references/lark-doc-search.md`:
- Around line 22-218: Add a brief pointer and compact mapping table in the
lark-doc-search reference that translates --filter JSON fields (e.g.,
creator_ids, doc_types, chat_ids, sharer_ids, folder_tokens, space_ids,
only_title, only_comment, open_time, create_time) to equivalent flat CLI flags
(suggested symbols: --creator-ids, --doc-types, --chat-ids, --sharer-ids,
--folder-tokens, --space-ids, --only-title, --only-comment,
--opened-since/--opened-until, --created-since/--created-until, plus
--page-size/--page-token/--format), and add a short note near the top (around
the --filter / 参数 section) pointing readers to the drive +search flag
equivalents and the migration guidance when folder_tokens vs space_ids are used;
keep it concise and non-blocking for maintenance-mode docs.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: defaults
Review profile: CHILL
Plan: Pro
Run ID: ec89308b-bed2-4595-8f5e-d9824fcb6c38
📒 Files selected for processing (9)
internal/core/types.goshortcuts/drive/drive_search.goshortcuts/drive/drive_search_test.goshortcuts/drive/shortcuts.goshortcuts/drive/shortcuts_test.goskills/lark-doc/SKILL.mdskills/lark-doc/references/lark-doc-search.mdskills/lark-drive/SKILL.mdskills/lark-drive/references/lark-drive-search.md
✅ Files skipped from review due to trivial changes (2)
- shortcuts/drive/shortcuts.go
- internal/core/types.go
🚧 Files skipped from review as they are similar to previous changes (4)
- shortcuts/drive/shortcuts_test.go
- skills/lark-drive/SKILL.md
- skills/lark-doc/SKILL.md
- skills/lark-drive/references/lark-drive-search.md
zgz2048
reviewed
Apr 24, 2026
liujinkun2025
force-pushed
the
feat/drive-search
branch
from
fb8c104 to
c433df3
Compare
Codecov Report❌ Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## main #658 +/- ##
==========================================
+ Coverage 62.78% 62.95% +0.16%
==========================================
Files 434 436 +2
Lines 38112 38593 +481
==========================================
+ Hits 23928 24295 +367
- Misses 12047 12150 +103
- Partials 2137 2148 +11 ☔ View full report in Codecov by Sentry. 🚀 New features to boost your workflow:
|
There was a problem hiding this comment.
Actionable comments posted: 3
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Inline comments:
In `@shortcuts/drive/drive_search.go`:
- Around line 660-663: The current rawTitle assignment uses fmt.Sprintf("%v",
u["title_highlighted"]) which produces the literal "<nil>" when the key is
missing; change the extraction to the same guarded pattern used for typeStr:
check u["title_highlighted"] for existence and that it's a non-nil string, else
fall back to u["title"] or the empty string, then run
htmlTagRe.ReplaceAllString(...) and common.TruncateStr on that safe string;
update the variables rawTitle/title accordingly (references: rawTitle, title,
htmlTagRe, common.TruncateStr, and u["title_highlighted"]/u["title"]).
In `@skills/lark-drive/references/lark-drive-search.md`:
- Line 78: The doc string for `--page-size` is inaccurate:
`parseDriveSearchPageSize` (in drive_search.go) returns a validation error for
non-numeric input and only clamps non-positive numbers to the default 15, so
update the docs to reflect that non-numeric values cause a validation error and
that only zero/negative values fall back to 15 (and values >20 are clamped to
20). Locate the `--page-size` row in lark-drive-search.md and change the phrase
"非法/非正数自动回落 15" to explicitly state "非数值将返回校验错误；非正数（例如0或负数）回落到15；超过20将被 clamp
到20".
- Line 114: The docs list eight --sort values but the CLI flag uses the enum
variable driveSearchSortValues (in drive_search.go) which only accepts five
values: default, edit_time, edit_time_asc, open_time, create_time; update the
table in lark-drive-search.md to remove create_time_asc,
entity_create_time_desc, and entity_create_time_asc so the documented allowed
values exactly match driveSearchSortValues (or alternately add any missing
values to driveSearchSortValues if you intend to support them).
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: defaults
Review profile: CHILL
Plan: Pro
Run ID: 99d97005-8b23-4c48-a5f4-07a32b89f494
📒 Files selected for processing (8)
shortcuts/drive/drive_search.goshortcuts/drive/drive_search_test.goshortcuts/drive/shortcuts.goshortcuts/drive/shortcuts_test.goskills/lark-doc/SKILL.mdskills/lark-doc/references/lark-doc-search.mdskills/lark-drive/SKILL.mdskills/lark-drive/references/lark-drive-search.md
✅ Files skipped from review due to trivial changes (3)
- shortcuts/drive/shortcuts_test.go
- shortcuts/drive/shortcuts.go
- skills/lark-doc/references/lark-doc-search.md
🚧 Files skipped from review as they are similar to previous changes (1)
- skills/lark-doc/SKILL.md
liujinkun2025
force-pushed
the
feat/drive-search
branch
2 times, most recently
from
1ee107f to
c2d1ac2
Compare
There was a problem hiding this comment.
Actionable comments posted: 1
🧹 Nitpick comments (4)
shortcuts/drive/drive_search.go (2)
678-695: Minor:urlandedit_timecolumns can still render the literal"<nil>"whenresult_metakeys are missing.The earlier title fix (lines 660–666) deliberately scoped the
<nil>guard totitleonly, and the test at line 870 explicitly notes the same%vpattern is still in place forurl/update_time_iso. Since these end up in the pretty table, consider applying the same guarded-extraction pattern (typed assertion + fallback to empty string) here too:♻️ Suggested guard
- url := "" - editTime := "" - if resultMeta != nil { - url = fmt.Sprintf("%v", resultMeta["url"]) - editTime = fmt.Sprintf("%v", resultMeta["update_time_iso"]) - } + var url, editTime string + if resultMeta != nil { + if s, ok := resultMeta["url"].(string); ok { + url = s + } + if s, ok := resultMeta["update_time_iso"].(string); ok { + editTime = s + } + }🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@shortcuts/drive/drive_search.go` around lines 678 - 695, The url and edit_time values are extracted from resultMeta using fmt.Sprintf("%v", ...) which renders "<nil>" when keys are missing; update the extraction in the code that sets url and editTime (before the rows = append(...) call) to perform typed assertions with fallbacks (e.g. if v, ok := resultMeta["url"].(string); ok && v != "" { url = v } else { url = "" } and similarly for resultMeta["update_time_iso"] into editTime) so the pretty table never shows "<nil>".
729-742: Minor: non-deterministic output if a*_timeand matching*_time_isokey coexist in the same map.Map iteration order is random, so for an input like
{"create_time": 1, "create_time_iso": "preserved"}the final value ofout["create_time_iso"]depends on whether the_time_isopassthrough or the_timeannotation runs last. The server doesn't currently emit*_time_isoso this is theoretical, but a small fix is to skip the annotation when the matching_isokey is already present inv:♻️ Suggested guard
out[key] = addDriveSearchIsoTimeFieldsOne(item) if strings.HasSuffix(key, "_time") { + if _, exists := v[key+"_iso"]; exists { + continue + } if iso := unixToISO8601(item); iso != "" { out[key+"_iso"] = iso } }🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@shortcuts/drive/drive_search.go` around lines 729 - 742, The map branch in addDriveSearchIsoTimeFieldsOne iterates v and may overwrite a pre-existing "<key>_iso" entry due to random map iteration; change the logic so when processing a "<key>_time" entry you first check whether the original input map v already contains "<key>_time_iso" and skip creating out["<key>_time_iso"] if it does. Specifically, in the case map[string]interface{} block (variables v, key, item) before calling unixToISO8601(item) for keys with suffix "_time", guard against adding out[key+"_iso"] when v already has that key, preserving any original "<...>_time_iso" value produced by the input.skills/lark-drive/references/lark-drive-search.md (2)
86-86: Nit: clarify--mineopen_id source.The doc reads "从 config 里读 open_id，取不到直接报错", but the implementation actually pulls from
runtime.UserOpenId()(line 116 ofdrive_search.go), which resolves through the auth/session layer rather than a literal config field. A small wording tweak avoids agents going to grepconfig.jsonwhen--mineerrors:📝 Suggested wording
-| `--mine` | `creator_ids = [当前用户 open_id]` | bool。一键"我创建的"；从 config 里读 open_id，取不到直接报错 | +| `--mine` | `creator_ids = [当前用户 open_id]` | bool。一键"我创建的"；从当前登录用户身份（`runtime.UserOpenId()`）解析 open_id，取不到直接报错（提示运行 `lark-cli auth login`） |🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@skills/lark-drive/references/lark-drive-search.md` at line 86, Update the `--mine` docs to state that the open_id is obtained via the auth/session layer by calling runtime.UserOpenId() (used in drive_search.go) rather than read from a static config field; revise the wording for the `--mine` row to reference runtime.UserOpenId() and note that an error will be thrown if that resolver cannot produce an open_id.
139-149: Optional: add a language tag to the bare fenced code blocks.Static analysis flags MD040 (
Fenced code blocks should have a language specified) at lines 139, 146 (this file), and 210. Since these blocks display CLI hints / stderr notices rather than code,textis the conventional choice and quiets the linter without changing rendering:📝 Suggested fix
-``` +```text start: floor 到整点 16:23:45 → 16:00:00 end: ceil 到整点 16:23:45 → 17:00:00@@
-+text
notice: my_edit_time has hour-level granularity server-side;
start 2026-04-22 16:23:00 → 2026-04-22 16:00:00
end 2026-04-22 16:28:00 → 2026-04-22 17:00:00The same change applies to the multi-slice notice block at line 210.
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@skills/lark-drive/references/lark-drive-search.md` around lines 139 - 149, The fenced code blocks showing CLI hints/stderr notices in lark-drive-search.md are missing language tags and trigger MD040; update each bare triple-backtick block (the blocks containing "start: floor 到整点...", the "notice: my_edit_time..." block, and the multi-slice notice block referenced around line 210) by changing ``` to ```text so the linter treats them as plain text and the content/formatting remains unchanged.
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Inline comments:
In `@shortcuts/drive/drive_search.go`:
- Around line 597-604: parseTimeValue currently treats any digit string length
>=10 as seconds, which silently misparses 13-digit epoch-millis into far-future
seconds; update parseTimeValue to mirror unixToISO8601's ms-vs-s heuristic: when
the parsed integer n is >= 1e12 or the digit length is >=13, divide n by 1000
(or otherwise normalize from milliseconds to seconds) before returning, or
alternatively return an explicit error for values that exceed a reasonable
seconds threshold; implement this normalization/rejection in parseTimeValue and
ensure downstream callers like clampOpenedTimeWindow receive normalized epoch
seconds.
---
Nitpick comments:
In `@shortcuts/drive/drive_search.go`:
- Around line 678-695: The url and edit_time values are extracted from
resultMeta using fmt.Sprintf("%v", ...) which renders "<nil>" when keys are
missing; update the extraction in the code that sets url and editTime (before
the rows = append(...) call) to perform typed assertions with fallbacks (e.g. if
v, ok := resultMeta["url"].(string); ok && v != "" { url = v } else { url = "" }
and similarly for resultMeta["update_time_iso"] into editTime) so the pretty
table never shows "<nil>".
- Around line 729-742: The map branch in addDriveSearchIsoTimeFieldsOne iterates
v and may overwrite a pre-existing "<key>_iso" entry due to random map
iteration; change the logic so when processing a "<key>_time" entry you first
check whether the original input map v already contains "<key>_time_iso" and
skip creating out["<key>_time_iso"] if it does. Specifically, in the case
map[string]interface{} block (variables v, key, item) before calling
unixToISO8601(item) for keys with suffix "_time", guard against adding
out[key+"_iso"] when v already has that key, preserving any original
"<...>_time_iso" value produced by the input.
In `@skills/lark-drive/references/lark-drive-search.md`:
- Line 86: Update the `--mine` docs to state that the open_id is obtained via
the auth/session layer by calling runtime.UserOpenId() (used in drive_search.go)
rather than read from a static config field; revise the wording for the `--mine`
row to reference runtime.UserOpenId() and note that an error will be thrown if
that resolver cannot produce an open_id.
- Around line 139-149: The fenced code blocks showing CLI hints/stderr notices
in lark-drive-search.md are missing language tags and trigger MD040; update each
bare triple-backtick block (the blocks containing "start: floor 到整点...", the
"notice: my_edit_time..." block, and the multi-slice notice block referenced
around line 210) by changing ``` to ```text so the linter treats them as plain
text and the content/formatting remains unchanged.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: defaults
Review profile: CHILL
Plan: Pro
Run ID: 01a96b7a-443f-462b-8070-57733b74ad0d
📒 Files selected for processing (9)
shortcuts/drive/drive_search.goshortcuts/drive/drive_search_test.goshortcuts/drive/shortcuts.goshortcuts/drive/shortcuts_test.goskills/lark-doc/SKILL.mdskills/lark-doc/references/lark-doc-search.mdskills/lark-drive/SKILL.mdskills/lark-drive/references/lark-drive-search.mdtests/cli_e2e/drive/drive_search_dryrun_test.go
✅ Files skipped from review due to trivial changes (2)
- shortcuts/drive/shortcuts.go
- skills/lark-doc/references/lark-doc-search.md
🚧 Files skipped from review as they are similar to previous changes (1)
- shortcuts/drive/shortcuts_test.go
liujinkun2025
force-pushed
the
feat/drive-search
branch
3 times, most recently
from
9538808 to
286a48b
Compare
Expose doc_wiki/search v2 under the drive domain via explicit flags (--query, --edited-since, --commented-since, --opened-since, --created-since, --mine, --creator-ids, --doc-types, --folder-tokens, --space-ids, ...) instead of a nested JSON filter, so natural-language queries from AI agents map 1:1 to discrete flags. Time handling: - my_edit_time and my_comment_time are snapped to the hour (floor/ceil) with a stderr notice, since those fields are aggregated at hour granularity server-side. create_time passes through as-is. - open_time has a server-side 3-month cap per request. When --opened-since / --opened-until span exceeds 90 days, the CLI narrows the request to the most recent 90-day slice and emits a stderr notice listing every remaining slice's --opened-* values so the agent can re-invoke for older ranges. Spans over 365 days are rejected up front to bound runaway slicing. Flag ergonomics: - --doc-types accepts mixed case; values are normalized to upper case before validation and before being sent to the server. - --sort default is translated to the server enum DEFAULT_TYPE (every other sort value upper-cases 1:1). Error hints: - Lark code 99992351 (referenced open_id outside the app's contact visibility) is enriched with a +search-specific hint that distinguishes API scope from contact visibility and points at --creator-ids / --sharer-ids as the likely source. Skill docs: - new reference at skills/lark-drive/references/lark-drive-search.md, including the open_time slicing protocol and the paginate-within- slice-before-switching agent playbook. - lark-drive/SKILL.md routes resource-discovery to drive +search. - lark-doc/SKILL.md and lark-doc-search.md mark docs +search as deprecated and point users at drive +search. Change-Id: I36d620045809b448446d4fdbdfa923b05794da19
liujinkun2025
force-pushed
the
feat/drive-search
branch
from
286a48b to
d9209a1
Compare
fangshuyu-768
approved these changes
Apr 25, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
drive +searchshortcut that exposes thedoc_wiki/searchv2 API through flat per-field flags (--edited-since,--opened-since,--mine,--creator-ids,--doc-types,--folder-tokens,--space-ids, etc.) instead of a nested JSON filter, so natural-language agent prompts map 1:1 to CLI flags.open_time3-month cap automatically: when--opened-since/--opened-untilspan exceeds 90 days, the CLI narrows the current request to the most recent 90-day slice and emits a stderr notice listing every remaining slice's--opened-*values so the agent can walk older ranges deterministically. Spans beyond 365 days are rejected up front to bound runaway slicing.99992351— a referencedopen_idfalls outside the app's contact visibility) with a+search-specific hint that distinguishes API scope from contact visibility and points at--creator-ids/--sharer-idsas the likely source.--doc-typesaccepts mixed case and normalizes to upper case;--sort defaultis translated to the server enumDEFAULT_TYPE(all other sort values upper-case 1:1).docs +searchin skill docs; routes all resource discovery (docs / wikis / sheets / bitable / folders) todrive +search, and includes a paginate-within-slice-before-switching playbook for agents.Test plan
go test ./shortcuts/drive/... ./shortcuts/common/... ./internal/output/...passes--opened-since 2mproduces no notice;--opened-since 8memits a 3-slice notice and the clamped request body covers exactly 90 days;--opened-since 2yfails validation before any API calllark-cli drive +search --as user --query '测试' --page-size 3returns resultslark-cli drive +search --as user --opened-since 1y --page-size 3returns slice 1 results plus a 5-slice notice; feeding the slice 2 flag values back in returns a different window's results with no re-notice--doc-types docx,SHEET,Bitable(mixed case) succeeds; the request carries uppercase values--sort defaultreaches the server asDEFAULT_TYPE(the previous literalDEFAULTis rejected by the server enum)Summary by CodeRabbit
New Features
Behavioral Improvements
Documentation
Tests