Conversation
…atches and normalized then merge
📝 WalkthroughWalkthroughThis pull request enriches search documents with exact and normalized morphology fields, implements a two-pass exact+fuzzy search in the search library, adapts tokenization for exact Arabic fields, updates hydrate/rehydrate document transformation across web/mobile, refactors web search hooks to use the centralized search, and adds a session-backed tag combobox and suggested-tags atom. Changes
Sequence Diagram(s)sequenceDiagram
participant Client as useSearch Hook / UI
participant SearchFn as searchDictionary()
participant ExactPass as Exact Pass<br/>(EXACT_PROPERTIES)
participant FuzzyPass as Normalized Pass<br/>(NORMALIZED_PROPERTIES + BOOST)
participant Merge as Merge & Deduplicate
Client->>SearchFn: search(term, { limit?, offset?, language? })
alt term is empty
SearchFn->>FuzzyPass: single search (normalized)
FuzzyPass-->>SearchFn: fuzzy results
else
SearchFn->>ExactPass: first pass (exact fields, low tolerance)
ExactPass-->>SearchFn: exact results
SearchFn->>FuzzyPass: second pass (normalized fields, higher tolerance + boost)
FuzzyPass-->>SearchFn: fuzzy results
end
SearchFn->>Merge: combine exact + fuzzy
Merge->>Merge: deduplicate by id, apply offset/limit
Merge-->>SearchFn: paginated results
SearchFn-->>Client: Results<DictionaryDocument> (includes word_exact & morphology)
Estimated code review effort🎯 4 (Complex) | ⏱️ ~45 minutes Possibly related PRs
Poem
Pre-merge checks and finishing touches❌ Failed checks (1 warning)
✅ Passed checks (2 passed)
✨ Finishing touches
📜 Recent review detailsConfiguration used: defaults Review profile: CHILL Plan: Pro Disabled knowledge base sources:
📒 Files selected for processing (2)
🧰 Additional context used📓 Path-based instructions (4)**/*.{ts,tsx}📄 CodeRabbit inference engine (CLAUDE.md)
Files:
**/*.{ts,tsx,js,jsx}📄 CodeRabbit inference engine (CLAUDE.md)
Files:
apps/web/**/*.{ts,tsx}📄 CodeRabbit inference engine (CLAUDE.md)
Files:
apps/{web,mobile}/**/*.{ts,tsx}📄 CodeRabbit inference engine (CLAUDE.md)
Files:
🧠 Learnings (5)📓 Common learnings📚 Learning: 2025-12-28T19:33:44.247ZApplied to files:
📚 Learning: 2025-11-27T06:48:31.996ZApplied to files:
📚 Learning: 2025-11-27T06:48:57.365ZApplied to files:
📚 Learning: 2025-11-27T06:49:36.986ZApplied to files:
🧬 Code graph analysis (1)packages/search/src/database.ts (3)
⏰ Context from checks skipped due to timeout of 90000ms. You can increase the timeout in your CodeRabbit configuration to a maximum of 15 minutes (900000ms). (1)
🔇 Additional comments (8)
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 |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 5
🧹 Nitpick comments (5)
packages/search/src/arabic.ts (1)
23-31: Good deprecation documentation, but consider removing the export.The deprecation warning is clear and well-documented. However, since this function is marked as deprecated and "not used in search normalization," you might consider:
- Removing it from exports to prevent accidental usage
- Or keeping it unexported for potential internal use
If external consumers might still need it, the current approach is acceptable.
apps/web/src/lib/search/index.ts (2)
139-164: Consider extracting the morphology mapping to a reusable helper.This morphology transformation logic is duplicated in
rehydrateOramaDb(lines 253-278). Extracting it to a helper function would improve maintainability.🔎 Proposed helper extraction
// Add at file level const mapMorphologyForIndex = ( morphology: z.infer<typeof MorphologySchema> | null | undefined ): DictionaryDocument["morphology"] => { if (!morphology) return undefined; return { ism: morphology.ism ? { singular: morphology.ism.singular, plurals: morphology.ism.plurals?.map((p) => p.word), singular_exact: morphology.ism.singular, plurals_exact: morphology.ism.plurals?.map((p) => p.word), } : undefined, verb: morphology.verb ? { past_tense: morphology.verb.past_tense, present_tense: morphology.verb.present_tense, masadir: morphology.verb.masadir?.map((m) => m.word), past_tense_exact: morphology.verb.past_tense, present_tense_exact: morphology.verb.present_tense, masadir_exact: morphology.verb.masadir?.map((m) => m.word), } : undefined, }; };Then use:
morphology: mapMorphologyForIndex(morphology),
229-240: Rehydration silently skips morphology parsing errors unlike hydration.In
hydrateOramaDb, morphology validation failures are logged to Sentry (lines 108-117), butrehydrateOramaDbdoesn't log these errors. Consider adding similar logging for consistency.apps/mobile/src/lib/search/index.ts (1)
85-141: Morphology mapping duplicated across mobile and web apps.The morphology transformation logic (lines 118-141 and 282-305) is identical to the web app implementation. Consider extracting this to a shared utility in
@bahar/searchor@bahar/db-operationsto avoid drift between platforms.packages/search/src/tokenizer.ts (1)
48-58: EXACT_PROPS duplicated across multiple files.This list is also defined in:
packages/search/src/database.ts(EXACT_PROPERTIES)apps/web/src/hooks/useSearch.ts(EXACT_PROPERTIES)Consider exporting this from a single location to ensure consistency.
🔎 Proposed consolidation
Export from
packages/search/src/schema.ts:export const EXACT_SEARCH_PROPERTIES = [ "word_exact", "morphology.ism.singular_exact", "morphology.ism.plurals_exact", "morphology.verb.past_tense_exact", "morphology.verb.present_tense_exact", "morphology.verb.masadir_exact", ] as const; export const NORMALIZED_SEARCH_PROPERTIES = [ "word", "translation", "definition", "tags", "morphology.ism.plurals", "morphology.ism.singular", "morphology.verb.masadir", "morphology.verb.past_tense", "morphology.verb.present_tense", ] as const;Then import where needed.
📜 Review details
Configuration used: defaults
Review profile: CHILL
Plan: Pro
Disabled knowledge base sources:
- Linear integration is disabled by default for public repositories
You can enable these sources in your CodeRabbit configuration.
📒 Files selected for processing (7)
apps/mobile/src/lib/search/index.tsapps/web/src/hooks/useSearch.tsapps/web/src/lib/search/index.tspackages/search/src/arabic.tspackages/search/src/database.tspackages/search/src/schema.tspackages/search/src/tokenizer.ts
🧰 Additional context used
📓 Path-based instructions (5)
**/*.{ts,tsx}
📄 CodeRabbit inference engine (CLAUDE.md)
**/*.{ts,tsx}: Use TypeScript with strict typing across entire codebase
Do not useanytype unless absolutely necessary
Error handling with try/catch blocks and structured error types
Use Drizzle ORM for database operations
UseDisplayErrorclass for user-friendly error messages andResult<T, E>type for explicit error handling (Rust-like pattern)
Files:
packages/search/src/tokenizer.tsapps/mobile/src/lib/search/index.tspackages/search/src/database.tspackages/search/src/arabic.tsapps/web/src/lib/search/index.tsapps/web/src/hooks/useSearch.tspackages/search/src/schema.ts
**/*.{ts,tsx,js,jsx}
📄 CodeRabbit inference engine (CLAUDE.md)
**/*.{ts,tsx,js,jsx}: Write self-documenting code and avoid overuse of comments
Component naming: PascalCase for components, camelCase for functions/variables
Files:
packages/search/src/tokenizer.tsapps/mobile/src/lib/search/index.tspackages/search/src/database.tspackages/search/src/arabic.tsapps/web/src/lib/search/index.tsapps/web/src/hooks/useSearch.tspackages/search/src/schema.ts
apps/mobile/**/*.{ts,tsx}
📄 CodeRabbit inference engine (CLAUDE.md)
apps/mobile/**/*.{ts,tsx}: Mobile app uses UniWind (Tailwind for React Native) with Tailwind CSS v4
Mobile app uses Expo with file-based routing (Expo Router)
Files:
apps/mobile/src/lib/search/index.ts
apps/{web,mobile}/**/*.{ts,tsx}
📄 CodeRabbit inference engine (CLAUDE.md)
Use Lingui (v5) for internationalization in both web and mobile applications
Files:
apps/mobile/src/lib/search/index.tsapps/web/src/lib/search/index.tsapps/web/src/hooks/useSearch.ts
apps/web/**/*.{ts,tsx}
📄 CodeRabbit inference engine (CLAUDE.md)
apps/web/**/*.{ts,tsx}: Web app uses Tanstack Router for client-side routing
Orama is used for client-side WASM search engine with multi-language support (Arabic + English), indexed from local database on app initialization
Files:
apps/web/src/lib/search/index.tsapps/web/src/hooks/useSearch.ts
🧠 Learnings (7)
📓 Common learnings
Learnt from: CR
Repo: Shunseii/bahar PR: 0
File: CLAUDE.md:0-0
Timestamp: 2025-12-28T19:33:44.247Z
Learning: Applies to apps/web/**/*.{ts,tsx} : Orama is used for client-side WASM search engine with multi-language support (Arabic + English), indexed from local database on app initialization
📚 Learning: 2025-11-27T06:51:53.688Z
Learnt from: Shunseii
Repo: Shunseii/bahar PR: 24
File: apps/web/src/lib/search/orama-tokenizer.ts:31-31
Timestamp: 2025-11-27T06:51:53.688Z
Learning: In Orama (search library), custom tokenizers must include a `normalizationCache: Map<string, string>` property as part of the required tokenizer interface, even if not explicitly used in the tokenize method implementation. This is part of Orama's documented API contract for custom tokenizers.
Applied to files:
packages/search/src/tokenizer.tspackages/search/src/schema.ts
📚 Learning: 2025-12-28T19:33:44.247Z
Learnt from: CR
Repo: Shunseii/bahar PR: 0
File: CLAUDE.md:0-0
Timestamp: 2025-12-28T19:33:44.247Z
Learning: Applies to apps/web/**/*.{ts,tsx} : Orama is used for client-side WASM search engine with multi-language support (Arabic + English), indexed from local database on app initialization
Applied to files:
packages/search/src/tokenizer.tsapps/mobile/src/lib/search/index.tspackages/search/src/database.tsapps/web/src/lib/search/index.tsapps/web/src/hooks/useSearch.tspackages/search/src/schema.ts
📚 Learning: 2025-11-27T06:48:31.996Z
Learnt from: Shunseii
Repo: Shunseii/bahar PR: 24
File: apps/web/src/hooks/db/index.ts:59-60
Timestamp: 2025-11-27T06:48:31.996Z
Learning: In the Bahar project, Orama methods (insert, update, remove, search) are synchronous by default since no async plugins are being used. Orama v3.0.0+ core methods are sync; they only become async when specific plugins like plugin-embeddings are added. Do not suggest awaiting Orama operations unless async plugins are confirmed to be in use.
Applied to files:
packages/search/src/database.ts
📚 Learning: 2025-11-27T06:48:57.365Z
Learnt from: Shunseii
Repo: Shunseii/bahar PR: 24
File: apps/web/src/hooks/db/index.ts:84-85
Timestamp: 2025-11-27T06:48:57.365Z
Learning: In the Orama search library, core operations like insert(), remove(), and update() are synchronous by default. They only become asynchronous when specific plugins (e.g., embedding generators) or Orama Cloud features are used. Don't suggest awaiting these operations unless async plugins are confirmed to be in use.
Applied to files:
packages/search/src/database.ts
📚 Learning: 2025-11-27T06:49:36.986Z
Learnt from: Shunseii
Repo: Shunseii/bahar PR: 24
File: apps/web/src/hooks/db/index.ts:114-115
Timestamp: 2025-11-27T06:49:36.986Z
Learning: In the bahar project, Orama database operations (insert, update, remove) are synchronous because only the plugin-qps plugin is used, which is a local scoring plugin. These operations should not be awaited.
Applied to files:
packages/search/src/database.ts
📚 Learning: 2025-12-25T22:07:11.870Z
Learnt from: Shunseii
Repo: Shunseii/bahar PR: 27
File: apps/api/src/db/schema/auth.ts:34-36
Timestamp: 2025-12-25T22:07:11.870Z
Learning: Files in apps/api/src/db/schema/auth.ts are auto-generated by better-auth and should not be manually modified as changes will be overwritten. Any schema issues should be tested and reported upstream to better-auth.
Applied to files:
packages/search/src/schema.ts
🧬 Code graph analysis (4)
packages/search/src/tokenizer.ts (1)
packages/search/src/arabic.ts (2)
stripArabicDiacritics(9-11)normalizeArabicForSearch(40-42)
apps/mobile/src/lib/search/index.ts (2)
packages/db-operations/src/converters.ts (1)
safeJsonParse(31-52)packages/drizzle-user-db-schemas/src/types.ts (1)
MorphologySchema(31-77)
apps/web/src/lib/search/index.ts (2)
packages/db-operations/src/converters.ts (1)
safeJsonParse(31-52)packages/drizzle-user-db-schemas/src/types.ts (1)
MorphologySchema(31-77)
apps/web/src/hooks/useSearch.ts (3)
packages/search/src/schema.ts (1)
DictionaryDocument(66-82)apps/mobile/src/lib/search/index.ts (1)
getOramaDb(37-42)apps/web/src/lib/utils.ts (1)
stripArabicDiacritics(120-123)
⏰ Context from checks skipped due to timeout of 90000ms. You can increase the timeout in your CodeRabbit configuration to a maximum of 15 minutes (900000ms). (1)
- GitHub Check: Cloudflare Pages
🔇 Additional comments (9)
packages/search/src/arabic.ts (1)
33-42: Normalization simplification looks correct.The removal of weak-letter normalization aligns with the two-pass search strategy. Preserving hamza variants and weak letters in exact fields enables better ranking of precise matches.
apps/web/src/lib/search/index.ts (1)
104-117: LGTM - Good conditional logging for morphology validation.The check
!morphologyResult.ok && entry.morphologycorrectly avoids logging when morphology is simply null/undefined (expected case).packages/search/src/database.ts (2)
99-137: Well-structured search configuration constants.The separation of exact and normalized properties with associated boost values is clean and maintainable.
154-162: Empty term search bypasses offset parameter.When
termis empty, the search correctly falls back to a simple search, butoptions?.propertiesis also ignored. This may be intentional for listing all entries, but worth noting for API consistency.apps/mobile/src/lib/search/index.ts (1)
90-99: Inconsistent skip logic between rehydrate and hydrate.In
rehydrateOramaDb, entries are skipped if any ofroot,tags,antonyms, orexamplesfail parsing (lines 90-99). However, inhydrateOramaDb, the same validation includesmorphologyResultin the error logging but not in the skip condition (lines 216-222). The skip logic is consistent, which is good.packages/search/src/tokenizer.ts (1)
23-31: Good separation of exact tokenizer.The exact tokenizer correctly disables stemming while still tokenizing Arabic text, enabling precise matching on
*_exactfields.packages/search/src/schema.ts (3)
11-39: Well-documented schema extension for exact matching.The JSDoc clearly explains the purpose of
_exactfields, and the schema structure logically separates normalized and exact variants.
41-60: IndexedMorphology interface is well-structured.The flattened structure correctly mirrors the schema fields and provides proper typing for both normalized and exact variants.
66-82: DictionaryDocument correctly updated with new fields.The addition of
word_exactand updatedmorphologytype aligns with the schema changes.
Deploying bahar with
|
| Latest commit: |
b95a350
|
| Status: | ✅ Deploy successful! |
| Preview URL: | https://71ccc1d1.bahar-5xu.pages.dev |
| Branch Preview URL: | https://refactor-two-pass-orama-sear.bahar-5xu.pages.dev |
There was a problem hiding this comment.
Actionable comments posted: 4
Caution
Some comments are outside the diff and can’t be posted inline due to platform limitations.
⚠️ Outside diff range comments (2)
apps/web/src/hooks/useSearch.ts (2)
137-153: MissingsearchQueryLanguagein dependency array.The
useEffectat line 137 usessearchQueryLanguagebut it's not in the dependency array. This could cause stale language values when the search term's language changes:🔎 Proposed fix
useEffect(() => { // Don't search when offset is 0 since it // was already handled in the other useEffect if (offset === 0) return; const { hits } = search( { ...params, offset, }, searchQueryLanguage ); setHits((previousHits) => previousHits ? [...previousHits, ...hits] : hits ); - }, [offset, setHits, search]); + }, [offset, setHits, search, searchQueryLanguage]);
155-170: MissingsearchQueryLanguagein dependency array.Similar to the previous effect, this
useEffectusessearchQueryLanguagebut omits it from dependencies:🔎 Proposed fix
useEffect(() => { const { hits, ...metadata } = search( { ...params, offset: 0, }, searchQueryLanguage ); setOffset(0); setHits(hits); setSearchResultsMetadata(metadata); setHasMore(hits.length < metadata.count); - }, [paramsKey, setOffset, setHits, setSearchResultsMetadata, search]); + }, [paramsKey, setOffset, setHits, setSearchResultsMetadata, search, searchQueryLanguage]);
🧹 Nitpick comments (5)
apps/web/src/hooks/db/index.ts (1)
40-42: Consider updating suggested tags on edit as well.The
useEditDictionaryEntryhook doesn't updatesuggestedTagsAtomwhen editing a word with tags, whileuseAddDictionaryEntrydoes. If this inconsistency is intentional (only track tags from new entries), consider adding a brief comment. Otherwise, you may want to apply the same logic in the edit flow.apps/web/src/lib/search/index.ts (1)
174-214: Consider extracting shared morphology transformation logic.The morphology transformation logic (lines 189-214) is duplicated in
toOramaDocument,hydrateOramaDb, andrehydrateOramaDb. While the input types differ (SelectDictionaryEntryvs parsedMorphologySchema), the transformation logic is identical once morphology is parsed.Consider extracting a helper function that transforms the parsed morphology object:
🔎 Proposed helper
const transformMorphology = (morphology: z.infer<typeof MorphologySchema> | undefined) => { if (!morphology) return undefined; return { ism: morphology.ism ? { singular: morphology.ism.singular, plurals: morphology.ism.plurals?.map((p) => p.word), singular_exact: morphology.ism.singular, plurals_exact: morphology.ism.plurals?.map((p) => p.word), } : undefined, verb: morphology.verb ? { past_tense: morphology.verb.past_tense, present_tense: morphology.verb.present_tense, masadir: morphology.verb.masadir?.map((m) => m.word), past_tense_exact: morphology.verb.past_tense, present_tense_exact: morphology.verb.present_tense, masadir_exact: morphology.verb.masadir?.map((m) => m.word), } : undefined, }; };apps/web/src/components/TagsCombobox.tsx (3)
268-277: Fragile timeout pattern in onBlur.The 150ms
setTimeoutdelay to allow dropdown clicks is a common workaround but can be unreliable on slower devices or under load. Consider usingonMouseDownwithpreventDefault()on dropdown options instead, which prevents the blur event entirely.🔎 Alternative approach
On the dropdown option elements, use
onMouseDowninstead ofonClick:onMouseDown={(e) => { e.preventDefault(); // Prevents input blur selectOption(optionValue); }}This eliminates the need for the setTimeout workaround.
309-312: Hardcoded "Searching..." string should be internationalized.Per coding guidelines, Lingui is used for i18n. This hardcoded string should use
<Trans>ort:🔎 Proposed fix
+import { Trans } from "@lingui/react/macro"; // ... {isSearching ? ( <div className="py-6 text-center text-muted-foreground text-sm"> - Searching... + <Trans>Searching...</Trans> </div>
370-372: Default "Add" text should also be internationalized.The fallback
Add "{trimmedInput}"text should use Lingui:🔎 Proposed fix
{renderCreateOption ? ( renderCreateOption(trimmedInput) ) : ( - <>Add "{trimmedInput}"</> + <Trans>Add "{trimmedInput}"</Trans> )}
📜 Review details
Configuration used: defaults
Review profile: CHILL
Plan: Pro
Disabled knowledge base sources:
- Linear integration is disabled by default for public repositories
You can enable these sources in your CodeRabbit configuration.
⛔ Files ignored due to path filters (1)
pnpm-lock.yamlis excluded by!**/pnpm-lock.yaml
📒 Files selected for processing (10)
apps/web/src/atoms/suggested-tags.tsapps/web/src/components/TagsCombobox.tsxapps/web/src/components/features/dictionary/add/TagsFormSection.tsxapps/web/src/hooks/db/index.tsapps/web/src/hooks/useSearch.tsapps/web/src/lib/search/index.tsapps/web/src/lib/utils.tsapps/web/src/routes/_authorized-layout/_app-layout/dictionary/edit/$wordId.tsxbiome.jsoncpackages/search/src/database.ts
💤 Files with no reviewable changes (1)
- apps/web/src/lib/utils.ts
🧰 Additional context used
📓 Path-based instructions (6)
**/*.{ts,tsx}
📄 CodeRabbit inference engine (CLAUDE.md)
**/*.{ts,tsx}: Use TypeScript with strict typing across entire codebase
Do not useanytype unless absolutely necessary
Error handling with try/catch blocks and structured error types
Use Drizzle ORM for database operations
UseDisplayErrorclass for user-friendly error messages andResult<T, E>type for explicit error handling (Rust-like pattern)
Files:
apps/web/src/atoms/suggested-tags.tsapps/web/src/routes/_authorized-layout/_app-layout/dictionary/edit/$wordId.tsxapps/web/src/hooks/useSearch.tsapps/web/src/components/features/dictionary/add/TagsFormSection.tsxapps/web/src/hooks/db/index.tsapps/web/src/components/TagsCombobox.tsxapps/web/src/lib/search/index.tspackages/search/src/database.ts
**/*.{ts,tsx,js,jsx}
📄 CodeRabbit inference engine (CLAUDE.md)
**/*.{ts,tsx,js,jsx}: Write self-documenting code and avoid overuse of comments
Component naming: PascalCase for components, camelCase for functions/variables
Files:
apps/web/src/atoms/suggested-tags.tsapps/web/src/routes/_authorized-layout/_app-layout/dictionary/edit/$wordId.tsxapps/web/src/hooks/useSearch.tsapps/web/src/components/features/dictionary/add/TagsFormSection.tsxapps/web/src/hooks/db/index.tsapps/web/src/components/TagsCombobox.tsxapps/web/src/lib/search/index.tspackages/search/src/database.ts
apps/web/**/*.{ts,tsx}
📄 CodeRabbit inference engine (CLAUDE.md)
apps/web/**/*.{ts,tsx}: Web app uses Tanstack Router for client-side routing
Orama is used for client-side WASM search engine with multi-language support (Arabic + English), indexed from local database on app initialization
Files:
apps/web/src/atoms/suggested-tags.tsapps/web/src/routes/_authorized-layout/_app-layout/dictionary/edit/$wordId.tsxapps/web/src/hooks/useSearch.tsapps/web/src/components/features/dictionary/add/TagsFormSection.tsxapps/web/src/hooks/db/index.tsapps/web/src/components/TagsCombobox.tsxapps/web/src/lib/search/index.ts
apps/{web,mobile}/**/*.{ts,tsx}
📄 CodeRabbit inference engine (CLAUDE.md)
Use Lingui (v5) for internationalization in both web and mobile applications
Files:
apps/web/src/atoms/suggested-tags.tsapps/web/src/routes/_authorized-layout/_app-layout/dictionary/edit/$wordId.tsxapps/web/src/hooks/useSearch.tsapps/web/src/components/features/dictionary/add/TagsFormSection.tsxapps/web/src/hooks/db/index.tsapps/web/src/components/TagsCombobox.tsxapps/web/src/lib/search/index.ts
**/*.{tsx,jsx}
📄 CodeRabbit inference engine (CLAUDE.md)
**/*.{tsx,jsx}: React components use functional style with hooks
Prefer using jotai atoms over React Context
State management: use Tanstack Query for async state and Jotai for atomic state
Use Tanstack Query for server state management
Files:
apps/web/src/routes/_authorized-layout/_app-layout/dictionary/edit/$wordId.tsxapps/web/src/components/features/dictionary/add/TagsFormSection.tsxapps/web/src/components/TagsCombobox.tsx
apps/web/**/*.{tsx,jsx}
📄 CodeRabbit inference engine (CLAUDE.md)
Web app uses Shadcn/UI components and Tailwind CSS v4 for styling, use the
cn()utility function for combining and conditionally applying Tailwind classes
Files:
apps/web/src/routes/_authorized-layout/_app-layout/dictionary/edit/$wordId.tsxapps/web/src/components/features/dictionary/add/TagsFormSection.tsxapps/web/src/components/TagsCombobox.tsx
🧠 Learnings (11)
📓 Common learnings
Learnt from: CR
Repo: Shunseii/bahar PR: 0
File: CLAUDE.md:0-0
Timestamp: 2025-12-28T19:33:44.247Z
Learning: Applies to apps/web/**/*.{ts,tsx} : Orama is used for client-side WASM search engine with multi-language support (Arabic + English), indexed from local database on app initialization
📚 Learning: 2025-12-28T19:33:44.247Z
Learnt from: CR
Repo: Shunseii/bahar PR: 0
File: CLAUDE.md:0-0
Timestamp: 2025-12-28T19:33:44.247Z
Learning: Applies to **/*.{tsx,jsx} : Prefer using jotai atoms over React Context
Applied to files:
apps/web/src/atoms/suggested-tags.tsapps/web/src/hooks/db/index.ts
📚 Learning: 2025-12-28T19:33:44.247Z
Learnt from: CR
Repo: Shunseii/bahar PR: 0
File: CLAUDE.md:0-0
Timestamp: 2025-12-28T19:33:44.247Z
Learning: Applies to apps/web/**/*.{ts,tsx} : Orama is used for client-side WASM search engine with multi-language support (Arabic + English), indexed from local database on app initialization
Applied to files:
apps/web/src/hooks/useSearch.tsapps/web/src/hooks/db/index.tsapps/web/src/lib/search/index.tspackages/search/src/database.ts
📚 Learning: 2025-12-28T19:33:44.247Z
Learnt from: CR
Repo: Shunseii/bahar PR: 0
File: CLAUDE.md:0-0
Timestamp: 2025-12-28T19:33:44.247Z
Learning: Applies to **/*.{tsx,jsx} : State management: use Tanstack Query for async state and Jotai for atomic state
Applied to files:
apps/web/src/hooks/db/index.ts
📚 Learning: 2025-12-28T19:33:44.247Z
Learnt from: CR
Repo: Shunseii/bahar PR: 0
File: CLAUDE.md:0-0
Timestamp: 2025-12-28T19:33:44.247Z
Learning: Applies to **/*.{tsx,jsx} : Use Tanstack Query for server state management
Applied to files:
apps/web/src/hooks/db/index.ts
📚 Learning: 2025-11-27T06:48:57.365Z
Learnt from: Shunseii
Repo: Shunseii/bahar PR: 24
File: apps/web/src/hooks/db/index.ts:84-85
Timestamp: 2025-11-27T06:48:57.365Z
Learning: In the Orama search library, core operations like insert(), remove(), and update() are synchronous by default. They only become asynchronous when specific plugins (e.g., embedding generators) or Orama Cloud features are used. Don't suggest awaiting these operations unless async plugins are confirmed to be in use.
Applied to files:
apps/web/src/hooks/db/index.tspackages/search/src/database.ts
📚 Learning: 2025-12-28T19:33:44.247Z
Learnt from: CR
Repo: Shunseii/bahar PR: 0
File: CLAUDE.md:0-0
Timestamp: 2025-12-28T19:33:44.247Z
Learning: Applies to **/*.{tsx,jsx} : React components use functional style with hooks
Applied to files:
apps/web/src/components/TagsCombobox.tsx
📚 Learning: 2025-11-27T06:48:31.996Z
Learnt from: Shunseii
Repo: Shunseii/bahar PR: 24
File: apps/web/src/hooks/db/index.ts:59-60
Timestamp: 2025-11-27T06:48:31.996Z
Learning: In the Bahar project, Orama methods (insert, update, remove, search) are synchronous by default since no async plugins are being used. Orama v3.0.0+ core methods are sync; they only become async when specific plugins like plugin-embeddings are added. Do not suggest awaiting Orama operations unless async plugins are confirmed to be in use.
Applied to files:
apps/web/src/lib/search/index.tspackages/search/src/database.ts
📚 Learning: 2025-12-28T19:33:44.247Z
Learnt from: CR
Repo: Shunseii/bahar PR: 0
File: CLAUDE.md:0-0
Timestamp: 2025-12-28T19:33:44.247Z
Learning: Applies to **/*.{ts,tsx} : Use TypeScript with strict typing across entire codebase
Applied to files:
biome.jsonc
📚 Learning: 2025-12-28T19:33:44.247Z
Learnt from: CR
Repo: Shunseii/bahar PR: 0
File: CLAUDE.md:0-0
Timestamp: 2025-12-28T19:33:44.247Z
Learning: Applies to **/*.{ts,tsx} : Do not use `any` type unless absolutely necessary
Applied to files:
biome.jsonc
📚 Learning: 2025-11-27T06:49:36.986Z
Learnt from: Shunseii
Repo: Shunseii/bahar PR: 24
File: apps/web/src/hooks/db/index.ts:114-115
Timestamp: 2025-11-27T06:49:36.986Z
Learning: In the bahar project, Orama database operations (insert, update, remove) are synchronous because only the plugin-qps plugin is used, which is a local scoring plugin. These operations should not be awaited.
Applied to files:
packages/search/src/database.ts
🧬 Code graph analysis (3)
apps/web/src/hooks/useSearch.ts (4)
packages/search/src/schema.ts (1)
DictionaryDocument(66-82)packages/search/src/database.ts (2)
SearchLanguage(139-139)searchDictionary(147-215)apps/web/src/lib/search/index.ts (1)
getOramaDb(23-23)apps/mobile/src/lib/search/index.ts (1)
getOramaDb(37-42)
apps/web/src/hooks/db/index.ts (3)
apps/web/src/hooks/useSearch.ts (1)
useSearch(26-101)apps/web/src/atoms/suggested-tags.ts (1)
suggestedTagsAtom(3-16)apps/web/src/lib/search/index.ts (2)
getOramaDb(23-23)toOramaDocument(29-72)
apps/web/src/lib/search/index.ts (4)
packages/drizzle-user-db-schemas/src/index.ts (1)
SelectDictionaryEntry(12-12)packages/search/src/schema.ts (1)
DictionaryDocument(66-82)packages/db-operations/src/converters.ts (1)
safeJsonParse(31-52)packages/drizzle-user-db-schemas/src/types.ts (1)
MorphologySchema(31-77)
⏰ Context from checks skipped due to timeout of 90000ms. You can increase the timeout in your CodeRabbit configuration to a maximum of 15 minutes (900000ms). (1)
- GitHub Check: Cloudflare Pages
🔇 Additional comments (8)
apps/web/src/hooks/db/index.ts (1)
66-67: LGTM!The transformation to
toOramaDocumentensures consistency with hydration flows, and the synchronous Oramainsertcall is correct per the project's configuration (no async plugins).apps/web/src/components/features/dictionary/add/TagsFormSection.tsx (2)
19-38: LGTM!Clean implementation using Jotai for state management (per coding guidelines), proper filtering logic for available suggestions, and defensive duplicate checking in
addTag. TheuseFieldArraywithreplaceis appropriate for this use case.
53-62: Good integration with TagsCombobox.The
TagsComboboxintegration is well-structured with proper value extraction, i18n support via Lingui, and clean data transformation inonValueChange.apps/web/src/lib/search/index.ts (1)
25-72: LGTM! Well-documented transformation function.The
toOramaDocumentfunction correctly handles the morphology transformation (DB{word: string}[]→ Oramastring[]) and populates_exactvariants for the two-pass search strategy.packages/search/src/database.ts (2)
157-170: LGTM on empty term handling.The early return for empty terms correctly bypasses the two-pass logic and applies the offset, which is appropriate for browsing/listing scenarios.
210-214: Good fix on count calculation.Using
mergedHits.lengthfor count addresses the previous review comment about potential inaccuracy. This now correctly reflects the actual number of results returned.apps/web/src/components/TagsCombobox.tsx (1)
77-105: LGTM on async data fetching.Good use of cancellation flag to prevent state updates on unmounted components, and
startTransitionfor non-urgent UI updates.apps/web/src/hooks/useSearch.ts (1)
34-46: LGTM! Clean delegation to searchDictionary.The refactor correctly delegates to
searchDictionaryfrom the shared package, eliminating the previously duplicated two-pass search logic. This addresses the past review comments about code duplication.
There was a problem hiding this comment.
Actionable comments posted: 0
Caution
Some comments are outside the diff and can’t be posted inline due to platform limitations.
⚠️ Outside diff range comments (2)
apps/web/src/hooks/useSearch.ts (2)
137-153: Missing dependencies in useEffect.The
useEffectat lines 137-153 referencesparamsandsearchQueryLanguagebut only includes[offset, setHits, search]in its dependency array. This could cause stale closures ifparamsorsearchQueryLanguagechange whileoffsetremains non-zero.🔎 Suggested fix
useEffect(() => { // Don't search when offset is 0 since it // was already handled in the other useEffect if (offset === 0) return; const { hits } = search( { ...params, offset, }, searchQueryLanguage ); setHits((previousHits) => previousHits ? [...previousHits, ...hits] : hits ); - }, [offset, setHits, search]); + }, [offset, setHits, search, paramsKey, searchQueryLanguage]);Note: Using
paramsKeyinstead ofparamsavoids unnecessary re-renders sinceparamsKeyis already a stable string representation.
157-170: MissingsearchQueryLanguagein dependency array.The effect uses
searchQueryLanguagebut it's not listed in dependencies. If the language changes (e.g., user switches from Arabic to English input), the search won't re-execute.🔎 Suggested fix
- }, [paramsKey, setOffset, setHits, setSearchResultsMetadata, search]); + }, [paramsKey, searchQueryLanguage, setOffset, setHits, setSearchResultsMetadata, search]);
📜 Review details
Configuration used: defaults
Review profile: CHILL
Plan: Pro
Disabled knowledge base sources:
- Linear integration is disabled by default for public repositories
You can enable these sources in your CodeRabbit configuration.
📒 Files selected for processing (2)
apps/web/src/hooks/useSearch.tspackages/search/src/database.ts
🧰 Additional context used
📓 Path-based instructions (4)
**/*.{ts,tsx}
📄 CodeRabbit inference engine (CLAUDE.md)
**/*.{ts,tsx}: Use TypeScript with strict typing across entire codebase
Do not useanytype unless absolutely necessary
Error handling with try/catch blocks and structured error types
Use Drizzle ORM for database operations
UseDisplayErrorclass for user-friendly error messages andResult<T, E>type for explicit error handling (Rust-like pattern)
Files:
apps/web/src/hooks/useSearch.tspackages/search/src/database.ts
**/*.{ts,tsx,js,jsx}
📄 CodeRabbit inference engine (CLAUDE.md)
**/*.{ts,tsx,js,jsx}: Write self-documenting code and avoid overuse of comments
Component naming: PascalCase for components, camelCase for functions/variables
Files:
apps/web/src/hooks/useSearch.tspackages/search/src/database.ts
apps/web/**/*.{ts,tsx}
📄 CodeRabbit inference engine (CLAUDE.md)
apps/web/**/*.{ts,tsx}: Web app uses Tanstack Router for client-side routing
Orama is used for client-side WASM search engine with multi-language support (Arabic + English), indexed from local database on app initialization
Files:
apps/web/src/hooks/useSearch.ts
apps/{web,mobile}/**/*.{ts,tsx}
📄 CodeRabbit inference engine (CLAUDE.md)
Use Lingui (v5) for internationalization in both web and mobile applications
Files:
apps/web/src/hooks/useSearch.ts
🧠 Learnings (5)
📓 Common learnings
Learnt from: CR
Repo: Shunseii/bahar PR: 0
File: CLAUDE.md:0-0
Timestamp: 2025-12-28T19:33:44.247Z
Learning: Applies to apps/web/**/*.{ts,tsx} : Orama is used for client-side WASM search engine with multi-language support (Arabic + English), indexed from local database on app initialization
📚 Learning: 2025-12-28T19:33:44.247Z
Learnt from: CR
Repo: Shunseii/bahar PR: 0
File: CLAUDE.md:0-0
Timestamp: 2025-12-28T19:33:44.247Z
Learning: Applies to apps/web/**/*.{ts,tsx} : Orama is used for client-side WASM search engine with multi-language support (Arabic + English), indexed from local database on app initialization
Applied to files:
apps/web/src/hooks/useSearch.tspackages/search/src/database.ts
📚 Learning: 2025-11-27T06:48:31.996Z
Learnt from: Shunseii
Repo: Shunseii/bahar PR: 24
File: apps/web/src/hooks/db/index.ts:59-60
Timestamp: 2025-11-27T06:48:31.996Z
Learning: In the Bahar project, Orama methods (insert, update, remove, search) are synchronous by default since no async plugins are being used. Orama v3.0.0+ core methods are sync; they only become async when specific plugins like plugin-embeddings are added. Do not suggest awaiting Orama operations unless async plugins are confirmed to be in use.
Applied to files:
packages/search/src/database.ts
📚 Learning: 2025-11-27T06:48:57.365Z
Learnt from: Shunseii
Repo: Shunseii/bahar PR: 24
File: apps/web/src/hooks/db/index.ts:84-85
Timestamp: 2025-11-27T06:48:57.365Z
Learning: In the Orama search library, core operations like insert(), remove(), and update() are synchronous by default. They only become asynchronous when specific plugins (e.g., embedding generators) or Orama Cloud features are used. Don't suggest awaiting these operations unless async plugins are confirmed to be in use.
Applied to files:
packages/search/src/database.ts
📚 Learning: 2025-11-27T06:49:36.986Z
Learnt from: Shunseii
Repo: Shunseii/bahar PR: 24
File: apps/web/src/hooks/db/index.ts:114-115
Timestamp: 2025-11-27T06:49:36.986Z
Learning: In the bahar project, Orama database operations (insert, update, remove) are synchronous because only the plugin-qps plugin is used, which is a local scoring plugin. These operations should not be awaited.
Applied to files:
packages/search/src/database.ts
🧬 Code graph analysis (1)
packages/search/src/database.ts (3)
packages/search/src/schema.ts (1)
DictionaryDocument(66-82)apps/web/src/lib/utils.ts (1)
stripArabicDiacritics(118-121)packages/search/src/arabic.ts (1)
stripArabicDiacritics(9-11)
⏰ Context from checks skipped due to timeout of 90000ms. You can increase the timeout in your CodeRabbit configuration to a maximum of 15 minutes (900000ms). (1)
- GitHub Check: Cloudflare Pages
🔇 Additional comments (8)
packages/search/src/database.ts (5)
5-24: LGTM - Clean imports and type alias.The imports are well-organized, and the
SearchResultstype alias improves readability by avoiding repetition of the verbose generic type throughout the file.
99-137: Well-structured search configuration constants.The separation of exact and normalized properties with corresponding boost configuration is a clean approach for the two-pass search strategy. The boost values (10) consistently prioritize key fields.
161-171: Early return for empty term handles edge case correctly.When no search term is provided, bypassing the two-pass logic and performing a simple search with pagination parameters is appropriate.
173-190: Tolerance scaling based on term length is well-designed.The adaptive tolerance strategy (0 for short terms ≤4 chars, 1 for longer) for exact matching provides a good balance between precision and recall for Arabic text search.
207-220: Offset handling and count estimation have been improved.The previous review comments have been addressed:
- Offset is now correctly applied via
fetchLimit = offset + limitand.slice(offset, offset + limit)for proper pagination- Count calculation now uses
Math.max(exactResults.count, fuzzyResults.count)which provides a reasonable estimateHowever, note that
estimatedCountmay still be imprecise in edge cases where there's significant overlap between exact and fuzzy results, but this is acceptable for UI purposes and avoids the complexity of computing exact deduplicated counts across all results.apps/web/src/hooks/useSearch.ts (3)
1-7: Good refactor - Uses shared search module.The previous review comments about duplicated two-pass logic have been addressed. The hook now correctly imports and delegates to
searchDictionaryfrom the shared package, maintaining a single source of truth.
34-46: Clean delegation to searchDictionary.The simplified search callback properly delegates to the shared
searchDictionaryfunction with appropriate parameters. This eliminates the previously duplicated two-pass search logic.
172-178: LGTM - Proper hasMore calculation.The condition
hits.length >= searchResultsMetadata.countcorrectly determines when all results have been loaded.
1 participant
Summary by CodeRabbit
New Features
Improvements
✏️ Tip: You can customize this high-level summary in your review settings.