Enter your search text in the box above

Select a result to preview

sqlite/glossary.go

sqlite/glossary.go

Module: Glossary Backend
Path: backend/internal/repository/sqlite/glossary.go

SQLite repository implementation for glossary CRUD operations. Handles slug generation, FTS5 indexing, reciprocal relationship management, and DTO conversion.

Exports

Methods on the Store struct:

Name Kind Description
CreateGlossaryEntry method Full create with slug generation, duplicate term check, reciprocal adds, FTS5 index
GetGlossaryEntry method Query by ID, convert to DTO
ListGlossaryEntries method Filtered query with post-query tag matching, pagination, sorting
UpdateGlossaryEntry method Partial update with reciprocal diffs, index text rebuild, FTS5 reindex
DeleteGlossaryEntry method Cleanup reciprocal relationships, delete entry, remove from FTS5
UpdateGlossaryEntryCounts method Set absolute usage/doc counts (called by TermScanHandler after scan)
DecrementGlossaryEntryCounts method Read-then-write decrement with floor at 0 (called on content deletion)

Internal Helpers

Name Description
getOrCreateGlossary Lazy Glossary entity creation -- queries the project with its glossary edge, creates one if missing
generateSlug Lowercase, spaces to hyphens, strip non-alphanumeric, collapse consecutive hyphens
ensureUniqueSlug Appends -2, -3, etc. if the base slug already exists within the glossary
buildIndexText Joins term + shortDefinition + aliases + tags into a single lowercased string for FTS5
glossaryEntryToDTO Converts Ent entity to repository.GlossaryEntryDTO with nil-safe slices (empty array, not null)
applyReciprocalAddsAfterCreate Adds source entry ID to each target's inverse relationship array
applyReciprocalDiffsAfterUpdate Diffs old vs new relationship arrays, adds/removes source ID on target entries
cleanupRelationshipReferences Removes deleted entry ID from all entries' relationship arrays across the entire glossary
getRelationshipPairs Returns 6 source-to-inverse mapping pairs for reciprocal synchronization
stripSelfReferences Removes an entry's own ID from its relationship arrays after save
deduplicateSlice Removes duplicate strings from a slice using a seen-set
addToStringSlice Appends an ID to a slice if not already present
removeFromStringSlice Removes all occurrences of an ID from a slice
containsString Checks if a string slice contains a specific value
indexGlossaryForSearch Sends entry fields to FTS5 index; no-op if search provider is nil
removeGlossaryFromSearch Removes entry from FTS5 index; no-op if search provider is nil
hasAnyTag Case-insensitive any-match tag filter for post-query filtering

Reciprocal Relationship System

When a glossary entry references another entry in a relationship array, the inverse relationship is automatically maintained on the target entry.

Source Field Target Inverse Field Type
broaderTerms narrowerTerms Asymmetric
narrowerTerms broaderTerms Asymmetric
relatedTerms relatedTerms Symmetric
synonyms synonyms Symmetric
antonyms antonyms Symmetric
seeAlso seeAlso Symmetric

Self-references are silently stripped after save. Duplicates are collapsed before persistence.

Create Flow

  1. Open per-project SQLite database via dbPathFor(projectSlug)
  2. Get or create the Glossary parent entity via getOrCreateGlossary()
  3. Check for duplicate term (case-insensitive) using TermEqualFold
  4. Generate slug from term via generateSlug(), ensure uniqueness via ensureUniqueSlug()
  5. Build composite indexText from term, shortDefinition, aliases, tags
  6. Build Ent create operation with all provided fields
  7. Deduplicate relationship arrays before save
  8. Save the entry to the database
  9. Strip self-references from relationship arrays (re-save; entry ID was unknown before first save)
  10. Apply reciprocal adds on all target entries via applyReciprocalAddsAfterCreate()
  11. Convert to DTO via glossaryEntryToDTO()
  12. Index for FTS5 search via indexGlossaryForSearch()
  13. Return the DTO

Update Flow

  1. Open per-project SQLite database
  2. Query existing entry by ID (404 if not found)
  3. Snapshot old entry for reciprocal relationship diffing
  4. Build Ent update operation with only non-nil fields
  5. Strip self-references and deduplicate relationship arrays
  6. Rebuild indexText if any searchable field changed (term, shortDefinition, aliases, tags)
  7. Save the updated entry
  8. Apply reciprocal diffs on target entries via applyReciprocalDiffsAfterUpdate() -- adds source ID to newly referenced targets, removes from de-referenced targets
  9. Convert to DTO
  10. Re-index in FTS5 (delete + insert pattern)
  11. Return the DTO

Delete Flow

  1. Open per-project SQLite database
  2. Clean up reciprocal references via cleanupRelationshipReferences() -- queries all entries, removes deleted ID from every relationship array
  3. Delete the entry by ID (404 if not found)
  4. Remove from FTS5 index via removeGlossaryFromSearch()

Imports / Dependencies

Import Source Purpose
ent backend/internal/ent Ent ORM client and entity types
glossary backend/internal/ent/glossary Glossary entity predicates
glossaryentry backend/internal/ent/glossaryentry GlossaryEntry entity predicates and field constants
entproject backend/internal/ent/project Project entity predicates (aliased to avoid collision)
repository backend/internal/repository DTO types and input/output structs

Side Effects

Notes

Tag filtering is performed post-query in Go code due to SQLite's lack of efficient JSON array-contains predicates. This means the total count returned by ListGlossaryEntries may not reflect tag-filtered results accurately -- it counts all entries matching the non-tag filters, while the returned entries slice is further filtered by tags. For glossaries with thousands of entries, this O(n) scan runs on every paginated page.