mirror of
https://github.com/SamyRai/tercul-backend.git
synced 2025-12-27 05:11:34 +00:00
4957117cb6
- Core Go application with GraphQL API using gqlgen - Comprehensive data models for literary works, authors, translations - Repository pattern with caching layer - Authentication and authorization system - Linguistics analysis capabilities with multiple adapters - Vector search integration with Weaviate - Docker containerization support - Python data migration and analysis scripts - Clean architecture with proper separation of concerns - Production-ready configuration and middleware - Proper .gitignore excluding vendor/, database files, and build artifacts
529 lines
17 KiB
Go
529 lines
17 KiB
Go
package weaviate
|
|
|
|
import (
|
|
"context"
|
|
"log"
|
|
|
|
"github.com/weaviate/weaviate-go-client/v5/weaviate"
|
|
"github.com/weaviate/weaviate/entities/models"
|
|
)
|
|
|
|
// CreateSchema initializes the schema in Weaviate
|
|
func CreateSchema(client *weaviate.Client) {
|
|
// Define all classes
|
|
classes := []*models.Class{
|
|
{
|
|
Class: "Work",
|
|
Properties: []*models.Property{
|
|
{Name: "name", DataType: []string{"text"}},
|
|
{Name: "language", DataType: []string{"string"}},
|
|
{Name: "content", DataType: []string{"text"}},
|
|
// Relationship properties (references)
|
|
{Name: "source", DataType: []string{"Source"}},
|
|
{Name: "embedding", DataType: []string{"Embedding"}},
|
|
{Name: "copyright", DataType: []string{"Copyright"}},
|
|
{Name: "collection", DataType: []string{"Collection"}},
|
|
{Name: "tag", DataType: []string{"Tag"}},
|
|
{Name: "readabilityScore", DataType: []string{"ReadabilityScore"}},
|
|
{Name: "media", DataType: []string{"Media"}},
|
|
{Name: "writingStyle", DataType: []string{"WritingStyle"}},
|
|
{Name: "emotion", DataType: []string{"Emotion"}},
|
|
{Name: "translation", DataType: []string{"Translation"}},
|
|
{Name: "category", DataType: []string{"Category"}},
|
|
{Name: "topicCluster", DataType: []string{"TopicCluster"}},
|
|
{Name: "mood", DataType: []string{"Mood"}},
|
|
{Name: "concept", DataType: []string{"Concept"}},
|
|
{Name: "linguisticLayer", DataType: []string{"LinguisticLayer"}},
|
|
{Name: "workStats", DataType: []string{"WorkStats"}},
|
|
{Name: "textMetadata", DataType: []string{"TextMetadata"}},
|
|
{Name: "poeticAnalysis", DataType: []string{"PoeticAnalysis"}},
|
|
{Name: "hybridEntityWork", DataType: []string{"HybridEntity_Work"}},
|
|
{Name: "copyrightClaim", DataType: []string{"CopyrightClaim"}},
|
|
{Name: "author", DataType: []string{"Author"}},
|
|
},
|
|
VectorIndexConfig: map[string]interface{}{
|
|
"distance": "cosine",
|
|
"efConstruction": 128,
|
|
"maxConnections": 64,
|
|
},
|
|
},
|
|
{
|
|
Class: "Translation",
|
|
Properties: []*models.Property{
|
|
{Name: "name", DataType: []string{"text"}},
|
|
{Name: "language", DataType: []string{"string"}},
|
|
{Name: "content", DataType: []string{"text"}},
|
|
// Relationships
|
|
{Name: "work", DataType: []string{"Work"}},
|
|
{Name: "embedding", DataType: []string{"Embedding"}},
|
|
{Name: "translationStats", DataType: []string{"TranslationStats"}},
|
|
{Name: "hybridEntityWork", DataType: []string{"HybridEntity_Work"}},
|
|
{Name: "copyright", DataType: []string{"Copyright"}},
|
|
{Name: "copyrightClaim", DataType: []string{"CopyrightClaim"}},
|
|
{Name: "translator", DataType: []string{"User"}},
|
|
},
|
|
VectorIndexConfig: map[string]interface{}{
|
|
"distance": "cosine",
|
|
},
|
|
},
|
|
{
|
|
Class: "TopicCluster",
|
|
Properties: []*models.Property{
|
|
{Name: "name", DataType: []string{"text"}},
|
|
{Name: "work", DataType: []string{"Work"}},
|
|
},
|
|
},
|
|
{
|
|
Class: "Emotion",
|
|
Properties: []*models.Property{
|
|
{Name: "name", DataType: []string{"text"}},
|
|
{Name: "language", DataType: []string{"string"}},
|
|
{Name: "user", DataType: []string{"User"}},
|
|
{Name: "work", DataType: []string{"Work"}},
|
|
{Name: "collection", DataType: []string{"Collection"}},
|
|
},
|
|
VectorIndexConfig: map[string]interface{}{"distance": "cosine"},
|
|
},
|
|
{
|
|
Class: "Embedding",
|
|
Properties: []*models.Property{
|
|
{Name: "vector", DataType: []string{"number[]"}},
|
|
{Name: "work", DataType: []string{"Work"}},
|
|
{Name: "translation", DataType: []string{"Translation"}},
|
|
},
|
|
},
|
|
{
|
|
Class: "Gamification",
|
|
Properties: []*models.Property{
|
|
{Name: "name", DataType: []string{"text"}},
|
|
{Name: "user", DataType: []string{"User"}},
|
|
},
|
|
},
|
|
{
|
|
Class: "Contribution",
|
|
Properties: []*models.Property{
|
|
{Name: "name", DataType: []string{"text"}},
|
|
{Name: "status", DataType: []string{"string"}},
|
|
{Name: "user", DataType: []string{"User"}},
|
|
{Name: "work", DataType: []string{"Work"}},
|
|
{Name: "translation", DataType: []string{"Translation"}},
|
|
},
|
|
},
|
|
{
|
|
Class: "Stats",
|
|
Properties: []*models.Property{
|
|
{Name: "name", DataType: []string{"text"}},
|
|
{Name: "user", DataType: []string{"User"}},
|
|
{Name: "work", DataType: []string{"Work"}},
|
|
},
|
|
},
|
|
{
|
|
Class: "LanguageAnalysis",
|
|
Properties: []*models.Property{
|
|
{Name: "name", DataType: []string{"text"}},
|
|
{Name: "work", DataType: []string{"Work"}},
|
|
},
|
|
},
|
|
{
|
|
Class: "WritingStyle",
|
|
Properties: []*models.Property{
|
|
{Name: "name", DataType: []string{"text"}},
|
|
{Name: "language", DataType: []string{"string"}},
|
|
{Name: "work", DataType: []string{"Work"}},
|
|
},
|
|
},
|
|
{
|
|
Class: "Media",
|
|
Properties: []*models.Property{
|
|
{Name: "name", DataType: []string{"text"}},
|
|
{Name: "language", DataType: []string{"string"}},
|
|
{Name: "author", DataType: []string{"Author"}},
|
|
{Name: "translation", DataType: []string{"Translation"}},
|
|
{Name: "country", DataType: []string{"Country"}},
|
|
{Name: "city", DataType: []string{"City"}},
|
|
{Name: "mediaStats", DataType: []string{"MediaStats"}},
|
|
{Name: "copyright", DataType: []string{"Copyright"}},
|
|
{Name: "copyrightClaim", DataType: []string{"CopyrightClaim"}},
|
|
},
|
|
},
|
|
{
|
|
Class: "Collection",
|
|
Properties: []*models.Property{
|
|
{Name: "name", DataType: []string{"text"}},
|
|
{Name: "description", DataType: []string{"text"}},
|
|
{Name: "work", DataType: []string{"Work"}},
|
|
{Name: "user", DataType: []string{"User"}},
|
|
{Name: "collectionStats", DataType: []string{"CollectionStats"}},
|
|
},
|
|
},
|
|
{
|
|
Class: "Bookmark",
|
|
Properties: []*models.Property{
|
|
{Name: "name", DataType: []string{"text"}},
|
|
{Name: "work", DataType: []string{"Work"}},
|
|
{Name: "user", DataType: []string{"User"}},
|
|
},
|
|
},
|
|
{
|
|
Class: "Word",
|
|
Properties: []*models.Property{
|
|
{Name: "name", DataType: []string{"text"}},
|
|
{Name: "concept", DataType: []string{"Concept"}},
|
|
{Name: "work", DataType: []string{"Work"}},
|
|
},
|
|
},
|
|
{
|
|
Class: "Copyright",
|
|
Properties: []*models.Property{
|
|
{Name: "name", DataType: []string{"text"}},
|
|
{Name: "language", DataType: []string{"string"}},
|
|
{Name: "workOwner", DataType: []string{"Author"}},
|
|
{Name: "work", DataType: []string{"Work"}},
|
|
{Name: "translation", DataType: []string{"Translation"}},
|
|
{Name: "book", DataType: []string{"Book"}},
|
|
{Name: "source", DataType: []string{"Source"}},
|
|
},
|
|
},
|
|
{
|
|
Class: "Admin",
|
|
Properties: []*models.Property{
|
|
{Name: "name", DataType: []string{"text"}},
|
|
{Name: "user", DataType: []string{"User"}},
|
|
{Name: "work", DataType: []string{"Work"}},
|
|
},
|
|
},
|
|
{
|
|
Class: "Author",
|
|
Properties: []*models.Property{
|
|
{Name: "name", DataType: []string{"text"}},
|
|
{Name: "language", DataType: []string{"string"}},
|
|
{Name: "biography", DataType: []string{"text"}},
|
|
{Name: "birthDate", DataType: []string{"date"}},
|
|
{Name: "deathDate", DataType: []string{"date"}},
|
|
{Name: "work", DataType: []string{"Work"}},
|
|
{Name: "book", DataType: []string{"Book"}},
|
|
{Name: "country", DataType: []string{"Country"}},
|
|
{Name: "city", DataType: []string{"City"}},
|
|
{Name: "place", DataType: []string{"Place"}},
|
|
{Name: "address", DataType: []string{"Address"}},
|
|
{Name: "copyrightClaim", DataType: []string{"CopyrightClaim"}},
|
|
{Name: "copyright", DataType: []string{"Copyright"}},
|
|
},
|
|
},
|
|
{
|
|
Class: "Category",
|
|
Properties: []*models.Property{
|
|
{Name: "name", DataType: []string{"text"}},
|
|
{Name: "work", DataType: []string{"Work"}},
|
|
},
|
|
},
|
|
{
|
|
Class: "User",
|
|
Properties: []*models.Property{
|
|
{Name: "name", DataType: []string{"text"}},
|
|
{Name: "email", DataType: []string{"string"}},
|
|
{Name: "role", DataType: []string{"string"}},
|
|
{Name: "work", DataType: []string{"Work"}},
|
|
{Name: "bookmark", DataType: []string{"Bookmark"}},
|
|
{Name: "translation", DataType: []string{"Translation"}},
|
|
{Name: "collection", DataType: []string{"Collection"}},
|
|
{Name: "like", DataType: []string{"Like"}},
|
|
{Name: "comment", DataType: []string{"Comment"}},
|
|
{Name: "author", DataType: []string{"Author"}},
|
|
{Name: "topicCluster", DataType: []string{"TopicCluster"}},
|
|
{Name: "country", DataType: []string{"Country"}},
|
|
{Name: "city", DataType: []string{"City"}},
|
|
{Name: "userStats", DataType: []string{"UserStats"}},
|
|
{Name: "book", DataType: []string{"Book"}},
|
|
{Name: "media", DataType: []string{"Media"}},
|
|
{Name: "address", DataType: []string{"Address"}},
|
|
{Name: "emotion", DataType: []string{"Emotion"}},
|
|
{Name: "copyrightClaim", DataType: []string{"CopyrightClaim"}},
|
|
{Name: "contribution", DataType: []string{"Contribution"}},
|
|
},
|
|
},
|
|
{
|
|
Class: "Book",
|
|
Properties: []*models.Property{
|
|
{Name: "name", DataType: []string{"text"}},
|
|
{Name: "language", DataType: []string{"string"}},
|
|
{Name: "work", DataType: []string{"Work"}},
|
|
{Name: "author", DataType: []string{"Author"}},
|
|
{Name: "bookStats", DataType: []string{"BookStats"}},
|
|
{Name: "copyright", DataType: []string{"Copyright"}},
|
|
{Name: "copyrightClaim", DataType: []string{"CopyrightClaim"}},
|
|
},
|
|
},
|
|
{
|
|
Class: "Source",
|
|
Properties: []*models.Property{
|
|
{Name: "name", DataType: []string{"text"}},
|
|
{Name: "language", DataType: []string{"string"}},
|
|
{Name: "copyright", DataType: []string{"Copyright"}},
|
|
{Name: "copyrightClaim", DataType: []string{"CopyrightClaim"}},
|
|
{Name: "work", DataType: []string{"Work"}},
|
|
},
|
|
},
|
|
{
|
|
Class: "Tag",
|
|
Properties: []*models.Property{
|
|
{Name: "name", DataType: []string{"text"}},
|
|
{Name: "work", DataType: []string{"Work"}},
|
|
},
|
|
},
|
|
{
|
|
Class: "Concept",
|
|
Properties: []*models.Property{
|
|
{Name: "name", DataType: []string{"text"}},
|
|
{Name: "work", DataType: []string{"Work"}},
|
|
{Name: "word", DataType: []string{"Word"}},
|
|
},
|
|
},
|
|
{
|
|
Class: "Comment",
|
|
Properties: []*models.Property{
|
|
{Name: "text", DataType: []string{"text"}},
|
|
{Name: "user", DataType: []string{"User"}},
|
|
{Name: "work", DataType: []string{"Work"}},
|
|
{Name: "translation", DataType: []string{"Translation"}},
|
|
{Name: "lineNumber", DataType: []string{"int"}},
|
|
{Name: "parentComment", DataType: []string{"Comment"}},
|
|
{Name: "like", DataType: []string{"Like"}},
|
|
},
|
|
},
|
|
{
|
|
Class: "ReadabilityScore",
|
|
Properties: []*models.Property{
|
|
{Name: "score", DataType: []string{"number"}},
|
|
{Name: "language", DataType: []string{"string"}},
|
|
{Name: "work", DataType: []string{"Work"}},
|
|
},
|
|
},
|
|
{
|
|
Class: "Language",
|
|
Properties: []*models.Property{
|
|
{Name: "name", DataType: []string{"text"}},
|
|
},
|
|
},
|
|
{
|
|
Class: "Vote",
|
|
Properties: []*models.Property{
|
|
{Name: "value", DataType: []string{"number"}},
|
|
{Name: "user", DataType: []string{"User"}},
|
|
{Name: "work", DataType: []string{"Work"}},
|
|
{Name: "translation", DataType: []string{"Translation"}},
|
|
{Name: "comment", DataType: []string{"Comment"}},
|
|
},
|
|
},
|
|
{
|
|
Class: "Edition",
|
|
Properties: []*models.Property{
|
|
{Name: "version", DataType: []string{"text"}},
|
|
{Name: "book", DataType: []string{"Book"}},
|
|
{Name: "work", DataType: []string{"Work"}},
|
|
},
|
|
},
|
|
{
|
|
Class: "LinguisticLayer",
|
|
Properties: []*models.Property{
|
|
{Name: "name", DataType: []string{"text"}},
|
|
{Name: "language", DataType: []string{"string"}},
|
|
{Name: "work", DataType: []string{"Work"}},
|
|
},
|
|
},
|
|
{
|
|
Class: "Mood",
|
|
Properties: []*models.Property{
|
|
{Name: "name", DataType: []string{"text"}},
|
|
{Name: "language", DataType: []string{"string"}},
|
|
{Name: "work", DataType: []string{"Work"}},
|
|
},
|
|
},
|
|
{
|
|
Class: "Like",
|
|
Properties: []*models.Property{
|
|
{Name: "user", DataType: []string{"User"}},
|
|
{Name: "work", DataType: []string{"Work"}},
|
|
{Name: "translation", DataType: []string{"Translation"}},
|
|
{Name: "comment", DataType: []string{"Comment"}},
|
|
},
|
|
},
|
|
{
|
|
Class: "Notification",
|
|
Properties: []*models.Property{
|
|
{Name: "message", DataType: []string{"text"}},
|
|
{Name: "language", DataType: []string{"string"}},
|
|
{Name: "user", DataType: []string{"User"}},
|
|
},
|
|
},
|
|
{
|
|
Class: "EditorialWorkflow",
|
|
Properties: []*models.Property{
|
|
{Name: "stage", DataType: []string{"text"}},
|
|
{Name: "language", DataType: []string{"string"}},
|
|
{Name: "work", DataType: []string{"Work"}},
|
|
{Name: "translation", DataType: []string{"Translation"}},
|
|
{Name: "user", DataType: []string{"User"}},
|
|
},
|
|
},
|
|
{
|
|
Class: "Monetization",
|
|
Properties: []*models.Property{
|
|
{Name: "amount", DataType: []string{"number"}},
|
|
{Name: "language", DataType: []string{"string"}},
|
|
{Name: "work", DataType: []string{"Work"}},
|
|
{Name: "translation", DataType: []string{"Translation"}},
|
|
{Name: "book", DataType: []string{"Book"}},
|
|
},
|
|
},
|
|
{
|
|
Class: "Country",
|
|
Properties: []*models.Property{
|
|
{Name: "name", DataType: []string{"text"}},
|
|
{Name: "language", DataType: []string{"string"}},
|
|
{Name: "author", DataType: []string{"Author"}},
|
|
{Name: "user", DataType: []string{"User"}},
|
|
},
|
|
},
|
|
{
|
|
Class: "City",
|
|
Properties: []*models.Property{
|
|
{Name: "name", DataType: []string{"text"}},
|
|
{Name: "language", DataType: []string{"string"}},
|
|
{Name: "country", DataType: []string{"Country"}},
|
|
{Name: "author", DataType: []string{"Author"}},
|
|
{Name: "user", DataType: []string{"User"}},
|
|
},
|
|
},
|
|
{
|
|
Class: "Address",
|
|
Properties: []*models.Property{
|
|
{Name: "street", DataType: []string{"text"}},
|
|
{Name: "city", DataType: []string{"City"}},
|
|
{Name: "country", DataType: []string{"Country"}},
|
|
{Name: "author", DataType: []string{"Author"}},
|
|
{Name: "user", DataType: []string{"User"}},
|
|
},
|
|
},
|
|
{
|
|
Class: "WorkStats",
|
|
Properties: []*models.Property{
|
|
{Name: "views", DataType: []string{"number"}},
|
|
{Name: "work", DataType: []string{"Work"}},
|
|
},
|
|
},
|
|
{
|
|
Class: "TranslationStats",
|
|
Properties: []*models.Property{
|
|
{Name: "views", DataType: []string{"number"}},
|
|
{Name: "translation", DataType: []string{"Translation"}},
|
|
},
|
|
},
|
|
{
|
|
Class: "MediaStats",
|
|
Properties: []*models.Property{
|
|
{Name: "views", DataType: []string{"number"}},
|
|
{Name: "media", DataType: []string{"Media"}},
|
|
},
|
|
},
|
|
{
|
|
Class: "UserStats",
|
|
Properties: []*models.Property{
|
|
{Name: "activity", DataType: []string{"number"}},
|
|
{Name: "user", DataType: []string{"User"}},
|
|
},
|
|
},
|
|
{
|
|
Class: "Place",
|
|
Properties: []*models.Property{
|
|
{Name: "name", DataType: []string{"text"}},
|
|
{Name: "language", DataType: []string{"string"}},
|
|
{Name: "city", DataType: []string{"City"}},
|
|
{Name: "country", DataType: []string{"Country"}},
|
|
{Name: "author", DataType: []string{"Author"}},
|
|
},
|
|
},
|
|
{
|
|
Class: "BookStats",
|
|
Properties: []*models.Property{
|
|
{Name: "sales", DataType: []string{"number"}},
|
|
{Name: "book", DataType: []string{"Book"}},
|
|
},
|
|
},
|
|
{
|
|
Class: "CollectionStats",
|
|
Properties: []*models.Property{
|
|
{Name: "items", DataType: []string{"number"}},
|
|
{Name: "collection", DataType: []string{"Collection"}},
|
|
},
|
|
},
|
|
{
|
|
Class: "TextMetadata",
|
|
Properties: []*models.Property{
|
|
{Name: "analysis", DataType: []string{"text"}},
|
|
{Name: "language", DataType: []string{"string"}},
|
|
{Name: "work", DataType: []string{"Work"}},
|
|
},
|
|
},
|
|
{
|
|
Class: "PoeticAnalysis",
|
|
Properties: []*models.Property{
|
|
{Name: "structure", DataType: []string{"text"}},
|
|
{Name: "language", DataType: []string{"string"}},
|
|
{Name: "work", DataType: []string{"Work"}},
|
|
},
|
|
},
|
|
{
|
|
Class: "HybridEntity_Work",
|
|
Properties: []*models.Property{
|
|
{Name: "name", DataType: []string{"text"}},
|
|
{Name: "work", DataType: []string{"Work"}},
|
|
{Name: "translation", DataType: []string{"Translation"}},
|
|
},
|
|
},
|
|
{
|
|
Class: "CopyrightClaim",
|
|
Properties: []*models.Property{
|
|
{Name: "details", DataType: []string{"text"}},
|
|
{Name: "work", DataType: []string{"Work"}},
|
|
{Name: "translation", DataType: []string{"Translation"}},
|
|
{Name: "book", DataType: []string{"Book"}},
|
|
{Name: "source", DataType: []string{"Source"}},
|
|
{Name: "author", DataType: []string{"Author"}},
|
|
{Name: "user", DataType: []string{"User"}},
|
|
},
|
|
},
|
|
{
|
|
Class: "Contributor",
|
|
Properties: []*models.Property{
|
|
{Name: "name", DataType: []string{"text"}},
|
|
{Name: "user", DataType: []string{"User"}},
|
|
{Name: "work", DataType: []string{"Work"}},
|
|
{Name: "translation", DataType: []string{"Translation"}},
|
|
},
|
|
},
|
|
{
|
|
Class: "Edge",
|
|
Properties: []*models.Property{
|
|
{Name: "sourceTable", DataType: []string{"string"}},
|
|
{Name: "sourceId", DataType: []string{"string"}},
|
|
{Name: "targetTable", DataType: []string{"string"}},
|
|
{Name: "targetId", DataType: []string{"string"}},
|
|
{Name: "relation", DataType: []string{"string"}},
|
|
{Name: "language", DataType: []string{"string"}},
|
|
{Name: "extra", DataType: []string{"object"}},
|
|
},
|
|
},
|
|
}
|
|
|
|
// Create each class in Weaviate
|
|
for _, class := range classes {
|
|
err := client.Schema().ClassCreator().WithClass(class).Do(context.Background())
|
|
if err != nil {
|
|
log.Printf("Failed to create class %s: %v", class.Class, err)
|
|
}
|
|
}
|
|
|
|
log.Println("Weaviate schema created successfully.")
|
|
}
|