domain

type AIProvider string

type AppSettings struct {
	// Search holds search behaviour settings.
	Search SearchSettings

	// Embedding holds embedding provider settings.
	Embedding EmbeddingSettings

	// LLM holds LLM provider settings.
	LLM LLMSettings

	// VectorIndex holds vector index settings.
	VectorIndex VectorIndexSettings
}

type AuthCapability uint8

type AuthMethod string

type AuthProvider struct {
	// ID is the unique identifier (UUID).
	ID string `json:"id"`
	// Name is the user-friendly name (e.g., "My Google App", "Work GitHub").
	Name string `json:"name"`
	// ProviderType identifies the provider (google, github, slack, etc.).
	ProviderType ProviderType `json:"provider_type"`
	// AuthMethod indicates how sources using this provider authenticate (oauth, pat, none).
	AuthMethod AuthMethod `json:"auth_method"`

	// OAuth holds OAuth application credentials (for AuthMethodOAuth).
	// Nil for PAT or no-auth providers.
	OAuth *OAuthProviderConfig `json:"oauth,omitempty"`

	// CreatedAt is when the provider was created.
	CreatedAt time.Time `json:"created_at"`
	// UpdatedAt is when the provider was last updated.
	UpdatedAt time.Time `json:"updated_at"`
}

type ChangeType int

type Chunk struct {
	// ID is the unique identifier for the chunk.
	ID string

	// DocumentID links to the parent Document.
	DocumentID string

	// Content is the text content of this chunk.
	Content string

	// Position is the ordinal position within the document.
	Position int

	// Embedding is the vector representation for semantic search.
	Embedding []float32

	// Metadata contains chunk-specific key-value pairs.
	Metadata map[string]any
}

type ConfigKey struct {
	// Key is the configuration key name.
	Key string
	// Label is the human-readable label for UI display.
	Label string
	// Description explains what this field is for.
	Description string
	// Default is the default value for this field (shown in placeholder).
	Default string
	// Required indicates whether this field must be provided.
	Required bool
	// Secret indicates whether this field should be masked in UI (e.g., tokens).
	Secret bool
}

type ConnectorType struct {
	// ID is the unique identifier (e.g., "filesystem", "github", "google-drive").
	ID string
	// Name is the human-readable display name.
	Name string
	// Description provides a brief explanation of the connector.
	Description string
	// ProviderType identifies which auth provider this connector uses.
	ProviderType ProviderType
	// AuthCapability specifies what authentication methods this connector supports.
	// Use this to determine if user should be given a choice of auth methods.
	AuthCapability AuthCapability
	// AuthMethod specifies how the connector authenticates (derived from provider).
	// Deprecated: Use AuthCapability instead. Kept for backward compatibility.
	AuthMethod AuthMethod
	// ConfigKeys lists the configuration fields required by this connector.
	ConfigKeys []ConfigKey
	// WebURLResolver converts document URIs to web-openable URLs.
	// If nil, falls back to legacy URI conversion.
	WebURLResolver WebURLResolver
}

type Credentials struct {
	// ID is the unique identifier (UUID).
	ID string `json:"id"`
	// SourceID links to the Source this credentials belongs to (1:1 relationship).
	SourceID string `json:"source_id"`

	// AccountIdentifier is the user's email or username from the provider.
	// Fetched from the provider's userinfo endpoint after authentication.
	// Examples: "user@gmail.com", "octocat", "user@company.slack.com"
	AccountIdentifier string `json:"account_identifier,omitempty"`

	// OAuth holds OAuth tokens (for OAuth authentication).
	// Nil for PAT authentication.
	OAuth *OAuthCredentials `json:"oauth,omitempty"`

	// PAT holds the Personal Access Token (for PAT authentication).
	// Nil for OAuth authentication.
	PAT *PATCredentials `json:"pat,omitempty"`

	// CreatedAt is when the credentials were created.
	CreatedAt time.Time `json:"created_at"`
	// UpdatedAt is when the credentials were last updated.
	UpdatedAt time.Time `json:"updated_at"`
}

type Document struct {
	// ID is the unique identifier for the document.
	ID string

	// SourceID links to the Source that produced this document.
	SourceID string

	// URI is the original location (file path, URL, etc).
	URI string

	// Title is the human-readable title.
	Title string

	// Content is the full text content after normalisation.
	// This is the complete document text before chunking.
	Content string

	// ParentID links to a parent document for hierarchical sources.
	ParentID *string

	// Metadata contains arbitrary key-value pairs.
	Metadata map[string]any

	// CreatedAt is when the document was first indexed.
	CreatedAt time.Time

	// UpdatedAt is when the document was last updated.
	UpdatedAt time.Time
}

type EmbeddingSettings struct {
	// Provider is the embedding service provider.
	Provider AIProvider

	// Model is the embedding model name.
	Model string

	// BaseURL is the API endpoint (for Ollama).
	BaseURL string

	// APIKey is the API key (for OpenAI).
	APIKey string
}

type Exclusion struct {
	// ID is the unique identifier for the exclusion.
	ID string

	// SourceID links to the Source this exclusion applies to.
	SourceID string

	// DocumentID is the ID of the excluded document.
	DocumentID string

	// URI is the original location for matching on re-sync.
	URI string

	// Reason is an optional explanation for the exclusion.
	Reason string

	// ExcludedAt is when the document was excluded.
	ExcludedAt time.Time
}

type LLMSettings struct {
	// Provider is the LLM service provider.
	Provider AIProvider

	// Model is the LLM model name.
	Model string

	// BaseURL is the API endpoint (for Ollama).
	BaseURL string

	// APIKey is the API key (for OpenAI/Anthropic).
	APIKey string
}

type OAuthCredentials struct {
	// AccessToken is the bearer token for API access.
	AccessToken string `json:"access_token"`
	// RefreshToken is used to obtain new access tokens.
	RefreshToken string `json:"refresh_token,omitempty"`
	// TokenType is typically "Bearer".
	TokenType string `json:"token_type"`
	// Expiry is when the access token expires.
	Expiry time.Time `json:"expiry,omitempty"`
}

type OAuthProviderConfig struct {
	// ClientID is the OAuth client ID from the developer console.
	ClientID string `json:"client_id"`
	// ClientSecret is the OAuth client secret from the developer console.
	ClientSecret string `json:"client_secret"`
	// Scopes are the OAuth scopes to request.
	Scopes []string `json:"scopes"`
	// AuthURL is the authorization endpoint (optional override for custom OAuth servers).
	AuthURL string `json:"auth_url,omitempty"`
	// TokenURL is the token exchange endpoint (optional override for custom OAuth servers).
	TokenURL string `json:"token_url,omitempty"`
	// RedirectURI is the callback URI (default: http://localhost:PORT/callback).
	RedirectURI string `json:"redirect_uri,omitempty"`
}

type OAuthToken struct {
	// AccessToken is the bearer token for API access.
	AccessToken string `json:"access_token"`
	// RefreshToken is used to obtain new access tokens.
	RefreshToken string `json:"refresh_token,omitempty"`
	// TokenType is typically "Bearer".
	TokenType string `json:"token_type"`
	// Expiry is when the access token expires.
	Expiry time.Time `json:"expiry,omitempty"`
}

type PATCredentials struct {
	// Token is the actual personal access token.
	Token string `json:"token"`
}

type PipelineConfig struct {
	// Processors is the ordered list of processor names to run.
	Processors []string

	// ProcessorConfigs holds per-processor configuration as generic maps.
	// Key is processor name, value is processor-specific config.
	ProcessorConfigs map[string]map[string]any
}

type ProviderType string

type RawDocument struct {
	// SourceID links to the Source that produced this document.
	SourceID string

	// URI is the original location (file path, URL, etc).
	URI string

	// MIMEType is the content type (e.g., "application/pdf").
	MIMEType string

	// Content is the raw bytes.
	Content []byte

	// ParentURI links to a parent for hierarchical sources.
	ParentURI *string

	// Metadata contains connector-specific key-value pairs.
	Metadata map[string]any
}

type RawDocumentChange struct {
	// Type is the kind of change.
	Type ChangeType

	// Document is the affected document.
	Document RawDocument
}

type ScheduledTask struct {
	// ID is the unique identifier for the task.
	ID string

	// Name is a human-readable name for the task.
	Name string

	// Interval defines how often the task should run.
	Interval time.Duration

	// LastRun is when the task last ran.
	LastRun time.Time

	// NextRun is when the task should run next.
	NextRun time.Time

	// LastError contains the last error message, if any.
	LastError string

	// LastSuccess is when the task last completed successfully.
	LastSuccess time.Time

	// Enabled indicates whether the task is active.
	Enabled bool
}

type SchedulerConfig struct {
	// Enabled is the master switch for the scheduler.
	Enabled bool

	// TaskConfigs holds per-task configuration.
	TaskConfigs map[string]TaskConfig
}

type SearchMode string

type SearchOptions struct {
	// Limit is the maximum number of results.
	Limit int

	// Offset is the number of results to skip.
	Offset int

	// SourceIDs filters to specific sources.
	SourceIDs []string

	// Semantic enables vector similarity search.
	Semantic bool

	// Hybrid enables combined keyword + semantic search.
	Hybrid bool
}

type SearchResult struct {
	// Document is the matched document.
	Document Document

	// Chunk is the specific chunk that matched.
	Chunk Chunk

	// Score is the relevance score.
	Score float64

	// Highlights contains snippets with matched terms.
	Highlights []string

	// SourceName is the display name of the source (includes account identifier).
	// Example: "Gmail - user@gmail.com" or "GitHub - octocat"
	SourceName string
}

type SearchSettings struct {
	// Mode is the search retrieval mode.
	Mode SearchMode
}

type Source struct {
	// ID is the unique identifier for the source.
	ID string

	// Type identifies the connector type (e.g., "filesystem", "gmail").
	Type string

	// Name is the human-readable name for this source.
	Name string

	// Config contains connector-specific configuration.
	Config map[string]string

	// AuthorizationID references the Authorization used by this source.
	// Deprecated: Use AuthProviderID and CredentialsID instead.
	// Kept for backward compatibility during migration.
	AuthorizationID string

	// AuthProviderID references the AuthProvider (OAuth app or PAT provider config).
	// Empty string for no-auth connectors (filesystem).
	AuthProviderID string

	// CredentialsID references this source's Credentials (tokens + account info).
	// Empty string for no-auth connectors.
	CredentialsID string

	// CreatedAt is when the source was created.
	CreatedAt time.Time

	// UpdatedAt is when the source was last updated.
	UpdatedAt time.Time
}

type SyncState struct {
	// SourceID links to the Source being synced.
	SourceID string

	// Cursor is an opaque token for incremental sync.
	Cursor string

	// LastSync is when the last successful sync completed.
	LastSync time.Time
}

type TaskConfig struct {
	// Enabled indicates whether this task should run.
	Enabled bool

	// Interval defines how often the task should run.
	Interval time.Duration
}

type TaskResult struct {
	// TaskID identifies which task was run.
	TaskID string

	// StartedAt is when the task started.
	StartedAt time.Time

	// EndedAt is when the task completed.
	EndedAt time.Time

	// Success indicates whether the task completed without error.
	Success bool

	// Error contains the error message if Success is false.
	Error string

	// ItemsProcessed is a count of items handled (e.g., documents synced).
	ItemsProcessed int
}

type VectorIndexSettings struct {
	// Enabled indicates whether vector indexing is active.
	Enabled bool

	// Dimensions is the embedding vector size.
	Dimensions int

	// Precision is the storage precision for vectors.
	// Default is float16 (best balance of size vs quality).
	Precision VectorPrecision
}

type VectorPrecision string

type WebURLResolver func(uri string, metadata map[string]any) string