frontend/src/types.ts

// ---- API Request types ----

export interface InitRequest {
  model_name: string;
  chunk_size: number;
  chunk_overlap: number;
  batch_size: number;
}

export interface DocumentRequest {
  doc_id: string;
  text: string;
}

export interface QueryRequest {
  text: string;
  top_k: number;
}

export interface CompareRequest {
  text_a: string;
  text_b: string;
}

export interface KeywordAnalysisRequest {
  keyword: string;
  top_k: number;
  cluster_threshold: number;
}

export interface KeywordMatchRequest {
  keyword: string;
  candidate_meanings: string[];
}

export interface BatchAnalysisRequest {
  keywords: string[];
  top_k: number;
  cluster_threshold: number;
  compare_across: boolean;
}

// ---- API Response types ----

export interface ChunkPreview {
  index: number;
  text: string;
}

export interface InitResponse {
  status: string;
  model: string;
  load_time_seconds: number;
}

export interface AddDocResponse {
  status: string;
  doc_id: string;
  num_chunks: number;
  chunks_preview: ChunkPreview[];
}

export interface BuildIndexResponse {
  status: string;
  total_chunks: number;
  embedding_dim: number;
  build_time_seconds: number;
}

export interface QueryResultItem {
  rank: number;
  score: number;
  doc_id: string;
  chunk_index: number;
  text: string;
}

export interface QueryResponse {
  query: string;
  results: QueryResultItem[];
}

export interface CompareResponse {
  text_a: string;
  text_b: string;
  similarity: number;
}

export interface ClusterContext {
  doc_id: string;
  chunk_index: number;
  text: string;
  highlight_positions: [number, number][];
}

export interface SimilarPassage {
  rank: number;
  score: number;
  doc_id: string;
  text: string;
}

export interface MeaningCluster {
  cluster_id: number;
  size: number;
  representative_text: string;
  contexts: ClusterContext[];
  similar_passages: SimilarPassage[];
}

export interface KeywordAnalysisResponse {
  keyword: string;
  total_occurrences: number;
  meaning_clusters: MeaningCluster[];
  cross_keyword_similarities: Record<string, number>;
}

export interface MatchResult {
  doc_id: string;
  chunk_index: number;
  text: string;
  best_match: string;
  best_score: number;
  all_scores: Record<string, number>;
}

export interface MatchResponse {
  keyword: string;
  candidate_meanings: string[];
  matches: MatchResult[];
}

export interface CorpusStats {
  total_chunks: number;
  total_documents: number;
  document_ids: string[];
  index_built: boolean;
  embedding_dim: number;
  model_name: string;
}

export interface SimilarityDistribution {
  sample_size: number;
  mean: number;
  std: number;
  min: number;
  max: number;
  percentiles: Record<string, number>;
  histogram: { bin_start: number; bin_end: number; count: number }[];
}

export interface DisambiguationMetric {
  keyword: string;
  accuracy: number;
  weighted_f1: number;
  per_meaning_precision: Record<string, number>;
  per_meaning_recall: Record<string, number>;
  per_meaning_f1: Record<string, number>;
  confusion_matrix: number[][];
  total_samples: number;
}

export interface RetrievalMetric {
  query: string;
  mrr: number;
  precision_at_k: Record<string, number>;
  recall_at_k: Record<string, number>;
  ndcg_at_k: Record<string, number>;
  avg_similarity: number;
  top_score: number;
}

// ---- Training types ----

export interface TrainResponse {
  strategy: string;
  model_path: string;
  training_pairs: number;
  epochs: number;
  seconds: number;
  keywords?: string[];
}

export interface TrainEvalResponse {
  pairs: {
    text_a: string;
    text_b: string;
    expected: number;
    base_score: number;
    trained_score: number;
    base_error: number;
    trained_error: number;
  }[];
  summary: {
    avg_base_error: number;
    avg_trained_error: number;
    error_reduction_pct: number;
    improved: number;
    degraded: number;
    total: number;
  };
}

// ---- Word2Vec types ----

export interface W2VInitResponse {
  vocab_size: number;
  sentences: number;
  vector_size: number;
  seconds: number;
}

export interface W2VQueryResult {
  rank: number;
  score: number;
  doc_id: string;
  text: string;
}

export interface W2VSimilarWord {
  word: string;
  score: number;
}

// ---- Dataset types ----

export interface DatasetSourceInfo {
  dataset_id: string;
  url: string;
  description: string;
  columns?: string[];
  size_mb?: number;
  model?: string;
  vector_dim?: number;
}

export interface DatasetInfo {
  raw_texts: DatasetSourceInfo;
  embeddings: DatasetSourceInfo;
}

export interface DatasetLoadRequest {
  source: "raw" | "embeddings";
  max_docs: number;
  min_text_length: number;
  source_filter?: string;
  build_index: boolean;
}

export interface DatasetLoadResponse {
  documents_loaded?: number;
  documents_skipped?: number;
  documents_created?: number;
  total_chunks?: number;
  chunks_indexed?: number;
  chromadb_vectors?: number;
  index_built: boolean;
  seconds: number;
  source?: string;
}

export interface DatasetPreviewDoc {
  doc_id: string;
  filename: string;
  text_preview: string;
  text_length: number;
}

export interface DatasetPreviewResponse {
  count: number;
  documents: DatasetPreviewDoc[];
}

// ---- Context Analysis types ----

export interface ContextAssociatedWord {
  word: string;
  score: number;
}

export interface ContextExample {
  doc_id: string;
  snippet: string;
}

export interface ContextMeaning {
  cluster_id: number;
  occurrences: number;
  confidence: number;
  associated_words: ContextAssociatedWord[];
  example_contexts: ContextExample[];
}

export interface ContextAnalysisResponse {
  keyword: string;
  total_occurrences: number;
  meanings: ContextMeaning[];
}

// ---- UI State ----

export type EvalSection = "distribution" | "disambiguation" | "retrieval";