Feat: rag tool

llm.go

@@ -11,7 +11,6 @@ import (
 
 var imageAttachmentPath string // Global variable to track image attachment for next message
 var lastImg string             // for ctrl+j
-var RAGMsg = "Retrieved context for user's query:\n"
 
 // containsToolSysMsg checks if the toolSysMsg already exists in the chat body
 func containsToolSysMsg() bool {
@@ -142,22 +141,6 @@ func (lcp LCPCompletion) FormMsg(msg, role string, resume bool) (io.Reader, erro
 		newMsg = *processMessageTag(&newMsg)
 		chatBody.Messages = append(chatBody.Messages, newMsg)
 	}
-	// if rag - add as system message to avoid conflicts with tool usage
-	if !resume && cfg.RAGEnabled {
-		um := chatBody.Messages[len(chatBody.Messages)-1].Content
-		logger.Debug("RAG is enabled, preparing RAG context", "user_message", um)
-		ragResp, err := chatRagUse(um)
-		if err != nil {
-			logger.Error("failed to form a rag msg", "error", err)
-			return nil, err
-		}
-		logger.Debug("RAG response received", "response_len", len(ragResp),
-			"response_preview", ragResp[:min(len(ragResp), 100)])
-		// Use system role for RAG context to avoid conflicts with tool usage
-		ragMsg := models.RoleMsg{Role: "system", Content: RAGMsg + ragResp}
-		chatBody.Messages = append(chatBody.Messages, ragMsg)
-		logger.Debug("RAG message added to chat body", "message_count", len(chatBody.Messages))
-	}
 	// sending description of the tools and how to use them
 	if cfg.ToolUse && !resume && role == cfg.UserRole && !containsToolSysMsg() {
 		chatBody.Messages = append(chatBody.Messages, models.RoleMsg{Role: cfg.ToolRole, Content: toolSysMsg})
@@ -301,23 +284,6 @@ func (op LCPChat) FormMsg(msg, role string, resume bool) (io.Reader, error) {
 		logger.Debug("LCPChat FormMsg: added message to chatBody", "role", newMsg.Role,
 			"content_len", len(newMsg.Content), "message_count_after_add", len(chatBody.Messages))
 	}
-	// if rag - add as system message to avoid conflicts with tool usage
-	if !resume && cfg.RAGEnabled {
-		um := chatBody.Messages[len(chatBody.Messages)-1].Content
-		logger.Debug("LCPChat: RAG is enabled, preparing RAG context", "user_message", um)
-		ragResp, err := chatRagUse(um)
-		if err != nil {
-			logger.Error("LCPChat: failed to form a rag msg", "error", err)
-			return nil, err
-		}
-		logger.Debug("LCPChat: RAG response received",
-			"response_len", len(ragResp), "response_preview", ragResp[:min(len(ragResp), 100)])
-		// Use system role for RAG context to avoid conflicts with tool usage
-		ragMsg := models.RoleMsg{Role: "system", Content: RAGMsg + ragResp}
-		chatBody.Messages = append(chatBody.Messages, ragMsg)
-		logger.Debug("LCPChat: RAG message added to chat body", "role", ragMsg.Role,
-			"rag_content_len", len(ragMsg.Content), "message_count_after_rag", len(chatBody.Messages))
-	}
 	filteredMessages, _ := filterMessagesForCurrentCharacter(chatBody.Messages)
 	// openai /v1/chat does not support custom roles; needs to be user, assistant, system
 	// Add persona suffix to the last user message to indicate who the assistant should reply as
@@ -389,22 +355,6 @@ func (ds DeepSeekerCompletion) FormMsg(msg, role string, resume bool) (io.Reader
 		newMsg = *processMessageTag(&newMsg)
 		chatBody.Messages = append(chatBody.Messages, newMsg)
 	}
-	// if rag - add as system message to avoid conflicts with tool usage
-	if !resume && cfg.RAGEnabled {
-		um := chatBody.Messages[len(chatBody.Messages)-1].Content
-		logger.Debug("DeepSeekerCompletion: RAG is enabled, preparing RAG context", "user_message", um)
-		ragResp, err := chatRagUse(um)
-		if err != nil {
-			logger.Error("DeepSeekerCompletion: failed to form a rag msg", "error", err)
-			return nil, err
-		}
-		logger.Debug("DeepSeekerCompletion: RAG response received",
-			"response_len", len(ragResp), "response_preview", ragResp[:min(len(ragResp), 100)])
-		// Use system role for RAG context to avoid conflicts with tool usage
-		ragMsg := models.RoleMsg{Role: "system", Content: RAGMsg + ragResp}
-		chatBody.Messages = append(chatBody.Messages, ragMsg)
-		logger.Debug("DeepSeekerCompletion: RAG message added to chat body", "message_count", len(chatBody.Messages))
-	}
 	// sending description of the tools and how to use them
 	if cfg.ToolUse && !resume && role == cfg.UserRole && !containsToolSysMsg() {
 		chatBody.Messages = append(chatBody.Messages, models.RoleMsg{Role: cfg.ToolRole, Content: toolSysMsg})
@@ -474,22 +424,6 @@ func (ds DeepSeekerChat) FormMsg(msg, role string, resume bool) (io.Reader, erro
 		newMsg = *processMessageTag(&newMsg)
 		chatBody.Messages = append(chatBody.Messages, newMsg)
 	}
-	// if rag - add as system message to avoid conflicts with tool usage
-	if !resume && cfg.RAGEnabled {
-		um := chatBody.Messages[len(chatBody.Messages)-1].Content
-		logger.Debug("RAG is enabled, preparing RAG context", "user_message", um)
-		ragResp, err := chatRagUse(um)
-		if err != nil {
-			logger.Error("failed to form a rag msg", "error", err)
-			return nil, err
-		}
-		logger.Debug("RAG response received", "response_len", len(ragResp),
-			"response_preview", ragResp[:min(len(ragResp), 100)])
-		// Use system role for RAG context to avoid conflicts with tool usage
-		ragMsg := models.RoleMsg{Role: "system", Content: RAGMsg + ragResp}
-		chatBody.Messages = append(chatBody.Messages, ragMsg)
-		logger.Debug("RAG message added to chat body", "message_count", len(chatBody.Messages))
-	}
 	// Create copy of chat body with standardized user role
 	filteredMessages, _ := filterMessagesForCurrentCharacter(chatBody.Messages)
 	// Add persona suffix to the last user message to indicate who the assistant should reply as
@@ -552,22 +486,6 @@ func (or OpenRouterCompletion) FormMsg(msg, role string, resume bool) (io.Reader
 		newMsg = *processMessageTag(&newMsg)
 		chatBody.Messages = append(chatBody.Messages, newMsg)
 	}
-	// if rag - add as system message to avoid conflicts with tool usage
-	if !resume && cfg.RAGEnabled {
-		um := chatBody.Messages[len(chatBody.Messages)-1].Content
-		logger.Debug("RAG is enabled, preparing RAG context", "user_message", um)
-		ragResp, err := chatRagUse(um)
-		if err != nil {
-			logger.Error("failed to form a rag msg", "error", err)
-			return nil, err
-		}
-		logger.Debug("RAG response received", "response_len",
-			len(ragResp), "response_preview", ragResp[:min(len(ragResp), 100)])
-		// Use system role for RAG context to avoid conflicts with tool usage
-		ragMsg := models.RoleMsg{Role: "system", Content: RAGMsg + ragResp}
-		chatBody.Messages = append(chatBody.Messages, ragMsg)
-		logger.Debug("RAG message added to chat body", "message_count", len(chatBody.Messages))
-	}
 	// sending description of the tools and how to use them
 	if cfg.ToolUse && !resume && role == cfg.UserRole && !containsToolSysMsg() {
 		chatBody.Messages = append(chatBody.Messages, models.RoleMsg{Role: cfg.ToolRole, Content: toolSysMsg})
@@ -670,22 +588,6 @@ func (or OpenRouterChat) FormMsg(msg, role string, resume bool) (io.Reader, erro
 		newMsg = *processMessageTag(&newMsg)
 		chatBody.Messages = append(chatBody.Messages, newMsg)
 	}
-	// if rag - add as system message to avoid conflicts with tool usage
-	if !resume && cfg.RAGEnabled {
-		um := chatBody.Messages[len(chatBody.Messages)-1].Content
-		logger.Debug("RAG is enabled, preparing RAG context", "user_message", um)
-		ragResp, err := chatRagUse(um)
-		if err != nil {
-			logger.Error("failed to form a rag msg", "error", err)
-			return nil, err
-		}
-		logger.Debug("RAG response received", "response_len", len(ragResp),
-			"response_preview", ragResp[:min(len(ragResp), 100)])
-		// Use system role for RAG context to avoid conflicts with tool usage
-		ragMsg := models.RoleMsg{Role: "system", Content: RAGMsg + ragResp}
-		chatBody.Messages = append(chatBody.Messages, ragMsg)
-		logger.Debug("RAG message added to chat body", "message_count", len(chatBody.Messages))
-	}
 	// Create copy of chat body with standardized user role
 	filteredMessages, _ := filterMessagesForCurrentCharacter(chatBody.Messages)
 	// Add persona suffix to the last user message to indicate who the assistant should reply as

rag/rag.go

@@ -9,6 +9,8 @@ import (
 	"log/slog"
 	"os"
 	"path"
+	"regexp"
+	"sort"
 	"strings"
 	"sync"
 
@@ -195,3 +197,309 @@ func (r *RAG) ListLoaded() ([]string, error) {
 func (r *RAG) RemoveFile(filename string) error {
 	return r.storage.RemoveEmbByFileName(filename)
 }
+
+var (
+	queryRefinementPattern = regexp.MustCompile(`(?i)(based on my (vector db|vector db|vector database|rags?|past (conversations?|chat|messages?))|from my (files?|documents?|data|information|memory)|search (in|my) (vector db|database|rags?)|rag search for)`)
+	importantKeywords      = []string{"project", "architecture", "code", "file", "chat", "conversation", "topic", "summary", "details", "history", "previous", "my", "user", "me"}
+	stopWords              = []string{"the", "a", "an", "and", "or", "but", "in", "on", "at", "to", "for", "of", "with", "by", "from", "up", "down", "left", "right"}
+)
+
+func (r *RAG) RefineQuery(query string) string {
+	original := query
+	query = strings.TrimSpace(query)
+
+	if len(query) == 0 {
+		return original
+	}
+
+	if len(query) <= 3 {
+		return original
+	}
+
+	query = strings.ToLower(query)
+
+	for _, stopWord := range stopWords {
+		wordPattern := `\b` + stopWord + `\b`
+		re := regexp.MustCompile(wordPattern)
+		query = re.ReplaceAllString(query, "")
+	}
+
+	query = strings.TrimSpace(query)
+
+	if len(query) < 5 {
+		return original
+	}
+
+	if queryRefinementPattern.MatchString(original) {
+		cleaned := queryRefinementPattern.ReplaceAllString(original, "")
+		cleaned = strings.TrimSpace(cleaned)
+		if len(cleaned) >= 5 {
+			return cleaned
+		}
+	}
+
+	query = r.extractImportantPhrases(query)
+
+	if len(query) < 5 {
+		return original
+	}
+
+	return query
+}
+
+func (r *RAG) extractImportantPhrases(query string) string {
+	words := strings.Fields(query)
+
+	var important []string
+	for _, word := range words {
+		word = strings.Trim(word, ".,!?;:'\"()[]{}")
+
+		isImportant := false
+		for _, kw := range importantKeywords {
+			if strings.Contains(strings.ToLower(word), kw) {
+				isImportant = true
+				break
+			}
+		}
+
+		if isImportant || len(word) > 3 {
+			important = append(important, word)
+		}
+	}
+
+	if len(important) == 0 {
+		return query
+	}
+
+	return strings.Join(important, " ")
+}
+
+func (r *RAG) GenerateQueryVariations(query string) []string {
+	variations := []string{query}
+
+	if len(query) < 5 {
+		return variations
+	}
+
+	parts := strings.Fields(query)
+	if len(parts) == 0 {
+		return variations
+	}
+
+	if len(parts) >= 2 {
+		trimmed := strings.Join(parts[:len(parts)-1], " ")
+		if len(trimmed) >= 5 {
+			variations = append(variations, trimmed)
+		}
+	}
+
+	if len(parts) >= 2 {
+		trimmed := strings.Join(parts[1:], " ")
+		if len(trimmed) >= 5 {
+			variations = append(variations, trimmed)
+		}
+	}
+
+	if !strings.HasSuffix(query, " explanation") {
+		variations = append(variations, query+" explanation")
+	}
+	if !strings.HasPrefix(query, "what is ") {
+		variations = append(variations, "what is "+query)
+	}
+	if !strings.HasSuffix(query, " details") {
+		variations = append(variations, query+" details")
+	}
+	if !strings.HasSuffix(query, " summary") {
+		variations = append(variations, query+" summary")
+	}
+
+	return variations
+}
+
+func (r *RAG) RerankResults(results []models.VectorRow, query string) []models.VectorRow {
+	type scoredResult struct {
+		row      models.VectorRow
+		distance float32
+	}
+
+	scored := make([]scoredResult, 0, len(results))
+
+	for i := range results {
+		row := results[i]
+
+		score := float32(0)
+
+		rawTextLower := strings.ToLower(row.RawText)
+		queryLower := strings.ToLower(query)
+
+		if strings.Contains(rawTextLower, queryLower) {
+			score += 10
+		}
+
+		queryWords := strings.Fields(queryLower)
+		matchCount := 0
+		for _, word := range queryWords {
+			if len(word) > 2 && strings.Contains(rawTextLower, word) {
+				matchCount++
+			}
+		}
+		if len(queryWords) > 0 {
+			score += float32(matchCount) / float32(len(queryWords)) * 5
+		}
+
+		if row.FileName == "chat" || strings.Contains(strings.ToLower(row.FileName), "conversation") {
+			score += 3
+		}
+
+		distance := row.Distance - score/100
+
+		scored = append(scored, scoredResult{row: row, distance: distance})
+	}
+
+	sort.Slice(scored, func(i, j int) bool {
+		return scored[i].distance < scored[j].distance
+	})
+
+	unique := make([]models.VectorRow, 0)
+	seen := make(map[string]bool)
+
+	for i := range scored {
+		if !seen[scored[i].row.Slug] {
+			seen[scored[i].row.Slug] = true
+			unique = append(unique, scored[i].row)
+		}
+	}
+
+	if len(unique) > 10 {
+		unique = unique[:10]
+	}
+
+	return unique
+}
+
+func (r *RAG) SynthesizeAnswer(results []models.VectorRow, query string) (string, error) {
+	if len(results) == 0 {
+		return "No relevant information found in the vector database.", nil
+	}
+
+	var contextBuilder strings.Builder
+	contextBuilder.WriteString("User Query: ")
+	contextBuilder.WriteString(query)
+	contextBuilder.WriteString("\n\nRetrieved Context:\n")
+
+	for i, row := range results {
+		contextBuilder.WriteString(fmt.Sprintf("[Source %d: %s]\n", i+1, row.FileName))
+		contextBuilder.WriteString(row.RawText)
+		contextBuilder.WriteString("\n\n")
+	}
+
+	contextBuilder.WriteString("Instructions: ")
+	contextBuilder.WriteString("Based on the retrieved context above, provide a concise, coherent answer to the user's query. ")
+	contextBuilder.WriteString("Extract only the most relevant information. ")
+	contextBuilder.WriteString("If no relevant information is found, state that clearly. ")
+	contextBuilder.WriteString("Cite sources by filename when relevant. ")
+	contextBuilder.WriteString("Do not include unnecessary preamble or explanations.")
+
+	synthesisPrompt := contextBuilder.String()
+
+	emb, err := r.LineToVector(synthesisPrompt)
+	if err != nil {
+		r.logger.Error("failed to embed synthesis prompt", "error", err)
+		return "", err
+	}
+
+	embResp := &models.EmbeddingResp{
+		Embedding: emb,
+		Index:     0,
+	}
+
+	topResults, err := r.SearchEmb(embResp)
+	if err != nil {
+		r.logger.Error("failed to search for synthesis context", "error", err)
+		return "", err
+	}
+
+	if len(topResults) > 0 && topResults[0].RawText != synthesisPrompt {
+		return topResults[0].RawText, nil
+	}
+
+	var finalAnswer strings.Builder
+	finalAnswer.WriteString("Based on the retrieved context:\n\n")
+
+	for i, row := range results {
+		if i >= 5 {
+			break
+		}
+		finalAnswer.WriteString(fmt.Sprintf("- From %s: %s\n", row.FileName, truncateString(row.RawText, 200)))
+	}
+
+	return finalAnswer.String(), nil
+}
+
+func truncateString(s string, maxLen int) string {
+	if len(s) <= maxLen {
+		return s
+	}
+	return s[:maxLen] + "..."
+}
+
+func (r *RAG) Search(query string, limit int) ([]models.VectorRow, error) {
+	refined := r.RefineQuery(query)
+	variations := r.GenerateQueryVariations(refined)
+
+	allResults := make([]models.VectorRow, 0)
+	seen := make(map[string]bool)
+
+	for _, q := range variations {
+		emb, err := r.LineToVector(q)
+		if err != nil {
+			r.logger.Error("failed to embed query variation", "error", err, "query", q)
+			continue
+		}
+
+		embResp := &models.EmbeddingResp{
+			Embedding: emb,
+			Index:     0,
+		}
+
+		results, err := r.SearchEmb(embResp)
+		if err != nil {
+			r.logger.Error("failed to search embeddings", "error", err, "query", q)
+			continue
+		}
+
+		for _, row := range results {
+			if !seen[row.Slug] {
+				seen[row.Slug] = true
+				allResults = append(allResults, row)
+			}
+		}
+	}
+
+	reranked := r.RerankResults(allResults, query)
+
+	if len(reranked) > limit {
+		reranked = reranked[:limit]
+	}
+
+	return reranked, nil
+}
+
+var (
+	ragInstance *RAG
+	ragOnce     sync.Once
+)
+
+func Init(c *config.Config, l *slog.Logger, s storage.FullRepo) error {
+	ragOnce.Do(func() {
+		if c == nil || l == nil || s == nil {
+			return
+		}
+		ragInstance = New(l, s, c)
+	})
+	return nil
+}
+
+func GetInstance() *RAG {
+	return ragInstance
+}

tools.go

@@ -16,6 +16,7 @@ import (
 	"sync"
 	"time"
 
+	"gf-lt/rag"
 	"github.com/GrailFinder/searchagent/searcher"
 )
 
@@ -58,9 +59,9 @@ Your current tools:
 "when_to_use": "when asked to search the web for information; returns clean summary without html,css and other web elements; limit is optional (default 3)"
 },
 {
-"name":"websearch_raw",
+"name":"rag_search",
 "args": ["query", "limit"],
-"when_to_use": "when asked to search the web for information; returns raw data as is without processing; limit is optional (default 3)"
+"when_to_use": "when asked to search the local document database for information; performs query refinement, semantic search, reranking, and synthesis; returns clean summary with sources; limit is optional (default 3)"
 },
 {
 "name":"read_url",
@@ -146,6 +147,7 @@ under the topic: Adam's number is stored:
 After that you are free to respond to the user.
 `
 	webSearchSysPrompt = `Summarize the web search results, extracting key information and presenting a concise answer. Provide sources and URLs where relevant.`
+	ragSearchSysPrompt = `Synthesize the document search results, extracting key information and presenting a concise answer. Provide sources and document IDs where relevant.`
 	readURLSysPrompt   = `Extract and summarize the content from the webpage. Provide key information, main points, and any relevant details.`
 	summarySysPrompt   = `Please provide a concise summary of the following conversation. Focus on key points, decisions, and actions. Provide only the summary, no additional commentary.`
 	basicCard          = &models.CharCard{
@@ -170,6 +172,10 @@ func init() {
 		panic("failed to init seachagent; error: " + err.Error())
 	}
 	WebSearcher = sa
+
+	if err := rag.Init(cfg, logger, store); err != nil {
+		logger.Warn("failed to init rag; rag_search tool will not be available", "error", err)
+	}
 }
 
 // getWebAgentClient returns a singleton AgentClient for web agents.
@@ -196,6 +202,8 @@ func getWebAgentClient() *agent.AgentClient {
 func registerWebAgents() {
 	webAgentsOnce.Do(func() {
 		client := getWebAgentClient()
+		// Register rag_search agent
+		agent.Register("rag_search", agent.NewWebAgentB(client, ragSearchSysPrompt))
 		// Register websearch agent
 		agent.Register("websearch", agent.NewWebAgentB(client, webSearchSysPrompt))
 		// Register read_url agent
@@ -239,6 +247,48 @@ func websearch(args map[string]string) []byte {
 	return data
 }
 
+// rag search (searches local document database)
+func ragsearch(args map[string]string) []byte {
+	query, ok := args["query"]
+	if !ok || query == "" {
+		msg := "query not provided to rag_search tool"
+		logger.Error(msg)
+		return []byte(msg)
+	}
+	limitS, ok := args["limit"]
+	if !ok || limitS == "" {
+		limitS = "3"
+	}
+	limit, err := strconv.Atoi(limitS)
+	if err != nil || limit == 0 {
+		logger.Warn("ragsearch limit; passed bad value; setting to default (3)",
+			"limit_arg", limitS, "error", err)
+		limit = 3
+	}
+
+	ragInstance := rag.GetInstance()
+	if ragInstance == nil {
+		msg := "rag not initialized; rag_search tool is not available"
+		logger.Error(msg)
+		return []byte(msg)
+	}
+
+	results, err := ragInstance.Search(query, limit)
+	if err != nil {
+		msg := "rag search failed; error: " + err.Error()
+		logger.Error(msg)
+		return []byte(msg)
+	}
+
+	data, err := json.Marshal(results)
+	if err != nil {
+		msg := "failed to marshal rag search result; error: " + err.Error()
+		logger.Error(msg)
+		return []byte(msg)
+	}
+	return data
+}
+
 // web search raw (returns raw data without processing)
 func websearchRaw(args map[string]string) []byte {
 	// make http request return bytes
@@ -997,6 +1047,7 @@ var fnMap = map[string]fnSig{
 	"recall":            recall,
 	"recall_topics":     recallTopics,
 	"memorise":          memorise,
+	"rag_search":        ragsearch,
 	"websearch":         websearch,
 	"websearch_raw":     websearchRaw,
 	"read_url":          readURL,
@@ -1033,6 +1084,28 @@ func callToolWithAgent(name string, args map[string]string) []byte {
 
 // openai style def
 var baseTools = []models.Tool{
+	// rag_search
+	models.Tool{
+		Type: "function",
+		Function: models.ToolFunc{
+			Name:        "rag_search",
+			Description: "Search local document database given query, limit of sources (default 3). Performs query refinement, semantic search, reranking, and synthesis.",
+			Parameters: models.ToolFuncParams{
+				Type:     "object",
+				Required: []string{"query", "limit"},
+				Properties: map[string]models.ToolArgProps{
+					"query": models.ToolArgProps{
+						Type:        "string",
+						Description: "search query",
+					},
+					"limit": models.ToolArgProps{
+						Type:        "string",
+						Description: "limit of the document results",
+					},
+				},
+			},
+		},
+	},
 	// websearch
 	models.Tool{
 		Type: "function",