Skip to content

RAGFlow is an agent-native database: a Retrieval-Augmented Generation (RAG) engine that grew into context infrastructure.

refactor: replace Split in loops with more efficient SplitSeq (#18248)

questfever52519e2fd7e5
diff --git a/internal/agent/tool/agentic_search.go b/internal/agent/tool/agentic_search.go
index 8510a6c431..3cbd2cb43c 100644
--- a/internal/agent/tool/agentic_search.go
+++ b/internal/agent/tool/agentic_search.go
@@ -173,14 +173,14 @@ func splitKeywords(keywords string) []string {
return nil
}
kwds := make([]string, 0, 8)
- for _, k := range strings.Split(keywords, ",") {
+ for k := range strings.SplitSeq(keywords, ",") {
if k = strings.TrimSpace(k); k != "" {
kwds = append(kwds, strings.ToLower(k))
}
}
if len(kwds) < 3 {
words := make([]string, 0, 8)
- for _, w := range strings.Split(keywords, " ") {
+ for w := range strings.SplitSeq(keywords, " ") {
if w = strings.TrimSpace(w); w != "" {
words = append(words, strings.ToLower(w))
}
diff --git a/internal/agent/tool/duckduckgo.go b/internal/agent/tool/duckduckgo.go
index c9a8ffe90a..191d9a06f4 100644
--- a/internal/agent/tool/duckduckgo.go
+++ b/internal/agent/tool/duckduckgo.go
@@ -513,7 +513,7 @@ func hasClassToken(n *xhtml.Node, want string) bool {
if a.Key != "class" {
continue
}
- for _, token := range strings.Fields(a.Val) {
+ for token := range strings.FieldsSeq(a.Val) {
if token == want {
return true
}
diff --git a/internal/deepdoc/parser/pdf/table/table_cells.go b/internal/deepdoc/parser/pdf/table/table_cells.go
index 1e6c0b1d8c..4c6d2dce9f 100644
--- a/internal/deepdoc/parser/pdf/table/table_cells.go
+++ b/internal/deepdoc/parser/pdf/table/table_cells.go
@@ -224,8 +224,8 @@ func simpleTokenCount(text string) int {
}
}
// Also count space-separated words.
- words := strings.Fields(text)
- for _, w := range words {
+ words := strings.FieldsSeq(text)
+ for w := range words {
if !containsCJK(w) {
count++
}
diff --git a/internal/deepdoc/parser/pdf/util/position.go b/internal/deepdoc/parser/pdf/util/position.go
index 9f0c95d095..100a75cb3a 100644
--- a/internal/deepdoc/parser/pdf/util/position.go
+++ b/internal/deepdoc/parser/pdf/util/position.go
@@ -45,7 +45,7 @@ func ExtractPositions(text string) []pdf.Position {
// Parse page range
var pageNums []int
- for _, p := range strings.Split(parts[0], "-") {
+ for p := range strings.SplitSeq(parts[0], "-") {
n, err := strconv.Atoi(p)
if err != nil {
slog.Warn("ExtractPositions: invalid page number in tag", "tag", tag, "part", p, "err", err)
diff --git a/internal/engine/elasticsearch/chunk.go b/internal/engine/elasticsearch/chunk.go
index 9255d1f29c..735d121d7f 100644
--- a/internal/engine/elasticsearch/chunk.go
+++ b/internal/engine/elasticsearch/chunk.go
@@ -2163,7 +2163,7 @@ func (e *Engine) GetAggregation(chunks []map[string]interface{}, fieldName strin
if fieldName == "tag_kwd" && strings.Contains(valueStr, "###") {
separator = "###"
}
- for _, tag := range strings.Split(valueStr, separator) {
+ for tag := range strings.SplitSeq(valueStr, separator) {
countElasticsearchAggregationTag(tagCounts, tag)
}
continue
diff --git a/internal/engine/elasticsearch/meta_filter.go b/internal/engine/elasticsearch/meta_filter.go
index 0cbf1eb9da..331c0a0902 100644
--- a/internal/engine/elasticsearch/meta_filter.go
+++ b/internal/engine/elasticsearch/meta_filter.go
@@ -482,8 +482,8 @@ func csvOrList(value interface{}, flt map[string]interface{}) []interface{} {
}
} else {
// Comma-separated
- parts := strings.Split(v, ",")
- for _, p := range parts {
+ parts := strings.SplitSeq(v, ",")
+ for p := range parts {
trimmed := strings.TrimSpace(p)
if trimmed != "" {
members = append(members, strings.ToLower(trimmed))
diff --git a/internal/engine/infinity/chunk.go b/internal/engine/infinity/chunk.go
index 6d1cecb243..c8a13e5953 100644
--- a/internal/engine/infinity/chunk.go
+++ b/internal/engine/infinity/chunk.go
@@ -1818,7 +1818,7 @@ func (e *Engine) GetAggregation(chunks []map[string]interface{}, fieldName strin
var tags []string
// Split by "###" for tag_kwd field
if fieldName == "tag_kwd" && strings.Contains(valueStr, "###") {
- for _, tag := range strings.Split(valueStr, "###") {
+ for tag := range strings.SplitSeq(valueStr, "###") {
tag = strings.TrimSpace(tag)
if tag != "" {
tags = append(tags, tag)
@@ -1826,7 +1826,7 @@ func (e *Engine) GetAggregation(chunks []map[string]interface{}, fieldName strin
}
} else {
// Fallback to comma separation
- for _, tag := range strings.Split(valueStr, ",") {
+ for tag := range strings.SplitSeq(valueStr, ",") {
tag = strings.TrimSpace(tag)
if tag != "" {
tags = append(tags, tag)
diff --git a/internal/engine/infinity/meta_filter.go b/internal/engine/infinity/meta_filter.go
index 1422e2bc5d..cca63b17cb 100644
--- a/internal/engine/infinity/meta_filter.go
+++ b/internal/engine/infinity/meta_filter.go
@@ -385,8 +385,8 @@ func csvOrList(value interface{}, flt map[string]interface{}) []interface{} {
members = parsed
}
} else {
- parts := strings.Split(v, ",")
- for _, p := range parts {
+ parts := strings.SplitSeq(v, ",")
+ for p := range parts {
trimmed := strings.TrimSpace(p)
if trimmed != "" {
members = append(members, strings.ToLower(trimmed))
diff --git a/internal/engine/infinity/sql.go b/internal/engine/infinity/sql.go
index 3c2b8745d0..e98a676c90 100644
--- a/internal/engine/infinity/sql.go
+++ b/internal/engine/infinity/sql.go
@@ -84,7 +84,7 @@ func loadFieldMapping(mappingFileName string) (aliasToActual map[string]string,
continue
}
var firstAlias string
- for _, raw := range strings.Split(info.Comment, ",") {
+ for raw := range strings.SplitSeq(info.Comment, ",") {
alias := strings.TrimSpace(raw)
if alias == "" {
continue
@@ -216,7 +216,7 @@ func parsePsqlTable(output string) *psqlResult {
return res
}
- for _, raw := range strings.Split(lines[0], "|") {
+ for raw := range strings.SplitSeq(lines[0], "|") {
if col := strings.TrimSpace(raw); col != "" {
res.Columns = append(res.Columns, col)
}
diff --git a/internal/engine/serenedb/common.go b/internal/engine/serenedb/common.go
index d504c1dcdb..31d510e970 100644
--- a/internal/engine/serenedb/common.go
+++ b/internal/engine/serenedb/common.go
@@ -197,7 +197,7 @@ func stripESQuery(matchingText string) string {
txt = esSyntaxRe.ReplaceAllString(txt, " ")
seen := map[string]struct{}{}
var out []string
- for _, t := range strings.Fields(txt) {
+ for t := range strings.FieldsSeq(txt) {
if _, ok := seen[t]; ok {
continue
}
diff --git a/internal/harness/core/middlewares/filesystem/filesystem.go b/internal/harness/core/middlewares/filesystem/filesystem.go
index f4d5d582fd..8658c1c937 100644
--- a/internal/harness/core/middlewares/filesystem/filesystem.go
+++ b/internal/harness/core/middlewares/filesystem/filesystem.go
@@ -253,7 +253,7 @@ func formatGrepResult(result, outputMode string) (string, error) {
return fmt.Sprintf("%d matches", lines), nil
case "files":
unique := make(map[string]bool)
- for _, line := range strings.Split(result, "\n") {
+ for line := range strings.SplitSeq(result, "\n") {
if line == "" {
continue
}
diff --git a/internal/harness/core/middlewares/skill/skill.go b/internal/harness/core/middlewares/skill/skill.go
index 5d259530f1..5b18eaf5fa 100644
--- a/internal/harness/core/middlewares/skill/skill.go
+++ b/internal/harness/core/middlewares/skill/skill.go
@@ -160,7 +160,7 @@ func parseSkill(content string) *Config {
if len(parts) == 2 {
front := strings.TrimSpace(parts[0])
body := strings.TrimSpace(parts[1])
- for _, line := range strings.Split(front, "\n") {
+ for line := range strings.SplitSeq(front, "\n") {
line = strings.TrimSpace(line)
if strings.HasPrefix(line, "name:") {
cfg.Name = strings.TrimSpace(line[5:])
diff --git a/internal/harness/core/middlewares/summarization/compactor.go b/internal/harness/core/middlewares/summarization/compactor.go
index dd51992385..141864ac9d 100644
--- a/internal/harness/core/middlewares/summarization/compactor.go
+++ b/internal/harness/core/middlewares/summarization/compactor.go
@@ -295,13 +295,13 @@ func ApplySupersede(msgs []*schema.Message) []*schema.Message {
}
func extractFilePath(content string) string {
- lines := strings.Split(content, "\n")
- for _, line := range lines {
+ lines := strings.SplitSeq(content, "\n")
+ for line := range lines {
line = strings.TrimSpace(line)
if strings.HasPrefix(line, "/") || strings.HasPrefix(line, "./") ||
strings.HasPrefix(line, "../") || strings.Contains(line, "/") {
- parts := strings.Fields(line)
- for _, p := range parts {
+ parts := strings.FieldsSeq(line)
+ for p := range parts {
if strings.Contains(p, ".") || strings.Contains(p, "/") {
return p
}
diff --git a/internal/harness/graph/graph/state.go b/internal/harness/graph/graph/state.go
index 85f67be4dd..c627d612f0 100644
--- a/internal/harness/graph/graph/state.go
+++ b/internal/harness/graph/graph/state.go
@@ -117,8 +117,8 @@ func parseAnnotation(tag string) (*Annotation, error) {
Metadata: make(map[string]interface{}),
}
- pairs := strings.Split(tag, ",")
- for _, pair := range pairs {
+ pairs := strings.SplitSeq(tag, ",")
+ for pair := range pairs {
kv := strings.SplitN(pair, "=", 2)
if len(kv) != 2 {
// Could be a boolean flag
diff --git a/internal/ingestion/component/knowledge_compiler/wiki/page.go b/internal/ingestion/component/knowledge_compiler/wiki/page.go
index 28cc60f203..74460cb83f 100644
--- a/internal/ingestion/component/knowledge_compiler/wiki/page.go
+++ b/internal/ingestion/component/knowledge_compiler/wiki/page.go
@@ -109,7 +109,7 @@ func ParseOutline(md string) []Section {
current = nil
}
}
- for _, line := range strings.Split(md, "\n") {
+ for line := range strings.SplitSeq(md, "\n") {
trimmed := strings.TrimRight(line, "\r")
if lvl := headingLevel(trimmed); lvl > 0 {
flush()
@@ -275,7 +275,7 @@ func buildWikiPageProducts(tenantID, docID string, pages []wikiPageResult) []com
// firstHeading returns the first Markdown "# "-prefixed line, trimmed.
func firstHeading(md string) string {
- for _, line := range strings.Split(md, "\n") {
+ for line := range strings.SplitSeq(md, "\n") {
s := strings.TrimSpace(line)
if strings.HasPrefix(s, "# ") {
return strings.TrimSpace(strings.TrimPrefix(s, "# "))
@@ -286,7 +286,7 @@ func firstHeading(md string) string {
// firstParagraph returns the first non-empty, non-heading line, capped.
func firstParagraph(md string) string {
- for _, line := range strings.Split(md, "\n") {
+ for line := range strings.SplitSeq(md, "\n") {
s := strings.TrimSpace(line)
if s == "" || strings.HasPrefix(s, "#") {
continue
diff --git a/internal/ingestion/component/knowledge_compiler/wiki/wiki.go b/internal/ingestion/component/knowledge_compiler/wiki/wiki.go
index d5caebdf07..a004e49223 100644
--- a/internal/ingestion/component/knowledge_compiler/wiki/wiki.go
+++ b/internal/ingestion/component/knowledge_compiler/wiki/wiki.go
@@ -1182,7 +1182,7 @@ func wikiMarkdownDropsContent(mergedMD, sourceMD string) bool {
}
func wikiMarkdownTitle(md string) string {
- for _, line := range strings.Split(md, "\n") {
+ for line := range strings.SplitSeq(md, "\n") {
line = strings.TrimSpace(line)
if strings.HasPrefix(line, "# ") {
return strings.TrimSpace(strings.TrimPrefix(line, "# "))
diff --git a/internal/service/nlp/reranker.go b/internal/service/nlp/reranker.go
index e9d23ed240..d0d1109562 100644
--- a/internal/service/nlp/reranker.go
+++ b/internal/service/nlp/reranker.go
@@ -536,7 +536,7 @@ func extractContentTokens(fields map[string]interface{}, cfield string) []string
// Split by whitespace to get individual tokens
seen := make(map[string]bool)
var result []string
- for _, t := range strings.Fields(v) {
+ for t := range strings.FieldsSeq(v) {
if !seen[t] {
seen[t] = true
result = append(result, t)
@@ -553,7 +553,7 @@ func extractTitleTokens(fields map[string]interface{}) []string {
}
// NOTE: Do NOT call RemoveRedundantSpaces here - it removes spaces between Chinese chars
var result []string
- for _, t := range strings.Fields(v) {
+ for t := range strings.FieldsSeq(v) {
if t != "" {
result = append(result, t)
}
@@ -568,7 +568,7 @@ func extractQuestionTokens(fields map[string]interface{}) []string {
return []string{}
}
var result []string
- for _, t := range strings.Fields(v) {
+ for t := range strings.FieldsSeq(v) {
if t != "" {
result = append(result, t)
}
diff --git a/internal/service/nlp/retrieval.go b/internal/service/nlp/retrieval.go
index 63b5a21653..0a09754d29 100644
--- a/internal/service/nlp/retrieval.go
+++ b/internal/service/nlp/retrieval.go
@@ -740,7 +740,7 @@ func (s *RetrievalService) Search(ctx context.Context, req *RetrievalSearchReque
for _, k := range keywords {
kwds[k] = struct{}{}
fgToken, _ := tokenizer.FineGrainedTokenize(k)
- for _, kk := range strings.Fields(fgToken) {
+ for kk := range strings.FieldsSeq(fgToken) {
if len(kk) < 2 {
continue
}
@@ -1088,7 +1088,7 @@ func (s *RetrievalService) PruneDeletedChunks(ctx context.Context, result *Retri
func buildIndexNames(tenantIDs []string) []string {
var indexNames []string
for _, tid := range tenantIDs {
- for _, part := range strings.Split(tid, ",") {
+ for part := range strings.SplitSeq(tid, ",") {
part = strings.TrimSpace(part)
if part != "" {
indexNames = append(indexNames, fmt.Sprintf("ragflow_%s", part))
diff --git a/internal/service/nlp/term_weight.go b/internal/service/nlp/term_weight.go
index 9c45b4ed39..362f656b34 100644
--- a/internal/service/nlp/term_weight.go
+++ b/internal/service/nlp/term_weight.go
@@ -134,7 +134,7 @@ func (d *TermWeightDealer) Pretoken(txt string, num bool, stpwd bool) []string {
tokenized = txt
}
- for _, t := range strings.Fields(tokenized) {
+ for t := range strings.FieldsSeq(tokenized) {
tk := t
// Check stop words
if stpwd {
@@ -258,7 +258,7 @@ func (d *TermWeightDealer) Split(txt string) []string {
txt = regexp.MustCompile("[ \\t]+").ReplaceAllString(txt, " ")
txt = strings.TrimSpace(txt)
- for _, t := range strings.Split(txt, " ") {
+ for t := range strings.SplitSeq(txt, " ") {
t = strings.TrimSpace(t)
if t == "" {
continue
diff --git a/internal/syncer/connector/google_drive.go b/internal/syncer/connector/google_drive.go
index cef4f821b6..9f3c712ad4 100644
--- a/internal/syncer/connector/google_drive.go
+++ b/internal/syncer/connector/google_drive.go
@@ -1240,7 +1240,7 @@ func (c *GoogleDriveConnector) effectiveIncludeSharedWithMe() bool {
func splitCommaList(value string) []string {
values := []string{}
- for _, part := range strings.Split(value, ",") {
+ for part := range strings.SplitSeq(value, ",") {
part = strings.TrimSpace(part)
if part != "" {
values = append(values, part)