summaryrefslogtreecommitdiff
path: root/core/engine.go
diff options
context:
space:
mode:
Diffstat (limited to 'core/engine.go')
-rw-r--r--core/engine.go56
1 files changed, 56 insertions, 0 deletions
diff --git a/core/engine.go b/core/engine.go
new file mode 100644
index 0000000..4b6b150
--- /dev/null
+++ b/core/engine.go
@@ -0,0 +1,56 @@
+package core
+
+import (
+ "fmt"
+ "searchEngine/core/analyzer"
+ "searchEngine/core/index"
+ "searchEngine/core/ingest"
+ "searchEngine/core/search"
+ "sync"
+)
+
+type Engine struct {
+ idx *index.InvertedIndex
+}
+
+func NewEngine() *Engine {
+ return &Engine{
+ idx: index.New(),
+ }
+}
+
+func (e *Engine) Index() *index.InvertedIndex {
+ return e.idx
+}
+
+func (e *Engine) IndexDump(filePath string, numWorkers int) error {
+ if numWorkers <= 0 {
+ numWorkers = 4
+ }
+
+ jobQueue := make(chan ingest.Page, 100)
+ var wg sync.WaitGroup
+
+ for i := 0; i < numWorkers; i++ {
+ wg.Go(func() {
+ for page := range jobQueue {
+ tokens := analyzer.ProcessText(page.Text)
+ e.idx.Add(page.Id, page.Title, tokens)
+ }
+ })
+ }
+
+ err := ingest.IngestWikiDump(filePath, jobQueue)
+ close(jobQueue)
+ wg.Wait()
+
+ if err != nil {
+ return fmt.Errorf("failed to ingest wiki dump: %w", err)
+ }
+
+ return nil
+}
+
+func (e *Engine) Search(query string) []string {
+ return search.Search(e.idx, query)
+}