gslin/go-ir-playground
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Use os.Args[1] as query string, also lowercase all the time.

Browse Source
This commit is contained in:
Gea-Suan Lin
2024-02-11 14:32:09 +08:00
parent d6e1c1dbf5
commit 1e5b1dcf9a
1 changed files with 4 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files

7
cmd/ir-tfidf/main.go
View File

@@ -18,7 +18,7 @@ func main() {
df := make(map[string]int) df := make(map[string]int)
for _, article := range articles { for _, article := range articles {
str := article.Title + "\n" + article.Body str := strings.ToLower(article.Title + "\n" + article.Body)
bag := tokenizer.Tokenize(str) bag := tokenizer.Tokenize(str)
tokens[article.Id] = bag tokens[article.Id] = bag
@@ -37,11 +37,12 @@ func main() {
fmt.Println("TF & DF Built") fmt.Println("TF & DF Built")
q := os.Args[1:] q := strings.ToLower(os.Args[1])
q_tokens := tokenizer.Tokenize(q)
for _, article := range articles { for _, article := range articles {
var score float64 = 0.0 var score float64 = 0.0
for _, w := range q { for _, w := range q_tokens {
if tf[w] != nil { if tf[w] != nil {
score += float64(tf[w][article.Id]) * math.Log2(float64(len(articles) / df[w])) score += float64(tf[w][article.Id]) * math.Log2(float64(len(articles) / df[w]))
} }