From 1ced451c2723c36b7e0cb70e6c0755cdbad457c3 Mon Sep 17 00:00:00 2001
From: sigoden <sigoden@gmail.com>
Date: Thu, 27 Jun 2024 07:51:30 +0800
Subject: refactor: agent rag use `documents` field other than `embeddings` dir
 (#658)

---
 src/rag/mod.rs | 12 ++++++++----
 1 file changed, 8 insertions(+), 4 deletions(-)

(limited to 'src/rag')

diff --git a/src/rag/mod.rs b/src/rag/mod.rs
index 3139b7e..f01fc0c 100644
--- a/src/rag/mod.rs
+++ b/src/rag/mod.rs
@@ -56,7 +56,7 @@ impl Rag {
         let mut rag = Self::create(config, name, save_path, data)?;
         let mut paths = doc_paths.to_vec();
         if paths.is_empty() {
-            paths = add_doc_paths()?;
+            paths = add_document_paths()?;
         };
         debug!("doc paths: {paths:?}");
         let loaders = config.read().rag_document_loaders.clone();
@@ -233,7 +233,7 @@ impl Rag {
         progress(&spinner, "Gathering paths".into());
         for path in paths {
             let path = path.as_ref();
-            if path.starts_with("http://") || path.starts_with("https://") {
+            if Self::is_url_path(path) {
                 if let Some(path) = path.strip_suffix("**") {
                     new_paths.push((path.to_string(), RECURSIVE_URL_LOADER.into()));
                 } else {
@@ -337,6 +337,10 @@ impl Rag {
         Ok(())
     }
 
+    pub fn is_url_path(path: &str) -> bool {
+        path.starts_with("http://") || path.starts_with("https://")
+    }
+
     async fn hybird_search(
         &self,
         query: &str,
@@ -628,8 +632,8 @@ fn set_chunk_overlay(default_value: usize) -> Result<usize> {
     value.parse().map_err(|_| anyhow!("Invalid chunk_overlay"))
 }
 
-fn add_doc_paths() -> Result<Vec<String>> {
-    let text = Text::new("Add document paths:")
+fn add_document_paths() -> Result<Vec<String>> {
+    let text = Text::new("Add documents:")
         .with_validator(required!("This field is required"))
         .with_help_message("e.g. file;dir/;dir/**/*.md;url;sites/**")
         .prompt()?;
-- 
cgit v1.2.3