From 9c8e7519f38850efcac4994543b7161a26449f75 Mon Sep 17 00:00:00 2001
From: Rin_0xTohsaka <89101179+Tachikoma000@users.noreply.github.com>
Date: Wed, 18 Sep 2024 19:11:38 -0500
Subject: [PATCH] minor comments to help review main feature files

- Comments added to pdf.rs, mod.rs, document_loaders.rs
---
 rig-core/examples/document_loaders.rs |  6 ++----
 rig-core/src/document_loaders/mod.rs  |  4 +++-
 rig-core/src/document_loaders/pdf.rs  | 13 ++++++++++++-
 3 files changed, 17 insertions(+), 6 deletions(-)

diff --git a/rig-core/examples/document_loaders.rs b/rig-core/examples/document_loaders.rs
index 4e04b217..40a41712 100644
--- a/rig-core/examples/document_loaders.rs
+++ b/rig-core/examples/document_loaders.rs
@@ -1,5 +1,3 @@
-// examples/document_loaders.rs
-
 use rig::{
     completion::Prompt,
     document_loaders::PdfLoader,
@@ -50,8 +48,8 @@ async fn main() -> Result<(), anyhow::Error> {
         embeddings.len()
     );
     for emb in &embeddings {
-        println!("Document ID: {}", emb.id);
-        println!("Document Content: {:?}", emb.document);
+        // println!("Document ID: {}", emb.id);
+        // println!("Document Content: {:?}", emb.document);
         println!("Number of embeddings: {}", emb.embeddings.len());
         println!(
             "First embedding vector length: {}",
diff --git a/rig-core/src/document_loaders/mod.rs b/rig-core/src/document_loaders/mod.rs
index ee375bc9..b25b15f5 100644
--- a/rig-core/src/document_loaders/mod.rs
+++ b/rig-core/src/document_loaders/mod.rs
@@ -1,4 +1,5 @@
-// src/document_loaders/mod.rs
+//! This module contains the implementation of document loaders for various file formats.
+//! Currently, it includes loaders for CSV and PDF files.
 
 mod csv;
 // mod directory;
@@ -14,6 +15,7 @@ use std::error::Error as StdError;
 
 #[async_trait]
 pub trait DocumentLoader {
+    /// Asynchronously loads the document and returns a vector of document embeddings.
     async fn load(&self) -> Result<Vec<DocumentEmbeddings>, Box<dyn StdError + Send + Sync>>;
 }
 
diff --git a/rig-core/src/document_loaders/pdf.rs b/rig-core/src/document_loaders/pdf.rs
index f8489ec3..6d073af1 100644
--- a/rig-core/src/document_loaders/pdf.rs
+++ b/rig-core/src/document_loaders/pdf.rs
@@ -1,14 +1,17 @@
+// Import necessary dependencies
 use super::DocumentLoader;
 use crate::embeddings::DocumentEmbeddings;
 use async_trait::async_trait;
 use lopdf::Document;
 use serde_json::json;
 
+// Define a struct for loading PDF documents
 pub struct PdfLoader {
     path: String,
 }
 
 impl PdfLoader {
+    // Implement a constructor for the PdfLoader struct
     pub fn new(path: &str) -> Self {
         Self {
             path: path.to_string(),
@@ -18,17 +21,25 @@ impl PdfLoader {
 
 #[async_trait]
 impl DocumentLoader for PdfLoader {
+    // Implement the load function for the DocumentLoader trait
     async fn load(
         &self,
     ) -> Result<Vec<DocumentEmbeddings>, Box<dyn std::error::Error + Send + Sync>> {
+        // Load the PDF document from the specified path
         let doc = Document::load(&self.path)?;
+
+        // Extract text from each page of the PDF document
         let mut text = String::new();
         for page in doc.get_pages() {
             if let Ok(content) = doc.extract_text(&[page.0]) {
                 text.push_str(&content);
             }
         }
-        println!("Extracted text from PDF: {}", text); // Debug print
+
+        // Print the extracted text for debugging purposes
+        println!("Extracted text from PDF: {}", text);
+
+        // Create a DocumentEmbeddings object with the extracted text
         Ok(vec![DocumentEmbeddings {
             id: self.path.clone(),
             document: json!({"text": text}),