refactor(thread): update start_line to be optional to represent entir… (

#3614)

* refactor(thread): update start_line to be optional to represent entire file

* fix unit test

* [autofix.ci] apply automated fixes

---------

Co-authored-by: autofix-ci[bot] <114827586+autofix-ci[bot]@users.noreply.github.com>

crates/tabby-common/src/api/code.rs

@@ -36,7 +36,9 @@ pub struct CodeSearchDocument {
     pub commit: Option<String>,
 
     pub language: String,
-    pub start_line: usize,
+
+    /// When start line is `None`, it represents the entire file.
+    pub start_line: Option<usize>,
 }
 
 #[derive(Error, Debug)]

crates/tabby/src/services/code.rs

@@ -205,11 +205,11 @@ async fn create_hit(
             code::fields::CHUNK_LANGUAGE,
         )
         .to_owned(),
-        start_line: get_json_number_field(
+        start_line: Some(get_json_number_field(
             &doc,
             schema.field_chunk_attributes,
             code::fields::CHUNK_START_LINE,
-        ) as usize,
+        ) as usize),
     };
     CodeSearchHit { scores, doc }
 }
@@ -325,7 +325,7 @@ mod tests {
                 git_url: "".to_owned(),
                 commit: Some("".to_owned()),
                 language: "".to_owned(),
-                start_line: 0,
+                start_line: Some(0),
             },
         };
 

ee/tabby-db/src/threads.rs

@@ -81,7 +81,9 @@ pub struct ThreadMessageAttachmentCode {
     pub language: String,
     pub filepath: String,
     pub content: String,
-    pub start_line: usize,
+
+    /// When start line is `None`, it represents the entire file.
+    pub start_line: Option<usize>,
 }
 
 #[derive(Serialize, Deserialize)]

ee/tabby-schema/graphql/schema.graphql

@@ -169,7 +169,7 @@ input EmailSettingInput {
 
 input MessageAttachmentCodeInput {
   filepath: String
-  startLine: Int
+  "When start line is `None`, it represents the entire file." startLine: Int
   content: String!
 }
 
@@ -502,7 +502,8 @@ type MessageAttachmentCode {
   filepath: String!
   language: String!
   content: String!
-  startLine: Int!
+  "When start line is `None`, it represents the entire file."
+  startLine: Int
 }
 
 type MessageAttachmentCodeScores {

ee/tabby-schema/src/dao.rs

@@ -206,7 +206,7 @@ impl From<ThreadMessageAttachmentCode> for thread::MessageAttachmentCode {
             filepath: value.filepath,
             language: value.language,
             content: value.content,
-            start_line: value.start_line as i32,
+            start_line: value.start_line.map(|x| x as i32),
         }
     }
 }
@@ -219,7 +219,7 @@ impl From<&thread::MessageAttachmentCode> for ThreadMessageAttachmentCode {
             filepath: val.filepath.clone(),
             language: val.language.clone(),
             content: val.content.clone(),
-            start_line: val.start_line as usize,
+            start_line: val.start_line.map(|x| x as usize),
         }
     }
 }

ee/tabby-schema/src/schema/thread/inputs.rs

@@ -143,7 +143,10 @@ pub struct MessageAttachmentInput {
 #[derive(GraphQLInputObject, Clone)]
 pub struct MessageAttachmentCodeInput {
     pub filepath: Option<String>,
+
+    /// When start line is `None`, it represents the entire file.
     pub start_line: Option<i32>,
+
     pub content: String,
 }
 

ee/tabby-schema/src/schema/thread/types.rs

@@ -77,7 +77,9 @@ pub struct MessageAttachmentCode {
     pub filepath: String,
     pub language: String,
     pub content: String,
-    pub start_line: i32,
+
+    /// When start line is `None`, it represents the entire file.
+    pub start_line: Option<i32>,
 }
 
 impl From<CodeSearchDocument> for MessageAttachmentCode {
@@ -88,7 +90,7 @@ impl From<CodeSearchDocument> for MessageAttachmentCode {
             filepath: doc.filepath,
             language: doc.language,
             content: doc.body,
-            start_line: doc.start_line as i32,
+            start_line: doc.start_line.map(|x| x as i32),
         }
     }
 }

ee/tabby-webserver/src/service/answer.rs

@@ -596,7 +596,9 @@ pub async fn merge_code_snippets(
                 insert_hit.scores.embedding /= num_files;
                 insert_hit.scores.rrf /= num_files;
                 insert_hit.doc.body = file_content;
-                insert_hit.doc.start_line = 1;
+
+                // When we use entire file content, mark start_line as None.
+                insert_hit.doc.start_line = None;
                 result.push(insert_hit);
             }
         } else {
@@ -772,7 +774,7 @@ mod tests {
                 filepath: "server.py".to_owned(),
                 language: "python".to_owned(),
                 content: "from flask import Flask\n\napp = Flask(__name__)\n\n@app.route('/')\ndef hello():\n    return 'Hello, World!'".to_owned(),
-                start_line: 1,
+                start_line: Some(1),
             }],
             client_code: vec![],
         };
@@ -806,7 +808,7 @@ mod tests {
                 filepath: "server.py".to_owned(),
                 language: "python".to_owned(),
                 content: "print('Hello, server!')".to_owned(),
-                start_line: 1,
+                start_line: Some(1),
             }],
             client_code: vec![tabby_schema::thread::MessageAttachmentClientCode {
                 filepath: Some("client.py".to_owned()),
@@ -983,7 +985,7 @@ mod tests {
                 filepath: "server.py".to_owned(),
                 language: "python".to_owned(),
                 content: "print('Hello, server!')".to_owned(),
-                start_line: 1,
+                start_line: Some(1),
             }],
             client_code: vec![tabby_schema::thread::MessageAttachmentClientCode {
                 filepath: Some("client.py".to_owned()),
@@ -1045,7 +1047,7 @@ mod tests {
                 filepath: "server.py".to_owned(),
                 language: "python".to_owned(),
                 content: "print('Hello, server!')".to_owned(),
-                start_line: 1,
+                start_line: Some(1),
             }],
             client_code: vec![tabby_schema::thread::MessageAttachmentClientCode {
                 filepath: Some("client.py".to_owned()),
@@ -1341,7 +1343,7 @@ mod tests {
                     git_url: "https://github.com/test/repo.git".to_string(),
                     commit: Some("commit".to_string()),
                     language: "rust".to_string(),
-                    start_line: 1,
+                    start_line: Some(1),
                 },
                 scores: CodeSearchScores {
                     bm25: 0.5,
@@ -1358,7 +1360,7 @@ mod tests {
                     git_url: "https://github.com/test/repo.git".to_string(),
                     commit: Some("commit".to_string()),
                     language: "rust".to_string(),
-                    start_line: 3,
+                    start_line: Some(3),
                 },
                 scores: CodeSearchScores {
                     bm25: 0.6,

ee/tabby-webserver/src/service/answer/testutils/mod.rs

@@ -143,7 +143,7 @@ impl CodeSearch for FakeCodeSearch {
                     doc: CodeSearchDocument {
                         filepath: "src/lib.rs".to_string(),
                         body: "fn add(a: i32, b: i32) -> i32 {\n    a + b\n}".to_string(),
-                        start_line: 1,
+                        start_line: Some(1),
                         language: "rust".to_string(),
                         file_id: "1".to_string(),
                         chunk_id: "chunk1".to_string(),
@@ -160,7 +160,7 @@ impl CodeSearch for FakeCodeSearch {
                     doc: CodeSearchDocument {
                         filepath: "src/main.rs".to_string(),
                         body: "fn main() {\n    println!(\"Hello World\");\n}".to_string(),
-                        start_line: 1,
+                        start_line: Some(1),
                         language: "rust".to_string(),
                         file_id: "2".to_string(),
                         chunk_id: "chunk2".to_string(),