Detect notebooks with the workspace's first-line rule

.nextchanges/bundles/notebook-detect-first-line.md

@@ -0,0 +1 @@
+* Detect notebooks with the same rule as the workspace: a `.py`, `.sql`, `.scala` or `.r` file is a notebook when its first line contains `Databricks notebook source`, for example after a byte order mark. Previously such a file was tracked as a plain file, so removing it locally left the remote notebook behind. ([#7290](https://github.com/databricks/cli/pull/7290))

acceptance/bundle/sync/replace-node-type/output.txt

@@ -24,12 +24,9 @@ FILE plain.sql
 === Remove both local files and sync again
 
 >>> [CLI] bundle sync --output text
-Action: DELETE: bom.sql, plain.sql
-Deleted bom.sql
+Action: DELETE: bom, plain.sql
+Deleted bom
 Deleted plain.sql
 Initial Sync Complete
 
-=== BUG: bom remains after sync reports deleting bom.sql
-
 >>> list
-NOTEBOOK bom

acceptance/bundle/sync/replace-node-type/script

@@ -27,6 +27,4 @@ trace list
 title "Remove both local files and sync again\n"
 rm bom.sql plain.sql
 trace $CLI bundle sync --output text 2>&1 | sort
-
-title "BUG: bom remains after sync reports deleting bom.sql\n"
 trace list

acceptance/bundle/sync/replace-node-type/test.toml

@@ -1,5 +1,3 @@
-Badness = "After a BOM-prefixed SQL notebook is removed locally, sync reports deleting bom.sql but leaves the remote notebook bom. The CLI does not detect the file as a notebook, so the snapshot keeps tracking it as bom.sql"
-
 Cloud = true
 Env.MSYS_NO_PATHCONV = "1"
 Ignore = [".databricks"]

libs/notebook/detect.go

@@ -22,8 +22,11 @@ type FileInfoWithWorkspaceObjectInfo interface {
 	WorkspaceObjectInfo() workspace.ObjectInfo
 }
 
-// Maximum length in bytes of the notebook header.
-const headerLength = 32
+// Marker the workspace looks for to import a source file as a notebook.
+const notebookSourceMarker = "Databricks notebook source"
+
+// The workspace reads the first line out of this many leading bytes.
+const headerLength = 200
 
 // file wraps an fs.File and implements a few helper methods such that
 // they don't need to be inlined in the [DetectWithFS] function below.
@@ -127,28 +130,20 @@ func DetectBytes(name string, content []byte) (notebook bool, language workspace
 
 // detectSource reports whether a source file with the given name and first line is a notebook.
 func detectSource(name, fileHeader string) (notebook bool, language workspace.Language) {
-	header := ""
-
-	// Determine which header to expect based on filename extension.
-	ext := strings.ToLower(filepath.Ext(name))
-	switch ext {
+	switch strings.ToLower(filepath.Ext(name)) {
 	case ExtensionPython:
-		header = `# Databricks notebook source`
 		language = workspace.LanguagePython
 	case ExtensionR:
-		header = `# Databricks notebook source`
 		language = workspace.LanguageR
 	case ExtensionScala:
-		header = "// Databricks notebook source"
 		language = workspace.LanguageScala
 	case ExtensionSql:
-		header = "-- Databricks notebook source"
 		language = workspace.LanguageSql
 	default:
 		return false, ""
 	}
 
-	if fileHeader != header {
+	if !strings.Contains(fileHeader, notebookSourceMarker) {
 		return false, ""
 	}
 

libs/notebook/detect_test.go

@@ -1,6 +1,7 @@
 package notebook
 
 import (
+	"bytes"
 	"io/fs"
 	"os"
 	"path/filepath"
@@ -140,6 +141,10 @@ func TestDetectBytes(t *testing.T) {
 	}{
 		{"nb.py", []byte("# Databricks notebook source\nprint(1)\n"), true, workspace.LanguagePython},
 		{"nb.sql", []byte("-- Databricks notebook source\nSELECT 1\n"), true, workspace.LanguageSql},
+		{"marker_mid_line.sql", []byte("-- MAGIC Databricks notebook source\nSELECT 1\n"), true, workspace.LanguageSql},
+		{"other_comment.py", []byte("// Databricks notebook source\nprint(1)\n"), true, workspace.LanguagePython},
+		{"second_line.py", []byte("print(1)\n# Databricks notebook source\n"), false, ""},
+		{"past_buffer.py", append(bytes.Repeat([]byte("#"), headerLength), []byte(" Databricks notebook source\n")...), false, ""},
 		{"nb.ipynb", ipynb, true, workspace.LanguagePython},
 		{"plain.py", []byte("print(1)\n"), false, ""},
 		{"space.geniespace.json", []byte("{}"), false, ""},