feat: implement hunk splitting strategy in diff application to better handle large diffs that might fail if the search fails

src/core/diff/strategies/new-unified/index.ts

@@ -1,4 +1,4 @@
-import { Diff, Hunk } from "./types"
+import { Diff, Hunk, Change } from "./types"
 import { findBestMatch, prepareSearchString } from "./search-strategies"
 import { applyEdit } from "./edit-strategies"
 import { DiffResult, DiffStrategy } from "../../types"
@@ -185,6 +185,59 @@ Your diff here
 </apply_diff>`
 	}
 
+	// Helper function to split a hunk into smaller hunks based on contiguous changes
+	private splitHunk(hunk: Hunk): Hunk[] {
+		const result: Hunk[] = []
+		let currentHunk: Hunk | null = null
+		let contextBefore: Change[] = []
+		let contextAfter: Change[] = []
+		const MAX_CONTEXT_LINES = 3 // Keep 3 lines of context before/after changes
+
+		for (let i = 0; i < hunk.changes.length; i++) {
+			const change = hunk.changes[i]
+
+			if (change.type === 'context') {
+				if (!currentHunk) {
+					contextBefore.push(change)
+					if (contextBefore.length > MAX_CONTEXT_LINES) {
+						contextBefore.shift()
+					}
+				} else {
+					contextAfter.push(change)
+					if (contextAfter.length > MAX_CONTEXT_LINES) {
+						// We've collected enough context after changes, create a new hunk
+						currentHunk.changes.push(...contextAfter)
+						result.push(currentHunk)
+						currentHunk = null
+						// Keep the last few context lines for the next hunk
+						contextBefore = contextAfter
+						contextAfter = []
+					}
+				}
+			} else {
+				if (!currentHunk) {
+					currentHunk = { changes: [...contextBefore] }
+					contextAfter = []
+				} else if (contextAfter.length > 0) {
+					// Add accumulated context to current hunk
+					currentHunk.changes.push(...contextAfter)
+					contextAfter = []
+				}
+				currentHunk.changes.push(change)
+			}
+		}
+
+		// Add any remaining changes
+		if (currentHunk) {
+			if (contextAfter.length > 0) {
+				currentHunk.changes.push(...contextAfter)
+			}
+			result.push(currentHunk)
+		}
+
+		return result
+	}
+
 	async applyDiff(
 		originalContent: string,
 		diffContent: string,
@@ -206,55 +259,82 @@ Your diff here
 			const contextStr = prepareSearchString(hunk.changes)
 			const { index: matchPosition, confidence, strategy } = findBestMatch(contextStr, result, 0, this.confidenceThreshold)
 
+			if (confidence < 1.1) {
+        console.log('Full hunk application failed, trying sub-hunks strategy')
+				// Try splitting the hunk into smaller hunks
+				const subHunks = this.splitHunk(hunk)
+				let subHunkSuccess = true
+				let subHunkResult = [...result]
+
+				for (const subHunk of subHunks) {
+					const subContextStr = prepareSearchString(subHunk.changes)
+          console.log(subContextStr)
+					const subSearchResult = findBestMatch(subContextStr, subHunkResult, 0, this.confidenceThreshold)
+
+					if (subSearchResult.confidence >= this.confidenceThreshold) {
+						const subEditResult = await applyEdit(subHunk, subHunkResult, subSearchResult.index, subSearchResult.confidence, this.confidenceThreshold)
+						if (subEditResult.confidence >= this.confidenceThreshold) {
+							subHunkResult = subEditResult.result
+							continue
+						}
+					}
+					subHunkSuccess = false
+					break
+				}
+
+				if (subHunkSuccess) {
+					result = subHunkResult
+					continue
+				}
+
+				// If sub-hunks also failed, return the original error
+				const contextLines = hunk.changes.filter(c => c.type === "context").length
+				const totalLines = hunk.changes.length
+				const contextRatio = contextLines / totalLines
+
+				let errorMsg = `Failed to find a matching location in the file (${Math.floor(confidence * 100)}% confidence, needs ${Math.floor(this.confidenceThreshold * 100)}%)\n\n`
+				errorMsg += "Debug Info:\n"
+				errorMsg += `- Search Strategy Used: ${strategy}\n`
+				errorMsg += `- Context Lines: ${contextLines} out of ${totalLines} total lines (${Math.floor(contextRatio * 100)}%)\n`
+				errorMsg += `- Attempted to split into ${subHunks.length} sub-hunks but still failed\n`
+				
+				if (contextRatio < 0.2) {
+					errorMsg += "\nPossible Issues:\n"
+					errorMsg += "- Not enough context lines to uniquely identify the location\n"
+					errorMsg += "- Add a few more lines of unchanged code around your changes\n"
+				} else if (contextRatio > 0.5) {
+					errorMsg += "\nPossible Issues:\n"
+					errorMsg += "- Too many context lines may reduce search accuracy\n"
+					errorMsg += "- Try to keep only 2-3 lines of context before and after changes\n"
+				} else {
+					errorMsg += "\nPossible Issues:\n"
+					errorMsg += "- The diff may be targeting a different version of the file\n"
+					errorMsg += "- There may be too many changes in a single hunk, try splitting the changes into multiple hunks\n"
+				}
+
+				if (startLine && endLine) {
+					errorMsg += `\nSearch Range: lines ${startLine}-${endLine}\n`
+				}
+
+				return { success: false, error: errorMsg }
+			}
+
 			const editResult = await applyEdit(hunk, result, matchPosition, confidence, this.confidenceThreshold)
 			if (editResult.confidence >= this.confidenceThreshold) {
 				result = editResult.result
 			} else {
-				// Determine if the failure is due to search or edit
+				// Edit failure - likely due to content mismatch
-				if (confidence < this.confidenceThreshold) {
+				let errorMsg = `Failed to apply the edit using ${editResult.strategy} strategy (${Math.floor(editResult.confidence * 100)}% confidence)\n\n`
-					// Search failure - likely due to context not matching
+				errorMsg += "Debug Info:\n"
-					const contextLines = hunk.changes.filter(c => c.type === "context").length
+				errorMsg += "- The location was found but the content didn't match exactly\n"
-					const totalLines = hunk.changes.length
+				errorMsg += "- This usually means the file has been modified since the diff was created\n"
-					const contextRatio = contextLines / totalLines
+				errorMsg += "- Or the diff may be targeting a different version of the file\n"
+				errorMsg += "\nPossible Solutions:\n"
+				errorMsg += "1. Refresh your view of the file and create a new diff\n"
+				errorMsg += "2. Double-check that the removed lines (-) match the current file content\n"
+				errorMsg += "3. Ensure your diff targets the correct version of the file"
 
-					let errorMsg = `Failed to find a matching location in the file (${Math.floor(confidence * 100)}% confidence, needs ${Math.floor(this.confidenceThreshold * 100)}%)\n\n`
+				return { success: false, error: errorMsg }
-					errorMsg += "Debug Info:\n"
-					errorMsg += `- Search Strategy Used: ${strategy}\n`
-					errorMsg += `- Context Lines: ${contextLines} out of ${totalLines} total lines (${Math.floor(contextRatio * 100)}%)\n`
-					
-					if (contextRatio < 0.2) {
-						errorMsg += "\nPossible Issues:\n"
-						errorMsg += "- Not enough context lines to uniquely identify the location\n"
-						errorMsg += "- Add a few more lines of unchanged code around your changes\n"
-					} else if (contextRatio > 0.5) {
-						errorMsg += "\nPossible Issues:\n"
-						errorMsg += "- Too many context lines may reduce search accuracy\n"
-						errorMsg += "- Try to keep only 2-3 lines of context before and after changes\n"
-					} else {
-						errorMsg += "\nPossible Issues:\n"
-						errorMsg += "- The diff may be targeting a different version of the file\n"
-            errorMsg += "- There may be too many changes in a single hunk, try splitting the changes into multiple hunks\n"
-					}
-
-					if (startLine && endLine) {
-						errorMsg += `\nSearch Range: lines ${startLine}-${endLine}\n`
-					}
-
-					return { success: false, error: errorMsg }
-				} else {
-					// Edit failure - likely due to content mismatch
-					let errorMsg = `Failed to apply the edit using ${editResult.strategy} strategy (${Math.floor(editResult.confidence * 100)}% confidence)\n\n`
-					errorMsg += "Debug Info:\n"
-					errorMsg += "- The location was found but the content didn't match exactly\n"
-					errorMsg += "- This usually means the file has been modified since the diff was created\n"
-					errorMsg += "- Or the diff may be targeting a different version of the file\n"
-					errorMsg += "\nPossible Solutions:\n"
-					errorMsg += "1. Refresh your view of the file and create a new diff\n"
-					errorMsg += "2. Double-check that the removed lines (-) match the current file content\n"
-					errorMsg += "3. Ensure your diff targets the correct version of the file"
-
-					return { success: false, error: errorMsg }
-				}
 			}
 		}