NoteMD Pro - Add Wiki-Links

Overview

This feature processes markdown files to add wiki-links ([[concept]]) between related concepts using an LLM. It analyzes the content and identifies core knowledge points, then adds appropriate backlinks.

When to Use

• Organize notes: Connect related concepts in your vault
• Build knowledge graph: Create interconnected notes for better navigation
• Enhance linking: Automatically add links that might be missed manually

Function Call Chain

processFile (fileUtils.ts)
├── read_file(inputPath)
├── splitContent(content, settings)
├── getProviderForTask('addLinks', settings)
├── call[Provider]API(prompt, chunk)
├── createConceptNotes(settings, concepts, filename)
│   ├── normalizeNameForFilePath(concept)
│   └── write_file(notePath, content)
├── handleDuplicates(content, settings)
├── cleanupLatexDelimiters(content)
├── refineMermaidBlocks(content)  ← AUTO-FIX
└── saveOrMoveProcessedFile(settings, inputPath, content)
    └── write_file(outputPath, content)

Key Functions (from fileUtils.ts)

processFile

Processes a single file: reads content, calls LLM, adds links, applies fixes, and saves.

export async function processFile(
  settings: NotemdSettings,
  inputPath: string,
  progressReporter: ProgressReporter,
  currentProcessingFileBasename: { value: string | null },
);

splitContent

Splits content into chunks based on word count for LLM processing.

export function splitContent(
  content: string,
  settings: NotemdSettings,
): string[];
// Returns: Array of content chunks

saveOrMoveProcessedFile

Handles saving or moving the processed file based on settings.

async function saveOrMoveProcessedFile(
  settings: NotemdSettings,
  inputPath: string,
  processedContent: string,
  progressReporter: ProgressReporter,
);

Settings

Core Settings

Output Settings

Mermaid/LaTeX Settings

Duplicate Detection

Prompt Template (addLinks)

From promptUtils.ts:

Completely decompose and structure the knowledge points in this markdown document,
outputting them in markdown format supported by Obsidian. Core knowledge points
should be labelled with Obsidian's backlink format [[]]. Do not output anything
other than the original text and the requested "Obsidian's backlink format [[]]".

Rules:
1. Only add Obsidian backlinks [[like this]] to core concepts. Do not modify the
   original text content or formatting otherwise.
2. Skip conventional names (common products, company names, dates, times, individual
   names) unless they represent a core technical or scientific concept within the
   text's context.
3. Output the *entire* original content of the chunk, preserving all formatting
   (headers, lists, code blocks, etc.), with only the added backlinks.
4. Handle duplicate concepts carefully:
   a. If both singular and plural forms of a word/concept appear (e.g., "model" and
      "models"), only add the backlink to the *first occurrence* of the *singular* form.
   b. If a single-word concept also appears as part of a multi-word concept, only
      add the backlink to the multi-word concept.
   c. Do not add duplicate backlinks for the exact same concept within this chunk.
5. Ignore any "References", "Bibliography", or similar sections.

Error Handling

Common Errors

1. No valid LLM provider
   • Error: "No valid LLM provider configured for 'Add Links' task"
   • Solution: Configure a provider in settings
2. API rate limiting
   • Error: Rate limit exceeded
   • Solution: Adjust apiCallInterval in settings
3. File permission errors
   • Error: Cannot modify/create file
   • Solution: Check folder permissions

Cancellation

• User can cancel processing via progress reporter
• Already processed chunks are preserved
• Progress saved in log

Auto-fix Integration

After LLM processing, the following fixes are automatically applied:

1. cleanupLatexDelimiters: Fix LaTeX math delimiters
2. refineMermaidBlocks: Fix Mermaid diagram syntax
3. Boxed content removal: Remove \boxed{ wrappers

See mermaid-healer for full Mermaid fixing code.