# Content Cache
LRU cache for file contents with mtime validation.
## Purpose
The Content Cache stores file contents in memory using a Least Recently Used (LRU) eviction policy. This reduces disk I/O for frequently accessed files and enables fast search operations.
## How It Works
```mermaid
flowchart TD
A[Read request] --> B{In memory cache?}
B -->|Yes| C{mtime matches?}
B -->|No| D{In persistent cache?}
C -->|Yes| E[Return cached content]
C -->|No| F[Invalidate entry]
D -->|Yes| G{mtime matches?}
D -->|No| H[Read from disk]
G -->|Yes| I[Promote to memory]
G -->|No| F
F --> H
H --> J[Parse metadata]
J --> K[Store in cache]
K --> E
I --> E
```
## Configuration
| Variable | Type | Default | Description |
| ----------------------------- | -------- | ------- | ------------------------------- |
| `OBSIDIAN_CONTENT_CACHE_SIZE` | `number` | `100` | Maximum entries in memory cache |
## Cached Data
Each cache entry contains:
| Field | Type | Description |
| ------------------- | ---------- | --------------------------- |
| `content` | `string` | Raw file content |
| `mtime` | `number` | File modification timestamp |
| `parsedFrontmatter` | `object` | Parsed YAML frontmatter |
| `wordCount` | `number` | Total words in file |
| `lineCount` | `number` | Total lines in file |
| `inlineTags` | `string[]` | Extracted inline `#tags` |
## LRU Eviction
When the cache reaches `maxSize`, the least recently used entry is evicted:
```mermaid
flowchart LR
subgraph cache [LRU Cache]
A[Most Recent] --> B --> C --> D[Least Recent]
end
E[New Entry] --> A
D --> F[Evicted]
```
## Mtime Validation
The cache validates entries against the file's modification time:
1. Get file stats from filesystem
2. Compare `mtime` with cached `mtime`
3. If different, invalidate and re-read
4. If same, return cached content
This ensures cache consistency without full file comparison.
## Persistence
When persistent caching is enabled:
- Cache entries are written to LMDB
- On startup, entries are loaded from LMDB to memory
- Memory cache acts as hot layer, LMDB as warm layer
See [Persistent Cache](persistent-cache.md) for details.
## Statistics
Available via [get_cache_stats](../tools/get_cache_stats.md):
| Stat | Description |
| ---------------- | ---------------------------- |
| `size` | Current entries in memory |
| `maxSize` | Maximum cache capacity |
| `files` | List of cached file paths |
| `persistentSize` | Entries in LMDB (if enabled) |
## Performance Impact
| Operation | Without Cache | With Cache |
| ---------------- | ------------- | ---------- |
| Read 10KB file | ~5-10ms | <0.1ms |
| Search 100 files | ~500-1000ms | ~10-50ms |
| Repeated read | ~5-10ms | <0.1ms |
## Source Files
- [`src/services/cache/ContentCache.ts`](../../src/services/cache/ContentCache.ts)
## Related Features
- [Persistent Cache](persistent-cache.md) - LMDB backing
- [File Watcher](file-watcher.md) - Automatic invalidation
- [Inverted Index](inverted-index.md) - Uses cached content for indexing
