WIP. A little custom music server
chore: update issues list
+20
-16
+20
-16
backend/ISSUES.md
+20
-16
backend/ISSUES.md
···
1
1
# Architectural Issues & Tasks
2
2
3
3
## High Priority (Stability & Performance)
4
-
- [ ] **Fix Critical Memory Usage in `FlacService`**
5
-
- **Location**: `src/flac/service.ts`
6
-
- **Issue**: `readMetadata` uses `fs.readFile` to load entire audio files (often 20-100MB) into memory just to read headers.
7
-
- **Task**: Implement partial reads (open file descriptor -> read buffer) to parse headers and Vorbis comments without loading the binary audio data.
4
+
- [ ] **Improve cover art logic**
5
+
- **Issue**: Right now we base our cover art loading on musicbrainzReleaseId using cover art api, but we should cache it on server. and use multiple fallbacks
6
+
- **Task**: Implement cover art fetching/fallbacking.
7
+
1. Loading from database / local cover art folder.
8
+
- If we have releasegroupid - try fetching from coverart api;
9
+
- Try parsing coverart from file.
10
+
2. If not - use link to coverart api based on relaseId
11
+
3. If not - use placeholder
8
12
9
-
- [ ] **Implement Pagination for API Endpoints**
10
-
- **Location**: `src/api.ts`
11
-
- **Issue**: `getAlbumList` and `getAllSongs` fetch the entire database table in a single query.
12
-
- **Task**: Add limit/offset or cursor-based pagination parameters to these endpoints and updated the database queries accordingly.
13
+
Cover art fetcher should be running on schedule to load this shit in background rather then block user flow.
14
+
13
15
14
-
- [ ] **Optimize Library Synchronization**
15
-
- **Location**: `src/file-parser.ts` & `src/sync-library.ts`
16
-
- **Issue**: `readDirectory` and the sync stream load all file paths into memory before processing. `alreadyIndexed` loads the entire file table.
17
-
- **Task**: Refactor to use a streaming directory iterator and batched database lookups to handle large libraries (50k+ songs) without OOM.
18
16
19
17
## Medium Priority (Reliability & Maintainability)
20
18
- [ ] **Standardize API Error Handling**
···
22
20
- **Issue**: Inconsistent error handling. Some endpoints map errors explicitly, others rely on `runtime.runPromise` which may cause unhandled promise rejections or generic 500s for domain errors.
23
21
- **Task**: Create a standardized wrapper or middleware that executes Effects and maps specific `TaggedError` types to HTTP status codes (404, 400, etc.).
24
22
25
-
- [ ] **Decouple API from Database Logic**
26
-
- **Location**: `src/api.ts`
27
-
- **Issue**: API handlers contain `db.query` and `db.select` calls directly ("Active Record" pattern in controllers).
28
-
- **Task**: Extract database queries into a dedicated Repository layer or specific Service methods to enable unit testing without a database connection.
23
+
- [ ] **Optimize Library Synchronization**
24
+
- **Location**: `src/file-parser.ts` & `src/sync-library.ts`
25
+
- **Issue**: `readDirectory` and the sync stream load all file paths into memory before processing. `alreadyIndexed` loads the entire file table.
26
+
- **Task**: Refactor to use a streaming directory iterator and batched database lookups to handle large libraries (50k+ songs) without OOM.
27
+
29
28
30
29
## Low Priority (Code Quality)
31
30
- [ ] **Refactor Effect <-> Elysia Adapter**
···
37
36
- **Location**: `src/index.ts`
38
37
- **Issue**: Server port `3003` is hardcoded.
39
38
- **Task**: Use Effect's `Config` module to make the port configurable via environment variables.
39
+
40
+
- [ ] **Implement Pagination for API Endpoints**
41
+
- **Location**: `src/api.ts`
42
+
- **Issue**: `getAlbumList` and `getAllSongs` fetch the entire database table in a single query.
43
+
- **Task**: Add limit/offset or cursor-based pagination parameters to these endpoints and updated the database queries accordingly.