From a6c43fe4684815cd3c26be5a9abd26087696d43b Mon Sep 17 00:00:00 2001
From: iou1name <iou1name@national.shitposting.agency>
Date: Thu, 19 Dec 2024 15:00:38 -0500
Subject: [PATCH] add build library functions

---
 README.md   |   2 +-
 build_db.py | 108 ++++++++++++++++++++++++++++++++++++++++++++++++++++
 pyrite.sql  |  13 +++++++
 3 files changed, 122 insertions(+), 1 deletion(-)
 create mode 100644 build_db.py
 create mode 100644 pyrite.sql

diff --git a/README.md b/README.md
index 95281f1..1a53de6 100644
--- a/README.md
+++ b/README.md
@@ -3,7 +3,7 @@ A naturally occurring iron disulfide mineral. The name comes from the Greek word
 
 ## Requirements
 Python 3.12+
-Python packages: `fastapi uvicorn[standard] httpx jinja2 asyncpg mutagen`
+Python packages: `fastapi uvicorn[standard] httpx jinja2 asyncpg tinytag`
 
 ## Install
 ```
diff --git a/build_db.py b/build_db.py
new file mode 100644
index 0000000..eafce89
--- /dev/null
+++ b/build_db.py
@@ -0,0 +1,108 @@
+#!/usr/bin/env python3
+"""
+Builds the music library database.
+"""
+import os
+import asyncio
+import multiprocessing
+
+import asyncpg
+import tinytag
+
+import config
+
+
+MUSIC_EXT = ['.flac', '.mp3', '.opus']
+
+def read_track(filepath):
+	"""
+	Reads the specified file and extracts relevant information from it.
+	"""
+	t = tinytag.TinyTag.get(filepath)
+
+	d = {
+		'filepath': filepath,
+		'artist': t.artist,
+		'albumartist': t.albumartist,
+		'album': t.album,
+		'title': t.title,
+		'date': t.year,
+		'discnumber': str(t.disc),
+		'tracknumber': str(t.track),
+		'genre': t.genre,
+		'duration': t.duration,
+		'last_modified': os.path.getmtime(filepath)
+	}
+
+	return d
+
+
+async def build_library(root_dir):
+	"""Walks the directory and builds a library from tracks discovered."""
+	print("Building library")
+
+	db_pool = await asyncpg.create_pool(**config.db)
+	async with db_pool.acquire() as conn:
+		with open('pyrite.sql', 'r') as file:
+			await conn.execute(file.read())
+
+	filepaths = []
+	for dir_name, sub_dirs, files in os.walk(root_dir):
+		for file in files:
+			if not os.path.splitext(file)[1] in MUSIC_EXT:
+				continue
+			filepath = os.path.join(root_dir, dir_name, file)
+			last_modified = os.path.getmtime(filepath)
+			filepaths.append((filepath, last_modified))
+
+	async with db_pool.acquire() as conn:
+		tracks_prev = await conn.fetch("SELECT filepath, last_modified FROM track")
+	tracks_prev = {track['filepath']: track for track in tracks_prev}
+
+	global worker
+	def worker(args):
+		"""Worker for multi-processing tracks."""
+		filepath, last_modified = args
+		track_prev = tracks_prev.get(filepath)
+		if track_prev:
+			if track_prev['last_modified'] >= last_modified:
+				return
+		data = read_track(filepath)
+		return data
+
+	with multiprocessing.Pool() as pool:
+		mapping = pool.imap(worker, filepaths)
+		tracks = []
+		prev_percent = 0
+		while True:
+			try:
+				track = mapping.next()
+				if track:
+					tracks.append(track)
+			except StopIteration:
+				break
+			percent = round(len(tracks) / len(filepaths) * 100, 2)
+			if percent >= prev_percent + 2.5:
+				print(f"{percent}%")
+				prev_percent = percent
+	if not tracks:
+		print("No new tracks found!")
+		return
+
+	cols = ', '.join(tracks[0].keys())
+	vals = ', '.join(['$'+str(i) for i in range(1, len(tracks[0])+1)])
+	tracks_data = [list(track.values()) for track in tracks]
+
+	async with db_pool.acquire() as conn:
+		p = f"INSERT INTO track ({cols}) VALUES ({vals}) "
+		p += "ON CONFLICT(filepath) DO UPDATE SET "
+		for col in tracks[0].keys():
+			p += col + " = EXCLUDED." + col + ", "
+		p = p[:-2]
+
+		cur = await conn.prepare(p)
+		await cur.executemany(tracks_data)
+	print("Done")
+
+if __name__ == "__main__":
+	asyncio.run(build_library(config.music_dir))
diff --git a/pyrite.sql b/pyrite.sql
new file mode 100644
index 0000000..f4e7607
--- /dev/null
+++ b/pyrite.sql
@@ -0,0 +1,13 @@
+CREATE TABLE IF NOT EXISTS track (
+	filepath TEXT PRIMARY KEY,
+	artist TEXT,
+	albumartist TEXT,
+	album TEXT,
+	title TEXT,
+	date TEXT,
+	discnumber TEXT,
+	tracknumber TEXT,
+	genre TEXT,
+	duration FLOAT,
+	last_modified FLOAT
+)