Update database schema for client tracking

- Add functions to manage client records (add, find, list)
- Require --confirm-add and --name flags when adding a new client

main.py

@@ -8,15 +8,20 @@ archival locations.
 
 """
 
-import argparse
 import os
-import sqlite3
 import sys
+import re
+import uuid
+import argparse
+import sqlite3
+
+from datetime import datetime, timezone
 
 import qbittorrent
 
 # SCHEMA format is YYYYMMDDX
-SCHEMA = 202410040
+SCHEMA = 202410060
+
 
 def init_db(conn):
     """
@@ -27,36 +32,77 @@ def init_db(conn):
     c.executescript(
         f"""
         PRAGMA user_version = {SCHEMA};
+
+        CREATE TABLE IF NOT EXISTS clients (
+            id INTEGER PRIMARY KEY AUTOINCREMENT,
+            name TEXT NOT NULL UNIQUE,
+            uuid TEXT NOT NULL UNIQUE,
+            endpoint TEXT NOT NULL,
+            last_seen DATETIME NOT NULL
+        );
+
         CREATE TABLE IF NOT EXISTS torrents (
             id INTEGER PRIMARY KEY AUTOINCREMENT,
             info_hash_v1 TEXT NOT NULL UNIQUE,
             info_hash_v2 TEXT UNIQUE,
-            name TEXT NOT NULL,
             file_count INTEGER NOT NULL,
+            completed_on DATETIME NOT NULL
+        );
+
+        CREATE TABLE IF NOT EXISTS torrent_clients (
+            id INTEGER PRIMARY KEY AUTOINCREMENT,
+            torrent_id INTEGER NOT NULL,
+            client_id INTEGER NOT NULL,
+            name TEXT NOT NULL,
             content_path TEXT NOT NULL,
-            completed_on DATETIME DEFAULT CURRENT_TIMESTAMP,
+            last_seen DATETIME NOT NULL,
-            tracker_ids TEXT
+            FOREIGN KEY (torrent_id) REFERENCES torrents(id),
+            FOREIGN KEY (client_id) REFERENCES clients(id),
+            UNIQUE (torrent_id, client_id)
         );
 
         CREATE TABLE IF NOT EXISTS trackers (
             id INTEGER PRIMARY KEY AUTOINCREMENT,
-            url TEXT NOT NULL UNIQUE
+            url TEXT NOT NULL UNIQUE,
+            last_seen DATETIME NOT NULL
+        );
+
+        CREATE TABLE IF NOT EXISTS torrent_trackers (
+            id INTEGER PRIMARY KEY AUTOINCREMENT,
+            client_id INTEGER NOT NULL,
+            torrent_id INTEGER NOT NULL,
+            tracker_id INTEGER NOT NULL,
+            last_seen DATETIME NOT NULL,
+            FOREIGN KEY (client_id) REFERENCES clients(id),
+            FOREIGN KEY (torrent_id) REFERENCES torrents(id),
+            FOREIGN KEY (tracker_id) REFERENCES trackers(id),
+            UNIQUE (client_id, torrent_id, tracker_id)
         );
 
         CREATE TABLE IF NOT EXISTS files (
             id INTEGER PRIMARY KEY AUTOINCREMENT,
-            torrent_id INTEGER NOT NULL,
-            file_index INTEGER NOT NULL,
-            file_path TEXT NOT NULL,
             size INTEGER NOT NULL,
-            is_downloaded BOOLEAN NOT NULL DEFAULT 0,
+            oshash TEXT NOT NULL UNIQUE,
-            last_checked DATETIME,
+            hash TEXT UNIQUE
-            FOREIGN KEY (torrent_id) REFERENCES torrents(id),
-            UNIQUE (torrent_id, file_index)
         );
 
-    """
+        CREATE TABLE IF NOT EXISTS torrent_files (
+            id INTEGER PRIMARY KEY AUTOINCREMENT,
+            file_id INTEGER NOT NULL,
+            torrent_id INTEGER NOT NULL,
+            client_id INTEGER NOT NULL,
+            file_index INTEGER NOT NULL,
+            file_path TEXT NOT NULL,
+            is_downloaded BOOLEAN NOT NULL,
+            last_checked DATETIME NOT NULL,
+            FOREIGN KEY (file_id) REFERENCES files(id),
+            FOREIGN KEY (torrent_id) REFERENCES torrents(id),
+            FOREIGN KEY (client_id) REFERENCES clients(id),
+            UNIQUE (file_id, torrent_id, client_id, file_index)
+        );
+        """
     )
+    conn.commit()
     c.close()
 
 
@@ -71,6 +117,43 @@ def list_tables(conn):
     return [table[0] for table in table_list]
 
 
+def add_client(conn, name, endpoint, last_seen):
+    """
+    Add a new client endpoint to database
+    """
+    c = conn.cursor()
+    c.execute(
+        f"""
+        INSERT INTO clients (uuid, name, endpoint, last_seen)
+        VALUES ("{uuid.uuid4()}", "{name}", "{endpoint}", "{last_seen}");
+        """
+    )
+    conn.commit()
+    c.close()
+
+
+def find_client(conn, endpoint):
+    """
+    Find existing client
+    """
+    c = conn.cursor()
+    c.execute(f'SELECT id, name, uuid FROM clients WHERE endpoint="{endpoint}";')
+    response = c.fetchall()
+    c.close()
+    return response
+
+
+def list_clients(conn):
+    """
+    List all stored clients
+    """
+    c = conn.cursor()
+    c.execute("SELECT * FROM clients;")
+    rows = c.fetchall()
+    c.close()
+    return rows
+
+
 parser = argparse.ArgumentParser(description="Manage BitTorrent datasets", prog="tarch")
 subparsers = parser.add_subparsers(
     dest="command", required=True, help="Available commands"
@@ -78,6 +161,10 @@ subparsers = parser.add_subparsers(
 
 scan_parser = subparsers.add_parser("scan", help="Scan command")
 scan_parser.add_argument("--debug", action="store_true", help="Enable debug mode")
+scan_parser.add_argument(
+    "--confirm-add", action="store_true", help="Confirm adding a new client"
+)
+scan_parser.add_argument("-n", "--name", help="Name of client")
 scan_parser.add_argument("-d", "--directory", help="Directory to scan")
 scan_parser.add_argument("-t", "--type", help="Scan type")
 scan_parser.add_argument("-e", "--endpoint", help="Endpoint URL")
@@ -96,7 +183,7 @@ if args.command == "scan":
         sqlitedb = sqlite3.connect(STORAGE)
         tables = list_tables(sqlitedb)
     except sqlite3.DatabaseError as e:
-        print(f"[ERROR]: Database \"{STORAGE}\" Error: {str(e)}")
+        print(f'[ERROR]: Database Error "{STORAGE}" ({str(e)})')
         sys.exit(1)
     if len(tables) == 0:
         print(f"[INFO]: Initializing database at {STORAGE}")
@@ -109,19 +196,67 @@ if args.command == "scan":
         print(f"[ERROR]: SCHEMA {SCHEMA_FOUND}, expected {SCHEMA}")
         sys.exit(1)
     if not args.directory is None:
-        print("[INFO]: --directory is not implemented\n")
+        print("[INFO]: --directory is not implemented")
         sys.exit(0)
-    if not args.endpoint is None:
+    elif not args.endpoint is None:
         qb = qbittorrent.Client(args.endpoint)
-        torrents = qb.torrents()
+        if qb.qbittorrent_version is None:
-        print(f"[INFO]: There are {len(torrents)} torrents\n")
+            print(f'[ERROR]: Couldn\'t find client version at "{args.endpoint}"')
-        for torrent in torrents[:10]:
+            sys.exit(1)
-            files = qb.get_torrent_files(torrent["hash"])
+        elif not re.match(r"^v?\d+(\.\d+)*$", qb.qbittorrent_version):
+            print(f'[ERROR]: Invalid version found at "{args.endpoint}"')
             if args.debug:
-                print(f"[DEBUG]: {repr(torrent)}")
+                print(f"[DEBUG]: {qb.qbittorrent_version}")
-            print(f"[name]: {torrent['name']}")
+            sys.exit(1)
-            print(f"[infohash_v1]: {torrent['infohash_v1']}")
+        else:
-            print(f"[content_path]: {torrent['content_path']}")
+            print(
-            print(f"[magent_uri]: {torrent['magnet_uri'][0:80]}")
+                f'[INFO]: Found qbittorrent {qb.qbittorrent_version} at "{args.endpoint}"'
-            print(f"[completed_on]: {torrent['completed']}")
+            )
-            print(f"[file_count]: {len(files)}\n")
+        clients = find_client(sqlitedb, args.endpoint)
+        if args.confirm_add:
+            if len(clients) == 0:
+                if not args.name is None:
+                    now = datetime.now(timezone.utc).isoformat(
+                        sep=" ", timespec="seconds"
+                    )
+                    add_client(sqlitedb, args.name, args.endpoint, now)
+                    print(f"[INFO]: Added client {args.name} ({args.endpoint})")
+                else:
+                    print("[ERROR]: Must specify --name for a new client")
+                    sys.exit(1)
+            elif len(clients) == 1:
+                print(f"[ERROR]: {clients[0][1]} ({clients[0][2]}) already exists")
+                sys.exit(1)
+            else:
+                print(
+                    f"[ERROR]: Multiple clients with the same endpoint: {args.endpoint}"
+                )
+                sys.exit(1)
+        elif len(clients) == 0:
+            print(f'[ERROR]: Client using endpoint "{args.endpoint}" not found')
+            print("[ERROR]: Use --confirm-add to add a new endpoint")
+            sys.exit(1)
+        elif len(clients) == 1:
+            torrents = qb.torrents()
+            print(f"[INFO]: There are {len(torrents)} torrents\n")
+            for torrent in torrents[:2]:
+                files = qb.get_torrent_files(torrent["hash"])
+                trackers = qb.get_torrent_trackers(torrent["hash"])
+                print(f"[name]: {torrent['name']}")
+                print(f"[infohash_v1]: {torrent['infohash_v1']}")
+                print(f"[content_path]: {torrent['content_path']}")
+                print(f"[magent_uri]: {torrent['magnet_uri'][0:80]}")
+                print(f"[completed_on]: {torrent['completed']}")
+                print(f"[trackers]: {len(trackers)}")
+                print(f"[file_count]: {len(files)}\n")
+                if args.debug:
+                    print(f"[DEBUG]: {repr(torrent)}")
+                    for elem in trackers:
+                        print(f"[DEBUG]: Tracker {repr(elem)}")
+                    print("\n", end="")
+        else:
+            print(f'[ERROR]: Multiple clients ({len(clients)}) using "{args.endpoint}"')
+            sys.exit(1)
+    else:
+        print("[ERROR]: Must specify directory OR client endpoint")
+        sys.exit(1)