almost finish

This commit is contained in:
Burgess Leo
2025-05-23 13:54:31 +08:00
parent 1fb457b67d
commit d4a411ce68
6 changed files with 559 additions and 242 deletions

199
test/folders_sort.py Normal file
View File

@@ -0,0 +1,199 @@
import os
import sqlite3
from files_sort import GetFilesDBNodeInfo, sort_fragments_by_start_byte
def GetFolderID(
folder_path: str,
db_path: str = "../src/db_ntfs_info.db",
table_name: str = "db_path"
) -> int | None:
"""
根据文件夹路径,查询数据库中该文件夹对应的 ID。
:param folder_path: 文件夹路径(如 r"CloudMusic\\"
:param db_path: 数据库文件路径
:param table_name: 要查询的数据表名称,默认为 'db_path'
:return: 成功则返回 IDint失败返回 None
"""
conn = sqlite3.connect(db_path)
cursor = conn.cursor()
try:
# 使用 table_name 构建 SQL 查询
sql = f"SELECT ID FROM {table_name} WHERE Path = ?"
cursor.execute(sql, (folder_path,))
result = cursor.fetchone()
if result:
return result[0]
else:
print(f"未找到路径:{folder_path} 在表 {table_name}")
return None
except sqlite3.Error as e:
print(f"数据库操作失败:{e}")
return None
finally:
conn.close()
def GetSubPathsByParentID(
parent_id: int,
db_path: str = "../src/db_ntfs_info.db",
table_name: str = "db_path"
) -> list:
"""
根据 ParentID 查询 db_path 表中对应的子项(文件/文件夹)。
:param parent_id: 父节点 ID
:param db_path: 数据库文件路径
:param table_name: 数据表名称
:return: 包含 ID、Path、Name 的字典列表
"""
conn = sqlite3.connect(db_path)
cursor = conn.cursor()
sql = f"""
SELECT ID, Path, Name
FROM {table_name}
WHERE ParentID = ?
"""
try:
cursor.execute(sql, (parent_id,))
rows = cursor.fetchall()
except Exception as e:
print(f"数据库查询失败:{e}")
return []
results = []
for row in rows:
item = {
'id': row[0],
'absolute_path': row[1],
'name': row[2]
}
results.append(item)
conn.close()
return results
if __name__ == "__main__":
test_folder_path = "pictures/"
parent_id_test = GetFolderID(test_folder_path)
# node_data = GetNodeFragmentsByParentID(parent_id_test)
path_data = GetSubPathsByParentID(parent_id_test)
node_data = GetFilesDBNodeInfo(path_records=path_data)
for data in node_data:
print(data)
def GetSortFragmentsByFolderPath(db_path: str = "../src/db_ntfs_info.db", folder_path: str = None) -> list:
"""
根据文件夹路径,查询数据库中该文件夹下的所有文件的分片信息。
:param db_path: 要查询的数据库
:param folder_path: 文件夹的绝对路径
:return list: 文件夹下所有文件按片段顺序排列的列表
"""
parent_id = GetFolderID(folder_path=folder_path, db_path=db_path)
path_data = GetSubPathsByParentID(parent_id=parent_id, db_path=db_path)
node_data = GetFilesDBNodeInfo(path_records=path_data)
result = sort_fragments_by_start_byte(node_data)
return result
# if __name__ == "__main__":
# folder_path_test = "pictures/"
# data = GetSortFragmentsByFolderPath(db_path="../src/db_ntfs_info.db", folder_path=folder_path_test)
# for item in data:
# print(item)
def ScanDirectory(root_dir, skip_system=True):
"""
递归扫描指定目录,返回相对于盘符的路径列表(使用 '/' 分隔),不包含盘符。
:param root_dir: 要扫描的根目录路径
:param skip_system: 是否跳过系统目录(默认 True
:return: 文件路径列表,格式为 relative/path/to/file.ext
"""
file_list = []
for root, dirs, files in os.walk(root_dir):
# 跳过系统目录
if skip_system:
dirs[:] = [d for d in dirs if not d.startswith('$') and d != "System Volume Information"]
for file in files:
full_path = os.path.join(root, file)
# 去掉盘符
_, relative_path = os.path.splitdrive(full_path)
# 替换 \ 为 /
relative_path = relative_path.lstrip("\\").replace("\\", "/")
file_list.append(relative_path)
return file_list
# if __name__ == "__main__":
# folder_path = r"Y:/folder1/"
# files_list = ScanDirectory(folder_path)
#
# print(f"共找到 {len(files_list)} 个文件:")
# for f in files_list:
# print(f)
def ScanMultiFolders(folder_paths, skip_system=True):
"""
扫描多个根目录,返回所有子目录中的文件路径列表。
:param folder_paths: 包含多个根目录的列表
:param skip_system: 是否跳过系统目录(默认 True
:return: 所有文件的相对路径列表(格式为 folder/file.ext
"""
all_files = []
for root_dir in folder_paths:
# 确保路径存在
if not os.path.exists(root_dir):
print(f"⚠️ 路径不存在:{root_dir}")
continue
for root, dirs, files in os.walk(root_dir):
# 跳过系统目录
if skip_system:
dirs[:] = [d for d in dirs if not d.startswith('$') and d != "System Volume Information"]
for file in files:
full_path = os.path.join(root, file)
# 去掉盘符
_, relative_path = os.path.splitdrive(full_path)
relative_path = relative_path.lstrip("\\").replace("\\", "/")
all_files.append(relative_path)
return all_files
if __name__ == "__main__":
folders = [
r"Y:\CloudMusic",
r"Y:\folder1"
]
files = ScanMultiFolders(folders)
print(f"共找到 {len(files)} 个文件:")
for f in files:
print(f)