Feat: start testing fs_repository

This commit is contained in:
Bertrand Benjamin 2024-08-14 10:44:38 +02:00
parent 08c7fbe4c5
commit 5450de8628
6 changed files with 75 additions and 17 deletions

View File

@ -10,55 +10,60 @@ class FSRepository(AbstractRepository):
self.name = name self.name = name
self.basepath = Path(basepath) self.basepath = Path(basepath)
assert self.basepath.exists()
self._metadata_engine = metadata_engine self._metadata_engine = metadata_engine
def ls(self, dir, only_files=False, only_directories=False, recursive=False) -> list[str]: def ls(
dirpath = Path(dir) self, dir="", only_files=False, only_directories=False, recursive=False
) -> list[str]:
dirpath = self.basepath / dir
if only_files: if only_files:
return [str(f.relative_to(dirpath)) for f in dirpath.iterdir() if not f.is_dir()] return [
str(f.relative_to(dirpath)) for f in dirpath.iterdir() if not f.is_dir()
]
if only_directories: if only_directories:
if recursive: if recursive:
return [str(f[0].relative_to(dirpath)) for f in dirpath.walk()] return [str(f[0].relative_to(dirpath)) for f in dirpath.walk()]
return [str(f.relative_to(dirpath)) for f in dirpath.iterdir() if f.is_dir()] return [
str(f.relative_to(dirpath)) for f in dirpath.iterdir() if f.is_dir()
]
return [str(f.relative_to(dirpath)) for f in dirpath.iterdir()] return [str(f.relative_to(dirpath)) for f in dirpath.iterdir()]
def schemas(self, recursive=True) -> list[str]: def schemas(self, recursive=True) -> list[str]:
dirpath = self.basepath return self.ls("", only_directories=True, recursive=True)
return self.ls(dirpath, only_directories=True, recursive=True)
def tables(self, schema:str) -> list[str]: def tables(self, schema: str = ".") -> list[str]:
dirpath = self.basepath / schema return self.ls(schema, only_files=True)
return self.ls(dirpath, only_files=True)
def build_table_path(self, table: str, schema: str): def build_table_path(self, table: str, schema: str):
table_path = self.basepath table_path = self.basepath
if schema == '.': if schema == ".":
return table_path / table return table_path / table
return table_path / schema / table return table_path / schema / table
def info(self, table:str, schema:str='.'): def infos(self, table: str, schema: str = "."):
table_path = self.build_table_path(table, schema) table_path = self.build_table_path(table, schema)
pass pass
def read(self, table:str, schema:str='.', read_options={}): def read(self, table: str, schema: str = ".", read_options={}):
table_path = self.build_table_path(table, schema) table_path = self.build_table_path(table, schema)
extension = table_path.suffix extension = table_path.suffix
if extension == '.csv': if extension == ".csv":
return pd.read_csv(table_path, **read_options) return pd.read_csv(table_path, **read_options)
if extension == '.xlsx': if extension == ".xlsx":
return pd.read_excel(table_path, **read_options) return pd.read_excel(table_path, **read_options)
raise ValueError("Can't open the table") raise ValueError("Can't open the table")
def write(self, content, table:str, schema:str='.'): def write(self, content, table: str, schema: str = "."):
table_path = self.build_table_path(table, schema) table_path = self.build_table_path(table, schema)
pass pass
def delete(self, table:str, schema:str='.'): def delete_table(self, table: str, schema: str = "."):
table_path = self.build_table_path(table, schema) table_path = self.build_table_path(table, schema)
pass pass

View File

Binary file not shown.

View File

@ -0,0 +1,7 @@
Username; Identifier;First name;Last name
booker12;9012;Rachel;Booker
grey07;2070;Laura;Grey
johnson81;4081;Craig;Johnson
jenkins46;9346;Mary;Jenkins
smith79;5079;Jamie;Smith
1 Username Identifier First name Last name
2 booker12 9012 Rachel Booker
3 grey07 2070 Laura Grey
4 johnson81 4081 Craig Johnson
5 jenkins46 9346 Mary Jenkins
6 smith79 5079 Jamie Smith

View File

@ -0,0 +1,46 @@
import shutil
from pathlib import Path
import pytest
from dashboard.libs.repository.fs_repository import FSRepository
EXAMPLE_DIR = "./tests/repository/fs_examples/"
@pytest.fixture
def location(tmp_path):
loc = tmp_path
username_loc = loc / "username"
username_loc.mkdir()
salary_loc = loc / "salary"
salary_loc.mkdir()
example_src = Path(EXAMPLE_DIR)
for f in example_src.glob("*"):
if "username" in str(f):
shutil.copy(f, username_loc)
else:
shutil.copy(f, salary_loc)
return loc
def test_init(location):
repo = FSRepository("example", location)
assert repo.ls() == [
"username",
"salary",
]
assert repo.schemas() == [
".",
"username",
"salary",
]
assert repo.tables() == []
assert repo.tables("username") == [
"username.csv",
"username-password-recovery-code.xlsx",
]
assert repo.tables("salary") == ["salary.pdf"]