Add check for file existence

docs/CHANGELOG.md

@@ -16,6 +16,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
  - Change CDS keys from `.cdsapirc` file to `.config/eracli.txt` file. This will avoid conflict with e.g. ADS.
  - If a user makes a request without `--splitmonths` they are warned that the behavior will change in the future, and that they have to choose between `--splitmonths False` and `--splitmonths True`.
  - When a request would encounter a Request Too Large error in the CDS API, they are warned, and given a suggestion to use `--splitmonths`.
+ - When a file already exists and would be overwritten, the user is prompted for confirmation. This should prevent accidental overwriting of files.
  - `cli.py` has been refactored to make the structure more clear. Seperate argument builders are now in their own modules.
  - The documentation has been overhauled, and now uses Markdown files & MkDocs.
 

era5cli/fetch.py

@@ -189,6 +189,9 @@ def __init__(
             )
 
     def _get_login(self):
+        # First check if the config exists, and guide the user if it does not.
+        key_management.check_era5cli_config()
+        # Only then load the keys (as they should be there now).
         self.url, self.key = key_management.load_era5cli_config()
 
     def fetch(self, dryrun=False):
@@ -263,11 +266,11 @@ def _split_variable(self):
         outputfiles = [
             self._define_outputfilename(var, self.years) for var in self.variables
         ]
+        era5cli.utils.assert_outputfiles_not_exist(outputfiles)
+
         years = len(outputfiles) * [self.years]
-        if not self.threads:
-            pool = Pool()
-        else:
-            pool = Pool(nodes=self.threads)
+
+        pool = Pool(nodes=self.threads) if self.threads else Pool()
         pool.map(self._getdata, self.variables, years, outputfiles)
 
     def _split_variable_yr(self):
@@ -277,7 +280,11 @@ def _split_variable_yr(self):
         for var in self.variables:
             outputfiles += [self._define_outputfilename(var, [yr]) for yr in self.years]
             variables += len(self.years) * [var]
+
+        era5cli.utils.assert_outputfiles_not_exist(outputfiles)
+
         years = len(self.variables) * self.years
+
         pool = Pool(nodes=self.threads) if self.threads else Pool()
         pool.map(self._getdata, variables, years, outputfiles)
 
@@ -296,6 +303,8 @@ def _split_variable_yr_month(self):
             years += [year]
             months += [month]
 
+        era5cli.utils.assert_outputfiles_not_exist(outputfiles)
+
         pool = Pool(nodes=self.threads) if self.threads else Pool()
         pool.map(self._getdata, variables, years, outputfiles, months)
 
@@ -493,6 +502,7 @@ def _exit(self):
     def _getdata(self, variables: list, years: list, outputfile: str, months=None):
         """Fetch variables using cds api call."""
         name, request = self._build_request(variables, years, months)
+
         if self.dryrun:
             print(name, request, outputfile)
         else:

era5cli/utils.py

@@ -4,6 +4,7 @@
 import shutil
 import textwrap
 from pathlib import Path
+from typing import List
 import prettytable
 from netCDF4 import Dataset
 import era5cli
@@ -205,3 +206,17 @@ def strtobool(value: str) -> bool:
         "Could not convert string to boolean. Valid inputs are:"
         f"{trues} and {falses} (case insensitive)."
     )
+
+
+def assert_outputfiles_not_exist(outputfiles: List[str]) -> None:
+    """Check if files already exist, and prompt the user if they do."""
+    if any(Path(file).exists() for file in outputfiles):
+        answer = input(
+            "\n  Some filenames already exists in this folder."
+            "\n  Do you want to overwrite them? (Y/N)"
+        )
+        if answer.lower() in ["n", "no", "nope"]:
+            raise FileExistsError(
+                "\n  One or more files already exist in this folder."
+                "\n  Please remove them, or change to a different folder to continue"
+            )

tests/test_fetch.py

@@ -1,5 +1,6 @@
 """Tests for era5cli Fetch class."""
 
+import pathlib
 import unittest.mock as mock
 import pytest
 from era5cli import _request_size
@@ -23,6 +24,14 @@
 # fmt: on
 
 
+@pytest.fixture(scope="module", autouse=True)
+def my_thing_mock():
+    with mock.patch(
+        "era5cli.fetch.key_management.check_era5cli_config", autospec=True
+    ) as _fixture:
+        yield _fixture
+
+
 def initialize(
     outputformat="netcdf",
     merge=False,
@@ -643,3 +652,15 @@ def test_area():
     with pytest.raises(ValueError):
         era5 = initialize(area=[-180, 180, -90])
         era5._build_request("total_precipitation", [2008])
+
+
+def test_file_exists():
+    with mock.patch.object(pathlib.Path, "exists", return_value=True):
+        era5 = initialize()
+
+        with mock.patch("builtins.input", return_value="Y"):
+            era5.fetch(dryrun=True)
+
+        with mock.patch("builtins.input", return_value="N"):
+            with pytest.raises(FileExistsError):
+                era5.fetch(dryrun=True)

tests/test_integration.py

@@ -7,6 +7,14 @@
 from era5cli.cli import main
 
 
+@pytest.fixture(scope="module", autouse=True)
+def my_thing_mock():
+    with mock.patch(
+        "era5cli.fetch.key_management.check_era5cli_config", autospec=True
+    ) as _fixture:
+        yield _fixture
+
+
 # combine calls with result and possible warning message
 call_result = [
     {