feat: add subcommand to clean cache files (#4)

* add silent subcommand to clean cache files

* add news fragment

* add tests for ww3 clean command

* prompt before removing all files

* remove lint

news/4.feature

@@ -0,0 +1,2 @@
+Added the ``ww3 clean`` subcommand that removes cached data files.
+

tests/test_cli.py

@@ -1,3 +1,5 @@
+import pathlib
+
 import pytest
 from click.testing import CliRunner
 
@@ -15,7 +17,7 @@ def test_command_line_interface():
     assert "version" in result.output
 
 
-@pytest.mark.parametrize("subcommand", ("url", "fetch"))
+@pytest.mark.parametrize("subcommand", ("url", "fetch", "clean"))
 def test_subcommand_help(subcommand):
     runner = CliRunner()
     result = runner.invoke(ww3, [subcommand, "--help"])
@@ -27,3 +29,18 @@ def test_noop(subcommand):
     runner = CliRunner()
     result = runner.invoke(ww3, [subcommand])
     assert result.exit_code == 0
+
+
+def test_clean(tmpdir):
+    runner = CliRunner()
+
+    with tmpdir.as_cwd():
+        data_file = pathlib.Path("multi_1.glo_30m.dp.201712.grb2")
+        data_file.touch()
+
+        assert data_file.is_file()
+        runner.invoke(ww3, ["clean", "--cache-dir=.", "--dry-run"])
+        assert data_file.is_file()
+
+        runner.invoke(ww3, ["clean", "--cache-dir=.", "--yes"])
+        assert not data_file.is_file()

wavewatch3/cli.py

@@ -1,3 +1,4 @@
+import itertools
 import os
 import pathlib
 import sys
@@ -96,6 +97,53 @@ def fetch(ctx, date, dry_run, force, file):
             print(local_file.absolute())
 
 
+@ww3.command()
+@click.option("--dry-run", is_flag=True, help="only display what would have been done")
+@click.option(
+    "--cache-dir",
+    type=click.Path(file_okay=False, path_type=pathlib.Path),
+    help="cache folder to clean",
+    default="~/.wavewatch3/data",
+)
+@click.option("--yes", is_flag=True, help="remove files without prompting")
+# @click.confirmation_option(prompt="Are you sure you want to remove all cached files?")
+@click.pass_context
+def clean(ctx, dry_run, cache_dir, yes):
+    verbose = ctx.parent.params["verbose"]
+    silent = ctx.parent.params["silent"]
+
+    region = "*"
+    quantity = "*"
+    date = "*"
+    pattern = f"multi_1.{region}.{quantity}.{date}.grb2"
+
+    cache_dir = cache_dir.expanduser()
+    cache_files = list(
+        itertools.chain(cache_dir.glob(pattern), cache_dir.glob(pattern + ".*.idx"))
+    )
+
+    total_bytes = sum([cache_file.stat().st_size for cache_file in cache_files])
+
+    if not silent and not dry_run:
+        for cache_file in cache_files:
+            out(f"{cache_file}")
+        out(f"Total size: {total_bytes // 2**20} MB")
+
+    if not dry_run and len(cache_files):
+        yes = yes or click.confirm(
+            "Are you sure you want to remove all files?", abort=True
+        )
+
+    for cache_file in cache_files:
+        if dry_run:
+            out(f"rm {cache_file}")
+        else:
+            cache_file.unlink()
+
+    if not dry_run and (verbose and not silent):
+        out(f"Removed {len(cache_files)} files ({total_bytes} bytes)")
+
+
 def _retreive_urls(urls, disable=False, force=False):
     tqdm.set_lock(RLock())
     p = Pool(initializer=tqdm.set_lock, initargs=(tqdm.get_lock(),))