From 1a5097b0105520b905dcab857976ab403d736a69 Mon Sep 17 00:00:00 2001
From: Kentaro Wada <www.kentaro.wada@gmail.com>
Date: Tue, 15 Jan 2019 18:30:04 +0000
Subject: [PATCH] Add gdown.cached_download

---
 gdown/__init__.py        |  3 ++
 gdown/cached_download.py | 59 ++++++++++++++++++++++++++++++++++++++++
 setup.py                 |  2 +-
 3 files changed, 63 insertions(+), 1 deletion(-)
 create mode 100644 gdown/cached_download.py

diff --git a/gdown/__init__.py b/gdown/__init__.py
index f2021605..3bdb96af 100644
--- a/gdown/__init__.py
+++ b/gdown/__init__.py
@@ -4,6 +4,9 @@
 
 from .download import download
 
+from .cached_download import cached_download
+from .cached_download import md5sum
+
 
 __author__ = 'Kentaro Wada <www.kentaro.wada@gmail.com>'
 __version__ = pkg_resources.get_distribution('gdown').version
diff --git a/gdown/cached_download.py b/gdown/cached_download.py
new file mode 100644
index 00000000..99002793
--- /dev/null
+++ b/gdown/cached_download.py
@@ -0,0 +1,59 @@
+import hashlib
+import os
+import os.path as osp
+import shutil
+import sys
+import tempfile
+
+import filelock
+
+from .download import download
+
+
+cache_root = osp.join(osp.expanduser('~'), '.cache/gdown')
+
+
+def md5sum(file):
+    with open(file, 'rb') as f:
+        return hashlib.md5(f.read()).hexdigest()
+
+
+def cached_download(url, path=None, md5=None, quiet=False, postprocess=None):
+    if path is None:
+        path = url.replace('/', '-SLASH-')\
+                  .replace(':', '-COLON-')\
+                  .replace('=', '-EQUAL-')\
+                  .replace('?', '-QUESTION-')
+        path = osp.join(cache_root, path)
+
+    # check existence
+    if osp.exists(path) and not md5:
+        if not quiet:
+            print('[%s] File exists.' % path)
+        return path
+    elif osp.exists(path) and md5 and md5sum(path) == md5:
+        return path
+
+    # download
+    lock_path = osp.join(cache_root, '_dl_lock')
+    try:
+        os.makedirs(osp.dirname(path))
+    except OSError:
+        pass
+    temp_root = tempfile.mkdtemp(dir=cache_root)
+    try:
+        temp_path = osp.join(temp_root, 'dl')
+        download(url, temp_path, quiet=quiet)
+        with filelock.FileLock(lock_path):
+            shutil.move(temp_path, path)
+        if not quiet:
+            print('Saved to: {}'.format(path), file=sys.stderr)
+    except Exception:
+        shutil.rmtree(temp_root)
+        raise
+
+    # postprocess
+    if postprocess is not None:
+        postprocess(path)
+
+    return path
diff --git a/setup.py b/setup.py
index 4f7f5f94..6ba781ab 100644
--- a/setup.py
+++ b/setup.py
@@ -44,7 +44,7 @@
     name='gdown',
     version=version,
     packages=find_packages(),
-    install_requires=['requests', 'six', 'tqdm'],
+    install_requires=['filelock', 'requests', 'six', 'tqdm'],
     description='Google Drive direct download of big files.',
     long_description=long_description,
     long_description_content_type='text/markdown',