From f0c12fbdce04d09bf13b141d549e5e385c64caad Mon Sep 17 00:00:00 2001
From: Hind-M <hind.montassif@gmail.com>
Date: Tue, 1 Jun 2021 18:39:31 +0200
Subject: First version allowing to fetch remote datasets

---
 src/python/test/test_remote_datasets.py | 22 ++++++++++++++++++++++
 1 file changed, 22 insertions(+)
 create mode 100644 src/python/test/test_remote_datasets.py

(limited to 'src/python/test')

diff --git a/src/python/test/test_remote_datasets.py b/src/python/test/test_remote_datasets.py
new file mode 100644
index 00000000..c4e752a7
--- /dev/null
+++ b/src/python/test/test_remote_datasets.py
@@ -0,0 +1,22 @@
+# This file is part of the Gudhi Library - https://gudhi.inria.fr/ - which is released under MIT.
+# See file LICENSE or go to https://gudhi.inria.fr/licensing/ for full license details.
+# Author(s):       Hind Montassif
+#
+# Copyright (C) 2021 Inria
+#
+# Modification(s):
+#   - YYYY/MM Author: Description of the modification
+
+
+from gudhi.datasets import remote
+
+def test_fetch_remote_datasets():
+    # Test files download from given urls
+    assert 'remote_datasets/spiral_2d.csv' == remote.fetch("https://raw.githubusercontent.com/Hind-M/gudhi-data/main/spiral_2d.csv", "spiral_2d.csv")
+    assert 'remote_datasets/sphere3D_pts_on_grid.off' == remote.fetch("https://raw.githubusercontent.com/Hind-M/gudhi-data/main/sphere3D_pts_on_grid.off", "sphere3D_pts_on_grid.off")
+
+    # Test files download with checksums provided
+    assert 'remote_datasets/spiral_2d.csv' == remote.fetch("https://raw.githubusercontent.com/Hind-M/gudhi-data/main/spiral_2d.csv", "spiral_2d.csv", checksum_flag = True,
+                                                     file_checksum = '37530355d980d957c4ec06b18c775f90a91e446107d06c6201c9b4000b077f38')
+    assert 'remote_datasets/sphere3D_pts_on_grid.off' == remote.fetch("https://raw.githubusercontent.com/Hind-M/gudhi-data/main/sphere3D_pts_on_grid.off", "sphere3D_pts_on_grid.off",
+                                                                      checksum_flag = True, file_checksum = '32f96d2cafb1177f0dd5e0a019b6ff5658e14a619a7815ae55ad0fc5e8bd3f88')
-- 
cgit v1.2.3


From c2f0cf79af04ea3586a70c0a121a200353e989ac Mon Sep 17 00:00:00 2001
From: Hind-M <hind.montassif@gmail.com>
Date: Wed, 2 Jun 2021 11:30:09 +0200
Subject: Add wrapping function to fecth spiral_2d.csv directly

---
 src/python/gudhi/datasets/remote.py     | 24 +++++++++++++++++++++---
 src/python/test/test_remote_datasets.py |  3 +++
 2 files changed, 24 insertions(+), 3 deletions(-)

(limited to 'src/python/test')

diff --git a/src/python/gudhi/datasets/remote.py b/src/python/gudhi/datasets/remote.py
index 27076785..4a300b15 100644
--- a/src/python/gudhi/datasets/remote.py
+++ b/src/python/gudhi/datasets/remote.py
@@ -46,11 +46,11 @@ def fetch(url, filename, dirname = "remote_datasets", checksum_flag = False, fil
     Parameters
     ----------
     url : string
-        The url to fetch the dataset from
+        The url to fetch the dataset from.
     filename : string
-        The filename to download
+        The name to give to downloaded file.
     dirname : string
-        The directory to save the file to.
+        The directory to save the file to. Default is "remote_datasets".
     checksum_flag : boolean
         To set if the user wants the file checksum. Default is 'False'.
         Note that if checksum_flag is set to 'True', the file_checksum must be provided.
@@ -83,3 +83,21 @@ def fetch(url, filename, dirname = "remote_datasets", checksum_flag = False, fil
                                                       file_checksum))
 
     return file_path
+
+def fetch_spiral_2d(filename = "spiral_2d.csv", dirname = "remote_datasets"):
+    """
+    Fetch spiral_2d.csv remotely
+
+    Parameters
+    ----------
+    filename : string
+        The name to give to downloaded file. Default is "spiral_2d.csv"
+    dirname : string
+        The directory to save the file to. Default is "remote_datasets".
+
+    Returns
+    -------
+    file_path: string
+        Full path of the created file.
+    """
+    return fetch("https://raw.githubusercontent.com/Hind-M/gudhi-data/main/spiral_2d.csv", filename, dirname, True, '37530355d980d957c4ec06b18c775f90a91e446107d06c6201c9b4000b077f38')
diff --git a/src/python/test/test_remote_datasets.py b/src/python/test/test_remote_datasets.py
index c4e752a7..dc854e25 100644
--- a/src/python/test/test_remote_datasets.py
+++ b/src/python/test/test_remote_datasets.py
@@ -20,3 +20,6 @@ def test_fetch_remote_datasets():
                                                      file_checksum = '37530355d980d957c4ec06b18c775f90a91e446107d06c6201c9b4000b077f38')
     assert 'remote_datasets/sphere3D_pts_on_grid.off' == remote.fetch("https://raw.githubusercontent.com/Hind-M/gudhi-data/main/sphere3D_pts_on_grid.off", "sphere3D_pts_on_grid.off",
                                                                       checksum_flag = True, file_checksum = '32f96d2cafb1177f0dd5e0a019b6ff5658e14a619a7815ae55ad0fc5e8bd3f88')
+
+    # Test spiral_2d.csv wrapping function
+    assert 'remote_datasets/spiral_2d.csv' == remote.fetch_spiral_2d()
-- 
cgit v1.2.3


From baa2e67036dae8ec63321a4d9ff4e913780a8757 Mon Sep 17 00:00:00 2001
From: Hind-M <hind.montassif@gmail.com>
Date: Wed, 2 Jun 2021 16:00:40 +0200
Subject: Modify test to consider both slash and backslash in the returned file
 path

---
 src/python/test/test_remote_datasets.py | 29 ++++++++++++++++++++++++-----
 1 file changed, 24 insertions(+), 5 deletions(-)

(limited to 'src/python/test')

diff --git a/src/python/test/test_remote_datasets.py b/src/python/test/test_remote_datasets.py
index dc854e25..a822ebaa 100644
--- a/src/python/test/test_remote_datasets.py
+++ b/src/python/test/test_remote_datasets.py
@@ -9,17 +9,36 @@
 
 
 from gudhi.datasets import remote
+import re
 
 def test_fetch_remote_datasets():
     # Test files download from given urls
-    assert 'remote_datasets/spiral_2d.csv' == remote.fetch("https://raw.githubusercontent.com/Hind-M/gudhi-data/main/spiral_2d.csv", "spiral_2d.csv")
-    assert 'remote_datasets/sphere3D_pts_on_grid.off' == remote.fetch("https://raw.githubusercontent.com/Hind-M/gudhi-data/main/sphere3D_pts_on_grid.off", "sphere3D_pts_on_grid.off")
+    path_file_dw = remote.fetch("https://raw.githubusercontent.com/Hind-M/gudhi-data/main/spiral_2d.csv", "spiral_2d.csv")
+    names_dw = re.split(r' |/|\\', path_file_dw)
+    assert 'remote_datasets' == names_dw[0]
+    assert 'spiral_2d.csv' == names_dw[1]
+
+    path_file_dw = remote.fetch("https://raw.githubusercontent.com/Hind-M/gudhi-data/main/sphere3D_pts_on_grid.off", "sphere3D_pts_on_grid.off")
+    names_dw = re.split(r' |/|\\', path_file_dw)
+    assert 'remote_datasets' == names_dw[0]
+    assert 'sphere3D_pts_on_grid.off' == names_dw[1]
+
 
     # Test files download with checksums provided
-    assert 'remote_datasets/spiral_2d.csv' == remote.fetch("https://raw.githubusercontent.com/Hind-M/gudhi-data/main/spiral_2d.csv", "spiral_2d.csv", checksum_flag = True,
+    path_file_dw = remote.fetch("https://raw.githubusercontent.com/Hind-M/gudhi-data/main/spiral_2d.csv", "spiral_2d.csv", checksum_flag = True,
                                                      file_checksum = '37530355d980d957c4ec06b18c775f90a91e446107d06c6201c9b4000b077f38')
-    assert 'remote_datasets/sphere3D_pts_on_grid.off' == remote.fetch("https://raw.githubusercontent.com/Hind-M/gudhi-data/main/sphere3D_pts_on_grid.off", "sphere3D_pts_on_grid.off",
+    names_dw = re.split(r' |/|\\', path_file_dw)
+    assert 'remote_datasets' == names_dw[0]
+    assert 'spiral_2d.csv' == names_dw[1]
+
+    path_file_dw = remote.fetch("https://raw.githubusercontent.com/Hind-M/gudhi-data/main/sphere3D_pts_on_grid.off", "sphere3D_pts_on_grid.off",
                                                                       checksum_flag = True, file_checksum = '32f96d2cafb1177f0dd5e0a019b6ff5658e14a619a7815ae55ad0fc5e8bd3f88')
+    names_dw = re.split(r' |/|\\', path_file_dw)
+    assert 'remote_datasets' == names_dw[0]
+    assert 'sphere3D_pts_on_grid.off' == names_dw[1]
 
     # Test spiral_2d.csv wrapping function
-    assert 'remote_datasets/spiral_2d.csv' == remote.fetch_spiral_2d()
+    path_file_dw = remote.fetch_spiral_2d()
+    names_dw = re.split(r' |/|\\', path_file_dw)
+    assert 'remote_datasets' == names_dw[0]
+    assert 'spiral_2d.csv' == names_dw[1]
-- 
cgit v1.2.3


From 3ee453718eebc7274b19caef4b79d8ec2754d583 Mon Sep 17 00:00:00 2001
From: Hind-M <hind.montassif@gmail.com>
Date: Thu, 3 Jun 2021 16:40:59 +0200
Subject: Modify urls to point to GUDHI/gudhi-data repo

---
 src/python/gudhi/datasets/remote.py     | 3 ++-
 src/python/test/test_remote_datasets.py | 8 ++++----
 2 files changed, 6 insertions(+), 5 deletions(-)

(limited to 'src/python/test')

diff --git a/src/python/gudhi/datasets/remote.py b/src/python/gudhi/datasets/remote.py
index 4a300b15..525a7b66 100644
--- a/src/python/gudhi/datasets/remote.py
+++ b/src/python/gudhi/datasets/remote.py
@@ -100,4 +100,5 @@ def fetch_spiral_2d(filename = "spiral_2d.csv", dirname = "remote_datasets"):
     file_path: string
         Full path of the created file.
     """
-    return fetch("https://raw.githubusercontent.com/Hind-M/gudhi-data/main/spiral_2d.csv", filename, dirname, True, '37530355d980d957c4ec06b18c775f90a91e446107d06c6201c9b4000b077f38')
+    return fetch("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/spiral_2d.csv", filename, dirname, True,
+                 '37530355d980d957c4ec06b18c775f90a91e446107d06c6201c9b4000b077f38')
diff --git a/src/python/test/test_remote_datasets.py b/src/python/test/test_remote_datasets.py
index a822ebaa..63ad7885 100644
--- a/src/python/test/test_remote_datasets.py
+++ b/src/python/test/test_remote_datasets.py
@@ -13,25 +13,25 @@ import re
 
 def test_fetch_remote_datasets():
     # Test files download from given urls
-    path_file_dw = remote.fetch("https://raw.githubusercontent.com/Hind-M/gudhi-data/main/spiral_2d.csv", "spiral_2d.csv")
+    path_file_dw = remote.fetch("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/spiral_2d.csv", "spiral_2d.csv")
     names_dw = re.split(r' |/|\\', path_file_dw)
     assert 'remote_datasets' == names_dw[0]
     assert 'spiral_2d.csv' == names_dw[1]
 
-    path_file_dw = remote.fetch("https://raw.githubusercontent.com/Hind-M/gudhi-data/main/sphere3D_pts_on_grid.off", "sphere3D_pts_on_grid.off")
+    path_file_dw = remote.fetch("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/sphere3D_pts_on_grid.off", "sphere3D_pts_on_grid.off")
     names_dw = re.split(r' |/|\\', path_file_dw)
     assert 'remote_datasets' == names_dw[0]
     assert 'sphere3D_pts_on_grid.off' == names_dw[1]
 
 
     # Test files download with checksums provided
-    path_file_dw = remote.fetch("https://raw.githubusercontent.com/Hind-M/gudhi-data/main/spiral_2d.csv", "spiral_2d.csv", checksum_flag = True,
+    path_file_dw = remote.fetch("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/spiral_2d.csv", "spiral_2d.csv", checksum_flag = True,
                                                      file_checksum = '37530355d980d957c4ec06b18c775f90a91e446107d06c6201c9b4000b077f38')
     names_dw = re.split(r' |/|\\', path_file_dw)
     assert 'remote_datasets' == names_dw[0]
     assert 'spiral_2d.csv' == names_dw[1]
 
-    path_file_dw = remote.fetch("https://raw.githubusercontent.com/Hind-M/gudhi-data/main/sphere3D_pts_on_grid.off", "sphere3D_pts_on_grid.off",
+    path_file_dw = remote.fetch("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/sphere3D_pts_on_grid.off", "sphere3D_pts_on_grid.off",
                                                                       checksum_flag = True, file_checksum = '32f96d2cafb1177f0dd5e0a019b6ff5658e14a619a7815ae55ad0fc5e8bd3f88')
     names_dw = re.split(r' |/|\\', path_file_dw)
     assert 'remote_datasets' == names_dw[0]
-- 
cgit v1.2.3


From bbe2e25a204be50eb422db71b4cf314b92797d4e Mon Sep 17 00:00:00 2001
From: Hind-M <hind.montassif@gmail.com>
Date: Fri, 4 Jun 2021 12:21:11 +0200
Subject: Remove checksum_flag parameter and use value of 'file_checksum is not
 None' instead

---
 src/python/gudhi/datasets/remote.py     | 13 +++----------
 src/python/test/test_remote_datasets.py |  7 ++++---
 2 files changed, 7 insertions(+), 13 deletions(-)

(limited to 'src/python/test')

diff --git a/src/python/gudhi/datasets/remote.py b/src/python/gudhi/datasets/remote.py
index 525a7b66..fdd20f74 100644
--- a/src/python/gudhi/datasets/remote.py
+++ b/src/python/gudhi/datasets/remote.py
@@ -39,7 +39,7 @@ def _checksum_sha256(file_path):
             sha256_hash.update(buffer)
     return sha256_hash.hexdigest()
 
-def fetch(url, filename, dirname = "remote_datasets", checksum_flag = False, file_checksum = None):
+def fetch(url, filename, dirname = "remote_datasets", file_checksum = None):
     """
     Fetch the wanted dataset from the given url and save it in file_path
 
@@ -51,12 +51,8 @@ def fetch(url, filename, dirname = "remote_datasets", checksum_flag = False, fil
         The name to give to downloaded file.
     dirname : string
         The directory to save the file to. Default is "remote_datasets".
-    checksum_flag : boolean
-        To set if the user wants the file checksum. Default is 'False'.
-        Note that if checksum_flag is set to 'True', the file_checksum must be provided.
     file_checksum : string
         The file checksum using sha256 to check against the one computed on the downloaded file.
-        To be considered, checksum_flag must be set to 'True'.
         Default is 'None'.
 
     Returns
@@ -71,10 +67,7 @@ def fetch(url, filename, dirname = "remote_datasets", checksum_flag = False, fil
 
     urlretrieve(url, file_path)
 
-    if (checksum_flag):
-        if file_checksum is None:
-            raise ValueError("The file checksum must be provided - different from None - for the check to be performed.")
-
+    if file_checksum is not None:
         checksum = _checksum_sha256(file_path)
         if file_checksum != checksum:
             raise IOError("{} has a SHA256 checksum : {}, "
@@ -100,5 +93,5 @@ def fetch_spiral_2d(filename = "spiral_2d.csv", dirname = "remote_datasets"):
     file_path: string
         Full path of the created file.
     """
-    return fetch("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/spiral_2d.csv", filename, dirname, True,
+    return fetch("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/spiral_2d.csv", filename, dirname,
                  '37530355d980d957c4ec06b18c775f90a91e446107d06c6201c9b4000b077f38')
diff --git a/src/python/test/test_remote_datasets.py b/src/python/test/test_remote_datasets.py
index 63ad7885..6c9217c8 100644
--- a/src/python/test/test_remote_datasets.py
+++ b/src/python/test/test_remote_datasets.py
@@ -25,14 +25,15 @@ def test_fetch_remote_datasets():
 
 
     # Test files download with checksums provided
-    path_file_dw = remote.fetch("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/spiral_2d.csv", "spiral_2d.csv", checksum_flag = True,
-                                                     file_checksum = '37530355d980d957c4ec06b18c775f90a91e446107d06c6201c9b4000b077f38')
+    path_file_dw = remote.fetch("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/spiral_2d.csv", "spiral_2d.csv",
+                                file_checksum = '37530355d980d957c4ec06b18c775f90a91e446107d06c6201c9b4000b077f38')
     names_dw = re.split(r' |/|\\', path_file_dw)
     assert 'remote_datasets' == names_dw[0]
     assert 'spiral_2d.csv' == names_dw[1]
 
     path_file_dw = remote.fetch("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/sphere3D_pts_on_grid.off", "sphere3D_pts_on_grid.off",
-                                                                      checksum_flag = True, file_checksum = '32f96d2cafb1177f0dd5e0a019b6ff5658e14a619a7815ae55ad0fc5e8bd3f88')
+                                file_checksum = '32f96d2cafb1177f0dd5e0a019b6ff5658e14a619a7815ae55ad0fc5e8bd3f88')
+
     names_dw = re.split(r' |/|\\', path_file_dw)
     assert 'remote_datasets' == names_dw[0]
     assert 'sphere3D_pts_on_grid.off' == names_dw[1]
-- 
cgit v1.2.3


From 82524c5b0a6ab02b020574b2200a8721f3ed424c Mon Sep 17 00:00:00 2001
From: Hind-M <hind.montassif@gmail.com>
Date: Mon, 7 Jun 2021 15:03:14 +0200
Subject: Add test with wrong checksum Add functions to avoid redundant code

---
 src/python/test/test_remote_datasets.py | 43 +++++++++++++++++----------------
 1 file changed, 22 insertions(+), 21 deletions(-)

(limited to 'src/python/test')

diff --git a/src/python/test/test_remote_datasets.py b/src/python/test/test_remote_datasets.py
index 6c9217c8..e252980d 100644
--- a/src/python/test/test_remote_datasets.py
+++ b/src/python/test/test_remote_datasets.py
@@ -10,36 +10,37 @@
 
 from gudhi.datasets import remote
 import re
+import os.path
+import pytest
 
-def test_fetch_remote_datasets():
-    # Test files download from given urls
-    path_file_dw = remote.fetch("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/spiral_2d.csv", "spiral_2d.csv")
-    names_dw = re.split(r' |/|\\', path_file_dw)
-    assert 'remote_datasets' == names_dw[0]
-    assert 'spiral_2d.csv' == names_dw[1]
+def check_dir_file_names(path_file_dw, filename, dirname):
+    assert os.path.isfile(path_file_dw)
 
-    path_file_dw = remote.fetch("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/sphere3D_pts_on_grid.off", "sphere3D_pts_on_grid.off")
     names_dw = re.split(r' |/|\\', path_file_dw)
-    assert 'remote_datasets' == names_dw[0]
-    assert 'sphere3D_pts_on_grid.off' == names_dw[1]
+    assert dirname == names_dw[0]
+    assert filename == names_dw[1]
 
+def check_fetch_output(url, filename, dirname = "remote_datasets", file_checksum = None):
+    path_file_dw = remote.fetch(url, filename, dirname, file_checksum)
+    check_dir_file_names(path_file_dw, filename, dirname)
+
+def test_fetch_remote_datasets():
+    # Test fetch with a wrong checksum
+    with pytest.raises(OSError):
+        check_fetch_output("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/spiral_2d.csv", "spiral_2d.csv", file_checksum = 'XXXXXXXXXX')
 
-    # Test files download with checksums provided
-    path_file_dw = remote.fetch("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/spiral_2d.csv", "spiral_2d.csv",
+    # Test files download from given urls with checksums provided
+    check_fetch_output("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/spiral_2d.csv", "spiral_2d.csv",
                                 file_checksum = '37530355d980d957c4ec06b18c775f90a91e446107d06c6201c9b4000b077f38')
-    names_dw = re.split(r' |/|\\', path_file_dw)
-    assert 'remote_datasets' == names_dw[0]
-    assert 'spiral_2d.csv' == names_dw[1]
 
-    path_file_dw = remote.fetch("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/sphere3D_pts_on_grid.off", "sphere3D_pts_on_grid.off",
+    check_fetch_output("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/sphere3D_pts_on_grid.off", "sphere3D_pts_on_grid.off",
                                 file_checksum = '32f96d2cafb1177f0dd5e0a019b6ff5658e14a619a7815ae55ad0fc5e8bd3f88')
 
-    names_dw = re.split(r' |/|\\', path_file_dw)
-    assert 'remote_datasets' == names_dw[0]
-    assert 'sphere3D_pts_on_grid.off' == names_dw[1]
+    # Test files download from given urls without checksums
+    check_fetch_output("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/spiral_2d.csv", "spiral_2d.csv")
+
+    check_fetch_output("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/sphere3D_pts_on_grid.off", "sphere3D_pts_on_grid.off")
 
     # Test spiral_2d.csv wrapping function
     path_file_dw = remote.fetch_spiral_2d()
-    names_dw = re.split(r' |/|\\', path_file_dw)
-    assert 'remote_datasets' == names_dw[0]
-    assert 'spiral_2d.csv' == names_dw[1]
+    check_dir_file_names(path_file_dw, 'spiral_2d.csv', 'remote_datasets')
-- 
cgit v1.2.3


From aa600c433e1f756bec4323e29e86786b937d9443 Mon Sep 17 00:00:00 2001
From: Hind-M <hind.montassif@gmail.com>
Date: Mon, 15 Nov 2021 11:12:27 +0100
Subject: Print files licenses when available Wrap bunny fetching Add
 corresponding tests

---
 src/python/gudhi/datasets/remote.py     | 38 ++++++++++++++++++++++-
 src/python/test/test_remote_datasets.py | 55 ++++++++++++++++++++++++++-------
 2 files changed, 81 insertions(+), 12 deletions(-)

(limited to 'src/python/test')

diff --git a/src/python/gudhi/datasets/remote.py b/src/python/gudhi/datasets/remote.py
index aef4b277..7e8f9ce7 100644
--- a/src/python/gudhi/datasets/remote.py
+++ b/src/python/gudhi/datasets/remote.py
@@ -39,7 +39,7 @@ def _checksum_sha256(file_path):
             sha256_hash.update(buffer)
     return sha256_hash.hexdigest()
 
-def fetch(url, filename, dirname = "remote_datasets", file_checksum = None):
+def fetch(url, filename, dirname = "remote_datasets", file_checksum = None, accept_license = False):
     """
     Fetch the wanted dataset from the given url and save it in file_path
 
@@ -54,6 +54,9 @@ def fetch(url, filename, dirname = "remote_datasets", file_checksum = None):
     file_checksum : string
         The file checksum using sha256 to check against the one computed on the downloaded file.
         Default is 'None'.
+    accept_license : boolean
+        Flag to specify if user accepts the file LICENSE and prevents from printing the corresponding license terms.
+        Default is False
 
     Returns
     -------
@@ -69,6 +72,7 @@ def fetch(url, filename, dirname = "remote_datasets", file_checksum = None):
         if not exists(dirname):
             makedirs(dirname)
 
+        # Get the file
         urlretrieve(url, file_path)
 
     if file_checksum is not None:
@@ -78,6 +82,13 @@ def fetch(url, filename, dirname = "remote_datasets", file_checksum = None):
                         "different from expected : {}."
                         "The file may be corrupted or the given url may be wrong !".format(file_path, checksum, file_checksum))
 
+    # Print license terms unless accept_license is set to True
+    if not accept_license:
+        license_file = join(dirname, "LICENSE")
+        if exists(license_file) and (file_path != license_file):
+            with open(license_file, 'r') as f:
+                print(f.read())
+
     return file_path
 
 def fetch_spiral_2d(filename = "spiral_2d.csv", dirname = "remote_datasets"):
@@ -98,3 +109,28 @@ def fetch_spiral_2d(filename = "spiral_2d.csv", dirname = "remote_datasets"):
     """
     return fetch("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/spiral_2d.csv", filename, dirname,
                  '37530355d980d957c4ec06b18c775f90a91e446107d06c6201c9b4000b077f38')
+
+def fetch_bunny(filename = "bunny.off", dirname = "remote_datasets/bunny", accept_license = False):
+    """
+    Fetch bunny.off remotely and its LICENSE file
+
+    Parameters
+    ----------
+    filename : string
+        The name to give to downloaded file. Default is "bunny.off"
+    dirname : string
+        The directory to save the file to. Default is "remote_datasets/bunny".
+    accept_license : boolean
+        Flag to specify if user accepts the file LICENSE and prevents from printing the corresponding license terms.
+        Default is False
+
+    Returns
+    -------
+    files_paths: list of strings
+        Full paths of the created file and its LICENSE.
+    """
+
+    return [fetch("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points//bunny/LICENSE", "LICENSE", dirname,
+                 'aeb1bad319b7d74fa0b8076358182f9c6b1284c67cc07dc67cbc9bc73025d956'),
+            fetch("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points//bunny/bunny.off", filename, dirname,
+                 '11852d5e73e2d4bd7b86a2c5cc8a5884d0fbb72539493e8cec100ea922b19f5b', accept_license)]
diff --git a/src/python/test/test_remote_datasets.py b/src/python/test/test_remote_datasets.py
index e252980d..e777abc6 100644
--- a/src/python/test/test_remote_datasets.py
+++ b/src/python/test/test_remote_datasets.py
@@ -11,36 +11,69 @@
 from gudhi.datasets import remote
 import re
 import os.path
+import io
+import sys
 import pytest
 
-def check_dir_file_names(path_file_dw, filename, dirname):
+def _check_dir_file_names(path_file_dw, filename, dirname):
     assert os.path.isfile(path_file_dw)
 
     names_dw = re.split(r' |/|\\', path_file_dw)
-    assert dirname == names_dw[0]
-    assert filename == names_dw[1]
+    # Case where inner directories are created in "remote_datasets/"; e.g: "remote_datasets/bunny"
+    if len(names_dw) >= 3:
+        for i in range(len(names_dw)-1):
+            assert re.split(r' |/|\\', dirname)[i] == names_dw[i]
+        assert filename == names_dw[i+1]
+    else:
+        assert dirname == names_dw[0]
+        assert filename == names_dw[1]
 
-def check_fetch_output(url, filename, dirname = "remote_datasets", file_checksum = None):
+def _check_fetch_output(url, filename, dirname = "remote_datasets", file_checksum = None):
     path_file_dw = remote.fetch(url, filename, dirname, file_checksum)
-    check_dir_file_names(path_file_dw, filename, dirname)
+    _check_dir_file_names(path_file_dw, filename, dirname)
+
+def _get_bunny_license_print(accept_license = False):
+    capturedOutput = io.StringIO()
+    # Redirect stdout
+    sys.stdout = capturedOutput
+    remote.fetch("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points//bunny/bunny.off", "bunny.off", "remote_datasets/bunny",
+                 '11852d5e73e2d4bd7b86a2c5cc8a5884d0fbb72539493e8cec100ea922b19f5b', accept_license)
+    # Reset redirect
+    sys.stdout = sys.__stdout__
+    return capturedOutput
 
 def test_fetch_remote_datasets():
     # Test fetch with a wrong checksum
     with pytest.raises(OSError):
-        check_fetch_output("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/spiral_2d.csv", "spiral_2d.csv", file_checksum = 'XXXXXXXXXX')
+        _check_fetch_output("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/spiral_2d.csv", "spiral_2d.csv", file_checksum = 'XXXXXXXXXX')
 
     # Test files download from given urls with checksums provided
-    check_fetch_output("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/spiral_2d.csv", "spiral_2d.csv",
+    _check_fetch_output("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/spiral_2d.csv", "spiral_2d.csv",
                                 file_checksum = '37530355d980d957c4ec06b18c775f90a91e446107d06c6201c9b4000b077f38')
 
-    check_fetch_output("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/sphere3D_pts_on_grid.off", "sphere3D_pts_on_grid.off",
+    _check_fetch_output("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/sphere3D_pts_on_grid.off", "sphere3D_pts_on_grid.off",
                                 file_checksum = '32f96d2cafb1177f0dd5e0a019b6ff5658e14a619a7815ae55ad0fc5e8bd3f88')
 
     # Test files download from given urls without checksums
-    check_fetch_output("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/spiral_2d.csv", "spiral_2d.csv")
+    _check_fetch_output("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/spiral_2d.csv", "spiral_2d.csv")
 
-    check_fetch_output("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/sphere3D_pts_on_grid.off", "sphere3D_pts_on_grid.off")
+    _check_fetch_output("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/sphere3D_pts_on_grid.off", "sphere3D_pts_on_grid.off")
 
     # Test spiral_2d.csv wrapping function
     path_file_dw = remote.fetch_spiral_2d()
-    check_dir_file_names(path_file_dw, 'spiral_2d.csv', 'remote_datasets')
+    _check_dir_file_names(path_file_dw, 'spiral_2d.csv', 'remote_datasets')
+
+    # Test printing existing LICENSE file when fetching bunny.off with accept_license = False (default)
+    # Fetch LICENSE file
+    remote.fetch("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points//bunny/LICENSE", "LICENSE", "remote_datasets/bunny",
+                 'aeb1bad319b7d74fa0b8076358182f9c6b1284c67cc07dc67cbc9bc73025d956')
+    with open("remote_datasets/bunny/LICENSE") as f:
+        assert f.read() == _get_bunny_license_print().getvalue().rstrip("\n")
+
+    # Test not printing bunny.off LICENSE when accept_license = True
+    assert "" == _get_bunny_license_print(accept_license = True).getvalue()
+
+    # Test fetch_bunny wrapping function
+    path_file_dw = remote.fetch_bunny()
+    _check_dir_file_names(path_file_dw[0], 'LICENSE', 'remote_datasets/bunny')
+    _check_dir_file_names(path_file_dw[1], 'bunny.off', 'remote_datasets/bunny')
-- 
cgit v1.2.3


From 8d1e7aeb3416194d00f45587d1ecea85ba218028 Mon Sep 17 00:00:00 2001
From: Hind-M <hind.montassif@gmail.com>
Date: Fri, 28 Jan 2022 16:21:33 +0100
Subject: Return arrays of points instead of files paths when fetching
 bunny.npy and spiral_2d.csv

---
 src/python/gudhi/datasets/remote.py     | 83 +++++++++++++++++++++------------
 src/python/test/test_remote_datasets.py | 33 +++++++------
 2 files changed, 72 insertions(+), 44 deletions(-)

(limited to 'src/python/test')

diff --git a/src/python/gudhi/datasets/remote.py b/src/python/gudhi/datasets/remote.py
index 7e8f9ce7..ef797417 100644
--- a/src/python/gudhi/datasets/remote.py
+++ b/src/python/gudhi/datasets/remote.py
@@ -7,17 +7,17 @@
 # Modification(s):
 #   - YYYY/MM Author: Description of the modification
 
-import hashlib
-
 from os.path import join, exists
 from os import makedirs
 
 from urllib.request import urlretrieve
+import hashlib
 
+import numpy as np
 
 def _checksum_sha256(file_path):
     """
-    Compute the file checksum using sha256
+    Compute the file checksum using sha256.
 
     Parameters
     ----------
@@ -26,7 +26,7 @@ def _checksum_sha256(file_path):
 
     Returns
     -------
-        The hex digest of file_path
+        The hex digest of file_path.
     """
     sha256_hash = hashlib.sha256()
     chunk_size = 4096
@@ -39,9 +39,9 @@ def _checksum_sha256(file_path):
             sha256_hash.update(buffer)
     return sha256_hash.hexdigest()
 
-def fetch(url, filename, dirname = "remote_datasets", file_checksum = None, accept_license = False):
+def _fetch_remote(url, filename, dirname = "remote_datasets", file_checksum = None, accept_license = False):
     """
-    Fetch the wanted dataset from the given url and save it in file_path
+    Fetch the wanted dataset from the given url and save it in file_path.
 
     Parameters
     ----------
@@ -56,7 +56,7 @@ def fetch(url, filename, dirname = "remote_datasets", file_checksum = None, acce
         Default is 'None'.
     accept_license : boolean
         Flag to specify if user accepts the file LICENSE and prevents from printing the corresponding license terms.
-        Default is False
+        Default is False.
 
     Returns
     -------
@@ -66,14 +66,8 @@ def fetch(url, filename, dirname = "remote_datasets", file_checksum = None, acce
 
     file_path = join(dirname, filename)
 
-    # Check for an already existing file at file_path
-    if not exists(file_path):
-        # Create directory if not existing
-        if not exists(dirname):
-            makedirs(dirname)
-
-        # Get the file
-        urlretrieve(url, file_path)
+    # Get the file
+    urlretrieve(url, file_path)
 
     if file_checksum is not None:
         checksum = _checksum_sha256(file_path)
@@ -93,44 +87,71 @@ def fetch(url, filename, dirname = "remote_datasets", file_checksum = None, acce
 
 def fetch_spiral_2d(filename = "spiral_2d.csv", dirname = "remote_datasets"):
     """
-    Fetch spiral_2d.csv remotely
+    Fetch "spiral_2d.csv" remotely.
 
     Parameters
     ----------
     filename : string
-        The name to give to downloaded file. Default is "spiral_2d.csv"
+        The name to give to downloaded file. Default is "spiral_2d.csv".
     dirname : string
         The directory to save the file to. Default is "remote_datasets".
 
     Returns
     -------
-    file_path: string
-        Full path of the created file.
+    points: array
+        Array of points stored in "spiral_2d.csv".
     """
-    return fetch("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/spiral_2d.csv", filename, dirname,
-                 '37530355d980d957c4ec06b18c775f90a91e446107d06c6201c9b4000b077f38')
+    file_url = "https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/spiral_2d.csv"
+    file_checksum = '37530355d980d957c4ec06b18c775f90a91e446107d06c6201c9b4000b077f38'
 
-def fetch_bunny(filename = "bunny.off", dirname = "remote_datasets/bunny", accept_license = False):
+    archive_path = join(dirname, filename)
+
+    if not exists(archive_path):
+        # Create directory if not existing
+        if not exists(dirname):
+            makedirs(dirname)
+
+        file_path_pkl = _fetch_remote(file_url, filename, dirname, file_checksum)
+
+        return np.loadtxt(file_path_pkl)
+    else:
+        return np.loadtxt(archive_path)
+
+def fetch_bunny(filename = "bunny.npy", dirname = "remote_datasets/bunny", accept_license = False):
     """
-    Fetch bunny.off remotely and its LICENSE file
+    Fetch "bunny.npy" remotely and its LICENSE file.
 
     Parameters
     ----------
     filename : string
-        The name to give to downloaded file. Default is "bunny.off"
+        The name to give to downloaded file. Default is "bunny.npy".
     dirname : string
         The directory to save the file to. Default is "remote_datasets/bunny".
     accept_license : boolean
         Flag to specify if user accepts the file LICENSE and prevents from printing the corresponding license terms.
-        Default is False
+        Default is False.
 
     Returns
     -------
-    files_paths: list of strings
-        Full paths of the created file and its LICENSE.
+    points: array
+        Array of points stored in "bunny.npy".
     """
 
-    return [fetch("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points//bunny/LICENSE", "LICENSE", dirname,
-                 'aeb1bad319b7d74fa0b8076358182f9c6b1284c67cc07dc67cbc9bc73025d956'),
-            fetch("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points//bunny/bunny.off", filename, dirname,
-                 '11852d5e73e2d4bd7b86a2c5cc8a5884d0fbb72539493e8cec100ea922b19f5b', accept_license)]
+    file_url = "https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/bunny/bunny.npy"
+    file_checksum = '13f7842ebb4b45370e50641ff28c88685703efa5faab14edf0bb7d113a965e1b'
+    license_url = "https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/bunny/LICENSE"
+    license_checksum = 'aeb1bad319b7d74fa0b8076358182f9c6b1284c67cc07dc67cbc9bc73025d956'
+
+    archive_path = join(dirname, filename)
+
+    if not exists(archive_path):
+        # Create directory if not existing
+        if not exists(dirname):
+            makedirs(dirname)
+
+        license_path = _fetch_remote(license_url, "LICENSE", dirname, license_checksum)
+        file_path_pkl = _fetch_remote(file_url, filename, dirname, file_checksum, accept_license)
+
+        return np.load(file_path_pkl, mmap_mode='r')
+    else:
+        return np.load(archive_path, mmap_mode='r')
diff --git a/src/python/test/test_remote_datasets.py b/src/python/test/test_remote_datasets.py
index e777abc6..56a273b4 100644
--- a/src/python/test/test_remote_datasets.py
+++ b/src/python/test/test_remote_datasets.py
@@ -10,13 +10,14 @@
 
 from gudhi.datasets import remote
 import re
-import os.path
+from os.path import isfile, exists
+from os import makedirs
 import io
 import sys
 import pytest
 
 def _check_dir_file_names(path_file_dw, filename, dirname):
-    assert os.path.isfile(path_file_dw)
+    assert isfile(path_file_dw)
 
     names_dw = re.split(r' |/|\\', path_file_dw)
     # Case where inner directories are created in "remote_datasets/"; e.g: "remote_datasets/bunny"
@@ -29,15 +30,20 @@ def _check_dir_file_names(path_file_dw, filename, dirname):
         assert filename == names_dw[1]
 
 def _check_fetch_output(url, filename, dirname = "remote_datasets", file_checksum = None):
-    path_file_dw = remote.fetch(url, filename, dirname, file_checksum)
+    if not exists(dirname):
+        makedirs(dirname)
+    path_file_dw = remote._fetch_remote(url, filename, dirname, file_checksum)
     _check_dir_file_names(path_file_dw, filename, dirname)
 
 def _get_bunny_license_print(accept_license = False):
     capturedOutput = io.StringIO()
     # Redirect stdout
     sys.stdout = capturedOutput
-    remote.fetch("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points//bunny/bunny.off", "bunny.off", "remote_datasets/bunny",
-                 '11852d5e73e2d4bd7b86a2c5cc8a5884d0fbb72539493e8cec100ea922b19f5b', accept_license)
+
+    if not exists("remote_datasets/bunny"):
+        makedirs("remote_datasets/bunny")
+    remote._fetch_remote("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/bunny/bunny.npy", "bunny.npy", "remote_datasets/bunny",
+                 '13f7842ebb4b45370e50641ff28c88685703efa5faab14edf0bb7d113a965e1b', accept_license)
     # Reset redirect
     sys.stdout = sys.__stdout__
     return capturedOutput
@@ -60,20 +66,21 @@ def test_fetch_remote_datasets():
     _check_fetch_output("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/sphere3D_pts_on_grid.off", "sphere3D_pts_on_grid.off")
 
     # Test spiral_2d.csv wrapping function
-    path_file_dw = remote.fetch_spiral_2d()
-    _check_dir_file_names(path_file_dw, 'spiral_2d.csv', 'remote_datasets')
+    spiral_2d_arr = remote.fetch_spiral_2d()
+    assert spiral_2d_arr.shape == (114562, 2)
 
-    # Test printing existing LICENSE file when fetching bunny.off with accept_license = False (default)
+    # Test printing existing LICENSE file when fetching bunny.npy with accept_license = False (default)
     # Fetch LICENSE file
-    remote.fetch("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points//bunny/LICENSE", "LICENSE", "remote_datasets/bunny",
+    if not exists("remote_datasets/bunny"):
+        makedirs("remote_datasets/bunny")
+    remote._fetch_remote("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/bunny/LICENSE", "LICENSE", "remote_datasets/bunny",
                  'aeb1bad319b7d74fa0b8076358182f9c6b1284c67cc07dc67cbc9bc73025d956')
     with open("remote_datasets/bunny/LICENSE") as f:
         assert f.read() == _get_bunny_license_print().getvalue().rstrip("\n")
 
-    # Test not printing bunny.off LICENSE when accept_license = True
+    # Test not printing bunny.npy LICENSE when accept_license = True
     assert "" == _get_bunny_license_print(accept_license = True).getvalue()
 
     # Test fetch_bunny wrapping function
-    path_file_dw = remote.fetch_bunny()
-    _check_dir_file_names(path_file_dw[0], 'LICENSE', 'remote_datasets/bunny')
-    _check_dir_file_names(path_file_dw[1], 'bunny.off', 'remote_datasets/bunny')
+    bunny_arr = remote.fetch_bunny()
+    assert bunny_arr.shape == (35947, 3)
-- 
cgit v1.2.3


From ad7a50fb87ed4237b9a02165eac39ae355dd5440 Mon Sep 17 00:00:00 2001
From: Hind-M <hind.montassif@gmail.com>
Date: Tue, 1 Feb 2022 10:32:03 +0100
Subject: Fetch spiral_2d.npy file instead of csv Add some modifications
 related to those done on files in gudhi-data

---
 src/python/gudhi/datasets/remote.py     | 20 ++++++++++----------
 src/python/test/test_remote_datasets.py | 14 +++++++-------
 2 files changed, 17 insertions(+), 17 deletions(-)

(limited to 'src/python/test')

diff --git a/src/python/gudhi/datasets/remote.py b/src/python/gudhi/datasets/remote.py
index ef797417..3498a645 100644
--- a/src/python/gudhi/datasets/remote.py
+++ b/src/python/gudhi/datasets/remote.py
@@ -85,24 +85,24 @@ def _fetch_remote(url, filename, dirname = "remote_datasets", file_checksum = No
 
     return file_path
 
-def fetch_spiral_2d(filename = "spiral_2d.csv", dirname = "remote_datasets"):
+def fetch_spiral_2d(filename = "spiral_2d.npy", dirname = "remote_datasets/spiral_2d"):
     """
-    Fetch "spiral_2d.csv" remotely.
+    Fetch "spiral_2d.npy" remotely.
 
     Parameters
     ----------
     filename : string
-        The name to give to downloaded file. Default is "spiral_2d.csv".
+        The name to give to downloaded file. Default is "spiral_2d.npy".
     dirname : string
-        The directory to save the file to. Default is "remote_datasets".
+        The directory to save the file to. Default is "remote_datasets/spiral_2d".
 
     Returns
     -------
     points: array
-        Array of points stored in "spiral_2d.csv".
+        Array of points stored in "spiral_2d.npy".
     """
-    file_url = "https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/spiral_2d.csv"
-    file_checksum = '37530355d980d957c4ec06b18c775f90a91e446107d06c6201c9b4000b077f38'
+    file_url = "https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/spiral_2d/spiral_2d.npy"
+    file_checksum = '88312ffd6df2e2cb2bde9c0e1f962d7d644c6f58dc369c7b377b298dacdc4eaf'
 
     archive_path = join(dirname, filename)
 
@@ -113,9 +113,9 @@ def fetch_spiral_2d(filename = "spiral_2d.csv", dirname = "remote_datasets"):
 
         file_path_pkl = _fetch_remote(file_url, filename, dirname, file_checksum)
 
-        return np.loadtxt(file_path_pkl)
+        return np.load(file_path_pkl, mmap_mode='r')
     else:
-        return np.loadtxt(archive_path)
+        return np.load(archive_path, mmap_mode='r')
 
 def fetch_bunny(filename = "bunny.npy", dirname = "remote_datasets/bunny", accept_license = False):
     """
@@ -140,7 +140,7 @@ def fetch_bunny(filename = "bunny.npy", dirname = "remote_datasets/bunny", accep
     file_url = "https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/bunny/bunny.npy"
     file_checksum = '13f7842ebb4b45370e50641ff28c88685703efa5faab14edf0bb7d113a965e1b'
     license_url = "https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/bunny/LICENSE"
-    license_checksum = 'aeb1bad319b7d74fa0b8076358182f9c6b1284c67cc07dc67cbc9bc73025d956'
+    license_checksum = 'b763dbe1b2fc6015d05cbf7bcc686412a2eb100a1f2220296e3b4a644c69633a'
 
     archive_path = join(dirname, filename)
 
diff --git a/src/python/test/test_remote_datasets.py b/src/python/test/test_remote_datasets.py
index 56a273b4..2057c63b 100644
--- a/src/python/test/test_remote_datasets.py
+++ b/src/python/test/test_remote_datasets.py
@@ -51,21 +51,21 @@ def _get_bunny_license_print(accept_license = False):
 def test_fetch_remote_datasets():
     # Test fetch with a wrong checksum
     with pytest.raises(OSError):
-        _check_fetch_output("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/spiral_2d.csv", "spiral_2d.csv", file_checksum = 'XXXXXXXXXX')
+        _check_fetch_output("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/spiral_2d/spiral_2d.npy", "spiral_2d.npy", file_checksum = 'XXXXXXXXXX')
 
     # Test files download from given urls with checksums provided
-    _check_fetch_output("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/spiral_2d.csv", "spiral_2d.csv",
-                                file_checksum = '37530355d980d957c4ec06b18c775f90a91e446107d06c6201c9b4000b077f38')
+    _check_fetch_output("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/spiral_2d/spiral_2d.npy", "spiral_2d.npy",
+                                file_checksum = '88312ffd6df2e2cb2bde9c0e1f962d7d644c6f58dc369c7b377b298dacdc4eaf')
 
     _check_fetch_output("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/sphere3D_pts_on_grid.off", "sphere3D_pts_on_grid.off",
                                 file_checksum = '32f96d2cafb1177f0dd5e0a019b6ff5658e14a619a7815ae55ad0fc5e8bd3f88')
 
     # Test files download from given urls without checksums
-    _check_fetch_output("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/spiral_2d.csv", "spiral_2d.csv")
+    _check_fetch_output("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/spiral_2d/spiral_2d.npy", "spiral_2d.npy")
 
     _check_fetch_output("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/sphere3D_pts_on_grid.off", "sphere3D_pts_on_grid.off")
 
-    # Test spiral_2d.csv wrapping function
+    # Test fetch_spiral_2d wrapping function
     spiral_2d_arr = remote.fetch_spiral_2d()
     assert spiral_2d_arr.shape == (114562, 2)
 
@@ -74,9 +74,9 @@ def test_fetch_remote_datasets():
     if not exists("remote_datasets/bunny"):
         makedirs("remote_datasets/bunny")
     remote._fetch_remote("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/bunny/LICENSE", "LICENSE", "remote_datasets/bunny",
-                 'aeb1bad319b7d74fa0b8076358182f9c6b1284c67cc07dc67cbc9bc73025d956')
+                 'b763dbe1b2fc6015d05cbf7bcc686412a2eb100a1f2220296e3b4a644c69633a')
     with open("remote_datasets/bunny/LICENSE") as f:
-        assert f.read() == _get_bunny_license_print().getvalue().rstrip("\n")
+        assert f.read().rstrip("\n") == _get_bunny_license_print().getvalue().rstrip("\n")
 
     # Test not printing bunny.npy LICENSE when accept_license = True
     assert "" == _get_bunny_license_print(accept_license = True).getvalue()
-- 
cgit v1.2.3


From 741f4f182479d1e5e78e9eb9180adce0a72e99b6 Mon Sep 17 00:00:00 2001
From: Hind-M <hind.montassif@gmail.com>
Date: Wed, 2 Feb 2022 10:38:15 +0100
Subject: Modify remote fetching test to increase its coverage

---
 src/python/test/test_remote_datasets.py | 21 +++++++--------------
 1 file changed, 7 insertions(+), 14 deletions(-)

(limited to 'src/python/test')

diff --git a/src/python/test/test_remote_datasets.py b/src/python/test/test_remote_datasets.py
index 2057c63b..dac9ee80 100644
--- a/src/python/test/test_remote_datasets.py
+++ b/src/python/test/test_remote_datasets.py
@@ -40,8 +40,6 @@ def _get_bunny_license_print(accept_license = False):
     # Redirect stdout
     sys.stdout = capturedOutput
 
-    if not exists("remote_datasets/bunny"):
-        makedirs("remote_datasets/bunny")
     remote._fetch_remote("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/bunny/bunny.npy", "bunny.npy", "remote_datasets/bunny",
                  '13f7842ebb4b45370e50641ff28c88685703efa5faab14edf0bb7d113a965e1b', accept_license)
     # Reset redirect
@@ -65,22 +63,17 @@ def test_fetch_remote_datasets():
 
     _check_fetch_output("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/sphere3D_pts_on_grid.off", "sphere3D_pts_on_grid.off")
 
-    # Test fetch_spiral_2d wrapping function
-    spiral_2d_arr = remote.fetch_spiral_2d()
-    assert spiral_2d_arr.shape == (114562, 2)
+    # Test fetch_spiral_2d and fetch_bunny wrapping functions (twice, to test case of already fetched files)
+    for i in range(2):
+        spiral_2d_arr = remote.fetch_spiral_2d()
+        assert spiral_2d_arr.shape == (114562, 2)
+
+        bunny_arr = remote.fetch_bunny()
+        assert bunny_arr.shape == (35947, 3)
 
     # Test printing existing LICENSE file when fetching bunny.npy with accept_license = False (default)
-    # Fetch LICENSE file
-    if not exists("remote_datasets/bunny"):
-        makedirs("remote_datasets/bunny")
-    remote._fetch_remote("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/bunny/LICENSE", "LICENSE", "remote_datasets/bunny",
-                 'b763dbe1b2fc6015d05cbf7bcc686412a2eb100a1f2220296e3b4a644c69633a')
     with open("remote_datasets/bunny/LICENSE") as f:
         assert f.read().rstrip("\n") == _get_bunny_license_print().getvalue().rstrip("\n")
 
     # Test not printing bunny.npy LICENSE when accept_license = True
     assert "" == _get_bunny_license_print(accept_license = True).getvalue()
-
-    # Test fetch_bunny wrapping function
-    bunny_arr = remote.fetch_bunny()
-    assert bunny_arr.shape == (35947, 3)
-- 
cgit v1.2.3


From 19689c712a1f5945e664f9c74c14b6994e7afaaf Mon Sep 17 00:00:00 2001
From: Hind-M <hind.montassif@gmail.com>
Date: Wed, 2 Feb 2022 16:14:17 +0100
Subject: Try to fix failing test in windows

---
 src/python/test/test_remote_datasets.py | 8 ++++++++
 1 file changed, 8 insertions(+)

(limited to 'src/python/test')

diff --git a/src/python/test/test_remote_datasets.py b/src/python/test/test_remote_datasets.py
index dac9ee80..643485f9 100644
--- a/src/python/test/test_remote_datasets.py
+++ b/src/python/test/test_remote_datasets.py
@@ -40,6 +40,9 @@ def _get_bunny_license_print(accept_license = False):
     # Redirect stdout
     sys.stdout = capturedOutput
 
+    if not exists("remote_datasets/bunny"):
+        makedirs("remote_datasets/bunny")
+
     remote._fetch_remote("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/bunny/bunny.npy", "bunny.npy", "remote_datasets/bunny",
                  '13f7842ebb4b45370e50641ff28c88685703efa5faab14edf0bb7d113a965e1b', accept_license)
     # Reset redirect
@@ -72,6 +75,11 @@ def test_fetch_remote_datasets():
         assert bunny_arr.shape == (35947, 3)
 
     # Test printing existing LICENSE file when fetching bunny.npy with accept_license = False (default)
+    # Fetch LICENSE file
+    if not exists("remote_datasets/bunny"):
+        makedirs("remote_datasets/bunny")
+    remote._fetch_remote("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/bunny/LICENSE", "LICENSE", "remote_datasets/bunny",
+                 'b763dbe1b2fc6015d05cbf7bcc686412a2eb100a1f2220296e3b4a644c69633a')
     with open("remote_datasets/bunny/LICENSE") as f:
         assert f.read().rstrip("\n") == _get_bunny_license_print().getvalue().rstrip("\n")
 
-- 
cgit v1.2.3


From a2d55f9bbf0f45e3ae4c147f734ce04f5bc87ab8 Mon Sep 17 00:00:00 2001
From: Hind-M <hind.montassif@gmail.com>
Date: Wed, 2 Feb 2022 21:32:55 +0100
Subject: Another attempt to fix windows failing test: move fetch_bunny to the
 end

---
 src/python/test/test_remote_datasets.py | 16 ++++++++--------
 1 file changed, 8 insertions(+), 8 deletions(-)

(limited to 'src/python/test')

diff --git a/src/python/test/test_remote_datasets.py b/src/python/test/test_remote_datasets.py
index 643485f9..5e607d73 100644
--- a/src/python/test/test_remote_datasets.py
+++ b/src/python/test/test_remote_datasets.py
@@ -66,14 +66,6 @@ def test_fetch_remote_datasets():
 
     _check_fetch_output("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/sphere3D_pts_on_grid.off", "sphere3D_pts_on_grid.off")
 
-    # Test fetch_spiral_2d and fetch_bunny wrapping functions (twice, to test case of already fetched files)
-    for i in range(2):
-        spiral_2d_arr = remote.fetch_spiral_2d()
-        assert spiral_2d_arr.shape == (114562, 2)
-
-        bunny_arr = remote.fetch_bunny()
-        assert bunny_arr.shape == (35947, 3)
-
     # Test printing existing LICENSE file when fetching bunny.npy with accept_license = False (default)
     # Fetch LICENSE file
     if not exists("remote_datasets/bunny"):
@@ -85,3 +77,11 @@ def test_fetch_remote_datasets():
 
     # Test not printing bunny.npy LICENSE when accept_license = True
     assert "" == _get_bunny_license_print(accept_license = True).getvalue()
+
+    # Test fetch_spiral_2d and fetch_bunny wrapping functions (twice, to test case of already fetched files)
+    for i in range(2):
+        spiral_2d_arr = remote.fetch_spiral_2d()
+        assert spiral_2d_arr.shape == (114562, 2)
+
+        bunny_arr = remote.fetch_bunny()
+        assert bunny_arr.shape == (35947, 3)
-- 
cgit v1.2.3


From 6109fd920ba477f89e83fea3df9803232c169463 Mon Sep 17 00:00:00 2001
From: Hind-M <hind.montassif@gmail.com>
Date: Thu, 3 Feb 2022 10:24:38 +0100
Subject: Remove archive folder before testing wrapping functions

---
 src/python/test/test_remote_datasets.py | 3 +++
 1 file changed, 3 insertions(+)

(limited to 'src/python/test')

diff --git a/src/python/test/test_remote_datasets.py b/src/python/test/test_remote_datasets.py
index 5e607d73..93a8a982 100644
--- a/src/python/test/test_remote_datasets.py
+++ b/src/python/test/test_remote_datasets.py
@@ -78,6 +78,9 @@ def test_fetch_remote_datasets():
     # Test not printing bunny.npy LICENSE when accept_license = True
     assert "" == _get_bunny_license_print(accept_license = True).getvalue()
 
+    # Remove "remote_datasets" directory and all its content
+    import shutil
+    shutil.rmtree("remote_datasets")
     # Test fetch_spiral_2d and fetch_bunny wrapping functions (twice, to test case of already fetched files)
     for i in range(2):
         spiral_2d_arr = remote.fetch_spiral_2d()
-- 
cgit v1.2.3


From a13282e4da9910a5d2bdadf97040095ae5b7880a Mon Sep 17 00:00:00 2001
From: Hind-M <hind.montassif@gmail.com>
Date: Fri, 4 Feb 2022 15:39:51 +0100
Subject: Store fetched datasets in user directory by default

---
 src/python/gudhi/datasets/remote.py     | 68 ++++++++++++++++++++++++++-------
 src/python/test/test_remote_datasets.py | 31 +++++++++++----
 2 files changed, 79 insertions(+), 20 deletions(-)

(limited to 'src/python/test')

diff --git a/src/python/gudhi/datasets/remote.py b/src/python/gudhi/datasets/remote.py
index 3498a645..3d6c01b0 100644
--- a/src/python/gudhi/datasets/remote.py
+++ b/src/python/gudhi/datasets/remote.py
@@ -7,14 +7,52 @@
 # Modification(s):
 #   - YYYY/MM Author: Description of the modification
 
-from os.path import join, exists
+from os.path import join, exists, expanduser
 from os import makedirs
 
 from urllib.request import urlretrieve
 import hashlib
+import shutil
 
 import numpy as np
 
+def get_data_home(data_home = None):
+    """
+    Return the path of the remote datasets directory.
+    This folder is used to store remotely fetched datasets.
+    By default the datasets directory is set to a folder named 'remote_datasets' in the user home folder.
+    Alternatively, it can be set by giving an explicit folder path. The '~' symbol is expanded to the user home folder.
+    If the folder does not already exist, it is automatically created.
+
+    Parameters
+    ----------
+    data_home : string
+        The path to remote datasets directory. Default is `None`, meaning that the data home directory will be set to "~/remote_datasets".
+
+    Returns
+    -------
+    data_home: string
+        The path to remote datasets directory.
+    """
+    if data_home is None:
+        data_home = join("~", "remote_datasets")
+    data_home = expanduser(data_home)
+    makedirs(data_home, exist_ok=True)
+    return data_home
+
+
+def clear_data_home(data_home = None):
+    """
+    Delete all the content of the data home cache.
+
+    Parameters
+    ----------
+    data_home : string, default is None.
+        The path to remote datasets directory. If `None`, the default directory to be removed is set to "~/remote_datasets".
+    """
+    data_home = get_data_home(data_home)
+    shutil.rmtree(data_home)
+
 def _checksum_sha256(file_path):
     """
     Compute the file checksum using sha256.
@@ -85,7 +123,7 @@ def _fetch_remote(url, filename, dirname = "remote_datasets", file_checksum = No
 
     return file_path
 
-def fetch_spiral_2d(filename = "spiral_2d.npy", dirname = "remote_datasets/spiral_2d"):
+def fetch_spiral_2d(filename = "spiral_2d.npy", dirname = None):
     """
     Fetch "spiral_2d.npy" remotely.
 
@@ -94,7 +132,7 @@ def fetch_spiral_2d(filename = "spiral_2d.npy", dirname = "remote_datasets/spira
     filename : string
         The name to give to downloaded file. Default is "spiral_2d.npy".
     dirname : string
-        The directory to save the file to. Default is "remote_datasets/spiral_2d".
+        The directory to save the file to. Default is None, meaning that the data home will be set to "~/remote_datasets/spiral_2d".
 
     Returns
     -------
@@ -104,20 +142,22 @@ def fetch_spiral_2d(filename = "spiral_2d.npy", dirname = "remote_datasets/spira
     file_url = "https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/spiral_2d/spiral_2d.npy"
     file_checksum = '88312ffd6df2e2cb2bde9c0e1f962d7d644c6f58dc369c7b377b298dacdc4eaf'
 
+    if dirname is None:
+        dirname = join(get_data_home(dirname), "spiral_2d")
+        makedirs(dirname, exist_ok=True)
+    else:
+        dirname = get_data_home(dirname)
+
     archive_path = join(dirname, filename)
 
     if not exists(archive_path):
-        # Create directory if not existing
-        if not exists(dirname):
-            makedirs(dirname)
-
         file_path_pkl = _fetch_remote(file_url, filename, dirname, file_checksum)
 
         return np.load(file_path_pkl, mmap_mode='r')
     else:
         return np.load(archive_path, mmap_mode='r')
 
-def fetch_bunny(filename = "bunny.npy", dirname = "remote_datasets/bunny", accept_license = False):
+def fetch_bunny(filename = "bunny.npy", dirname = None, accept_license = False):
     """
     Fetch "bunny.npy" remotely and its LICENSE file.
 
@@ -126,7 +166,7 @@ def fetch_bunny(filename = "bunny.npy", dirname = "remote_datasets/bunny", accep
     filename : string
         The name to give to downloaded file. Default is "bunny.npy".
     dirname : string
-        The directory to save the file to. Default is "remote_datasets/bunny".
+        The directory to save the file to. Default is None, meaning that the data home will be set to "~/remote_datasets/bunny".
     accept_license : boolean
         Flag to specify if user accepts the file LICENSE and prevents from printing the corresponding license terms.
         Default is False.
@@ -142,13 +182,15 @@ def fetch_bunny(filename = "bunny.npy", dirname = "remote_datasets/bunny", accep
     license_url = "https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/bunny/LICENSE"
     license_checksum = 'b763dbe1b2fc6015d05cbf7bcc686412a2eb100a1f2220296e3b4a644c69633a'
 
+    if dirname is None:
+        dirname = join(get_data_home(dirname), "bunny")
+        makedirs(dirname, exist_ok=True)
+    else:
+        dirname = get_data_home(dirname)
+
     archive_path = join(dirname, filename)
 
     if not exists(archive_path):
-        # Create directory if not existing
-        if not exists(dirname):
-            makedirs(dirname)
-
         license_path = _fetch_remote(license_url, "LICENSE", dirname, license_checksum)
         file_path_pkl = _fetch_remote(file_url, filename, dirname, file_checksum, accept_license)
 
diff --git a/src/python/test/test_remote_datasets.py b/src/python/test/test_remote_datasets.py
index 93a8a982..27eb51b0 100644
--- a/src/python/test/test_remote_datasets.py
+++ b/src/python/test/test_remote_datasets.py
@@ -10,7 +10,7 @@
 
 from gudhi.datasets import remote
 import re
-from os.path import isfile, exists
+from os.path import isfile, isdir, expanduser
 from os import makedirs
 import io
 import sys
@@ -30,8 +30,7 @@ def _check_dir_file_names(path_file_dw, filename, dirname):
         assert filename == names_dw[1]
 
 def _check_fetch_output(url, filename, dirname = "remote_datasets", file_checksum = None):
-    if not exists(dirname):
-        makedirs(dirname)
+    makedirs(dirname, exist_ok=True)
     path_file_dw = remote._fetch_remote(url, filename, dirname, file_checksum)
     _check_dir_file_names(path_file_dw, filename, dirname)
 
@@ -40,8 +39,7 @@ def _get_bunny_license_print(accept_license = False):
     # Redirect stdout
     sys.stdout = capturedOutput
 
-    if not exists("remote_datasets/bunny"):
-        makedirs("remote_datasets/bunny")
+    makedirs("remote_datasets/bunny", exist_ok=True)
 
     remote._fetch_remote("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/bunny/bunny.npy", "bunny.npy", "remote_datasets/bunny",
                  '13f7842ebb4b45370e50641ff28c88685703efa5faab14edf0bb7d113a965e1b', accept_license)
@@ -68,8 +66,7 @@ def test_fetch_remote_datasets():
 
     # Test printing existing LICENSE file when fetching bunny.npy with accept_license = False (default)
     # Fetch LICENSE file
-    if not exists("remote_datasets/bunny"):
-        makedirs("remote_datasets/bunny")
+    makedirs("remote_datasets/bunny", exist_ok=True)
     remote._fetch_remote("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/bunny/LICENSE", "LICENSE", "remote_datasets/bunny",
                  'b763dbe1b2fc6015d05cbf7bcc686412a2eb100a1f2220296e3b4a644c69633a')
     with open("remote_datasets/bunny/LICENSE") as f:
@@ -88,3 +85,23 @@ def test_fetch_remote_datasets():
 
         bunny_arr = remote.fetch_bunny()
         assert bunny_arr.shape == (35947, 3)
+
+    # Check that default dir was created
+    assert isdir(expanduser("~/remote_datasets")) == True
+
+    # Test clear_data_home
+    clear_data_home()
+    assert isdir(expanduser("~/remote_datasets")) == False
+
+    # Test fetch_spiral_2d and fetch_bunny wrapping functions with data directory different from default
+    spiral_2d_arr = remote.fetch_spiral_2d(dirname = "~/test")
+    assert spiral_2d_arr.shape == (114562, 2)
+
+    bunny_arr = remote.fetch_bunny(dirname = "~/test")
+    assert bunny_arr.shape == (35947, 3)
+
+    assert isdir(expanduser("~/test")) == True
+
+    # Test clear_data_home with data directory different from default
+    clear_data_home("~/test")
+    assert isdir(expanduser("~/test")) == False
-- 
cgit v1.2.3


From b0071de9ee7b6b4feb2eb9f19ceb759de21c997f Mon Sep 17 00:00:00 2001
From: Hind-M <hind.montassif@gmail.com>
Date: Fri, 4 Feb 2022 16:09:54 +0100
Subject: Add forgotten module name before func call

---
 src/python/test/test_remote_datasets.py | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

(limited to 'src/python/test')

diff --git a/src/python/test/test_remote_datasets.py b/src/python/test/test_remote_datasets.py
index 27eb51b0..9532b4ec 100644
--- a/src/python/test/test_remote_datasets.py
+++ b/src/python/test/test_remote_datasets.py
@@ -90,7 +90,7 @@ def test_fetch_remote_datasets():
     assert isdir(expanduser("~/remote_datasets")) == True
 
     # Test clear_data_home
-    clear_data_home()
+    remote.clear_data_home()
     assert isdir(expanduser("~/remote_datasets")) == False
 
     # Test fetch_spiral_2d and fetch_bunny wrapping functions with data directory different from default
@@ -103,5 +103,5 @@ def test_fetch_remote_datasets():
     assert isdir(expanduser("~/test")) == True
 
     # Test clear_data_home with data directory different from default
-    clear_data_home("~/test")
+    remote.clear_data_home("~/test")
     assert isdir(expanduser("~/test")) == False
-- 
cgit v1.2.3


From b5d7d6c2857d305ba2828065310c11edefb37c4e Mon Sep 17 00:00:00 2001
From: Hind-M <hind.montassif@gmail.com>
Date: Mon, 7 Feb 2022 13:03:45 +0100
Subject: Test get_data_home and clear_data_home on a separate folder

---
 src/python/test/test_remote_datasets.py | 21 ++++++++++-----------
 1 file changed, 10 insertions(+), 11 deletions(-)

(limited to 'src/python/test')

diff --git a/src/python/test/test_remote_datasets.py b/src/python/test/test_remote_datasets.py
index 9532b4ec..c160f270 100644
--- a/src/python/test/test_remote_datasets.py
+++ b/src/python/test/test_remote_datasets.py
@@ -87,21 +87,20 @@ def test_fetch_remote_datasets():
         assert bunny_arr.shape == (35947, 3)
 
     # Check that default dir was created
-    assert isdir(expanduser("~/remote_datasets")) == True
-
-    # Test clear_data_home
-    remote.clear_data_home()
-    assert isdir(expanduser("~/remote_datasets")) == False
+    assert isdir(expanduser("~/remote_datasets"))
 
     # Test fetch_spiral_2d and fetch_bunny wrapping functions with data directory different from default
-    spiral_2d_arr = remote.fetch_spiral_2d(dirname = "~/test")
+    spiral_2d_arr = remote.fetch_spiral_2d(dirname = "~/another_fetch_folder")
     assert spiral_2d_arr.shape == (114562, 2)
 
-    bunny_arr = remote.fetch_bunny(dirname = "~/test")
+    bunny_arr = remote.fetch_bunny(dirname = "~/another_fetch_folder")
     assert bunny_arr.shape == (35947, 3)
 
-    assert isdir(expanduser("~/test")) == True
+    assert isdir(expanduser("~/another_fetch_folder"))
+
+    # Test get_data_home and clear_data_home on new empty folder
+    empty_data_home = remote.get_data_home(data_home="empty_folder")
+    assert isdir(empty_data_home)
 
-    # Test clear_data_home with data directory different from default
-    remote.clear_data_home("~/test")
-    assert isdir(expanduser("~/test")) == False
+    remote.clear_data_home(data_home=empty_data_home)
+    assert not isdir(empty_data_home)
-- 
cgit v1.2.3


From e9b020adf11d48ce7a88932a5fe12cef011e72c9 Mon Sep 17 00:00:00 2001
From: Hind-M <hind.montassif@gmail.com>
Date: Mon, 7 Feb 2022 13:18:57 +0100
Subject: Separate tests into different functions and remove all test folders
 at the end

---
 src/python/test/test_remote_datasets.py | 19 +++++++++++++++----
 1 file changed, 15 insertions(+), 4 deletions(-)

(limited to 'src/python/test')

diff --git a/src/python/test/test_remote_datasets.py b/src/python/test/test_remote_datasets.py
index c160f270..2e595423 100644
--- a/src/python/test/test_remote_datasets.py
+++ b/src/python/test/test_remote_datasets.py
@@ -7,15 +7,17 @@
 # Modification(s):
 #   - YYYY/MM Author: Description of the modification
 
-
 from gudhi.datasets import remote
+
 import re
-from os.path import isfile, isdir, expanduser
-from os import makedirs
+import shutil
 import io
 import sys
 import pytest
 
+from os.path import isfile, isdir, expanduser
+from os import makedirs
+
 def _check_dir_file_names(path_file_dw, filename, dirname):
     assert isfile(path_file_dw)
 
@@ -76,8 +78,9 @@ def test_fetch_remote_datasets():
     assert "" == _get_bunny_license_print(accept_license = True).getvalue()
 
     # Remove "remote_datasets" directory and all its content
-    import shutil
     shutil.rmtree("remote_datasets")
+
+def test_fetch_remote_datasets_wrapped():
     # Test fetch_spiral_2d and fetch_bunny wrapping functions (twice, to test case of already fetched files)
     for i in range(2):
         spiral_2d_arr = remote.fetch_spiral_2d()
@@ -98,6 +101,14 @@ def test_fetch_remote_datasets():
 
     assert isdir(expanduser("~/another_fetch_folder"))
 
+    # Remove test folders
+    shutil.rmtree(expanduser("~/remote_datasets"))
+    shutil.rmtree(expanduser("~/another_fetch_folder"))
+
+    assert not isdir(expanduser("~/remote_datasets"))
+    assert not isdir(expanduser("~/another_fetch_folder"))
+
+def test_data_home():
     # Test get_data_home and clear_data_home on new empty folder
     empty_data_home = remote.get_data_home(data_home="empty_folder")
     assert isdir(empty_data_home)
-- 
cgit v1.2.3


From e964ec32247ce02fb12939cfcddaeabc04639869 Mon Sep 17 00:00:00 2001
From: Hind-M <hind.montassif@gmail.com>
Date: Mon, 7 Feb 2022 16:52:55 +0100
Subject: Del used variables before removing folders

---
 src/python/test/test_remote_datasets.py | 2 ++
 1 file changed, 2 insertions(+)

(limited to 'src/python/test')

diff --git a/src/python/test/test_remote_datasets.py b/src/python/test/test_remote_datasets.py
index 2e595423..cb53cb85 100644
--- a/src/python/test/test_remote_datasets.py
+++ b/src/python/test/test_remote_datasets.py
@@ -102,6 +102,8 @@ def test_fetch_remote_datasets_wrapped():
     assert isdir(expanduser("~/another_fetch_folder"))
 
     # Remove test folders
+    del spiral_2d_arr
+    del bunny_arr
     shutil.rmtree(expanduser("~/remote_datasets"))
     shutil.rmtree(expanduser("~/another_fetch_folder"))
 
-- 
cgit v1.2.3


From 5c0c731fdd2bc41c2a4833be1612dca5a082c337 Mon Sep 17 00:00:00 2001
From: Hind-M <hind.montassif@gmail.com>
Date: Wed, 2 Mar 2022 10:26:52 +0100
Subject: Modifications following PR review

---
 src/python/gudhi/datasets/remote.py     | 60 ++++++++++++++++++---------------
 src/python/test/test_remote_datasets.py | 38 ++++++++++-----------
 2 files changed, 51 insertions(+), 47 deletions(-)

(limited to 'src/python/test')

diff --git a/src/python/gudhi/datasets/remote.py b/src/python/gudhi/datasets/remote.py
index 3d6c01b0..618fa80e 100644
--- a/src/python/gudhi/datasets/remote.py
+++ b/src/python/gudhi/datasets/remote.py
@@ -20,14 +20,14 @@ def get_data_home(data_home = None):
     """
     Return the path of the remote datasets directory.
     This folder is used to store remotely fetched datasets.
-    By default the datasets directory is set to a folder named 'remote_datasets' in the user home folder.
+    By default the datasets directory is set to a folder named 'gudhi_data' in the user home folder.
     Alternatively, it can be set by giving an explicit folder path. The '~' symbol is expanded to the user home folder.
     If the folder does not already exist, it is automatically created.
 
     Parameters
     ----------
     data_home : string
-        The path to remote datasets directory. Default is `None`, meaning that the data home directory will be set to "~/remote_datasets".
+        The path to remote datasets directory. Default is `None`, meaning that the data home directory will be set to "~/gudhi_data".
 
     Returns
     -------
@@ -35,7 +35,7 @@ def get_data_home(data_home = None):
         The path to remote datasets directory.
     """
     if data_home is None:
-        data_home = join("~", "remote_datasets")
+        data_home = join("~", "gudhi_data")
     data_home = expanduser(data_home)
     makedirs(data_home, exist_ok=True)
     return data_home
@@ -43,12 +43,12 @@ def get_data_home(data_home = None):
 
 def clear_data_home(data_home = None):
     """
-    Delete all the content of the data home cache.
+    Delete the data home cache directory and all its content.
 
     Parameters
     ----------
     data_home : string, default is None.
-        The path to remote datasets directory. If `None`, the default directory to be removed is set to "~/remote_datasets".
+        The path to remote datasets directory. If `None`, the default directory to be removed is set to "~/gudhi_data".
     """
     data_home = get_data_home(data_home)
     shutil.rmtree(data_home)
@@ -77,7 +77,7 @@ def _checksum_sha256(file_path):
             sha256_hash.update(buffer)
     return sha256_hash.hexdigest()
 
-def _fetch_remote(url, filename, dirname = "remote_datasets", file_checksum = None, accept_license = False):
+def _fetch_remote(url, filename, dirname = "gudhi_data", file_checksum = None, accept_license = False):
     """
     Fetch the wanted dataset from the given url and save it in file_path.
 
@@ -88,10 +88,10 @@ def _fetch_remote(url, filename, dirname = "remote_datasets", file_checksum = No
     filename : string
         The name to give to downloaded file.
     dirname : string
-        The directory to save the file to. Default is "remote_datasets".
+        The directory to save the file to. Default is "gudhi_data".
     file_checksum : string
         The file checksum using sha256 to check against the one computed on the downloaded file.
-        Default is 'None'.
+        Default is 'None', which means the checksum is not checked.
     accept_license : boolean
         Flag to specify if user accepts the file LICENSE and prevents from printing the corresponding license terms.
         Default is False.
@@ -100,6 +100,11 @@ def _fetch_remote(url, filename, dirname = "remote_datasets", file_checksum = No
     -------
     file_path: string
         Full path of the created file.
+
+    Raises
+    ------
+    IOError
+        If the computed SHA256 checksum of file does not match the one given by the user.
     """
 
     file_path = join(dirname, filename)
@@ -123,32 +128,37 @@ def _fetch_remote(url, filename, dirname = "remote_datasets", file_checksum = No
 
     return file_path
 
+def _get_archive_and_dir(dirname, filename, label):
+    if dirname is None:
+        dirname = join(get_data_home(dirname), label)
+        makedirs(dirname, exist_ok=True)
+    else:
+        dirname = get_data_home(dirname)
+
+    archive_path = join(dirname, filename)
+
+    return archive_path, dirname
+
 def fetch_spiral_2d(filename = "spiral_2d.npy", dirname = None):
     """
-    Fetch "spiral_2d.npy" remotely.
+    Fetch spiral_2d dataset remotely.
 
     Parameters
     ----------
     filename : string
         The name to give to downloaded file. Default is "spiral_2d.npy".
     dirname : string
-        The directory to save the file to. Default is None, meaning that the data home will be set to "~/remote_datasets/spiral_2d".
+        The directory to save the file to. Default is None, meaning that the data home will be set to "~/gudhi_data/spiral_2d".
 
     Returns
     -------
     points: array
-        Array of points stored in "spiral_2d.npy".
+        Array of points.
     """
     file_url = "https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/spiral_2d/spiral_2d.npy"
     file_checksum = '88312ffd6df2e2cb2bde9c0e1f962d7d644c6f58dc369c7b377b298dacdc4eaf'
 
-    if dirname is None:
-        dirname = join(get_data_home(dirname), "spiral_2d")
-        makedirs(dirname, exist_ok=True)
-    else:
-        dirname = get_data_home(dirname)
-
-    archive_path = join(dirname, filename)
+    archive_path, dirname = _get_archive_and_dir(dirname, filename, "spiral_2d")
 
     if not exists(archive_path):
         file_path_pkl = _fetch_remote(file_url, filename, dirname, file_checksum)
@@ -159,14 +169,14 @@ def fetch_spiral_2d(filename = "spiral_2d.npy", dirname = None):
 
 def fetch_bunny(filename = "bunny.npy", dirname = None, accept_license = False):
     """
-    Fetch "bunny.npy" remotely and its LICENSE file.
+    Fetch Stanford bunny dataset remotely and its LICENSE file.
 
     Parameters
     ----------
     filename : string
         The name to give to downloaded file. Default is "bunny.npy".
     dirname : string
-        The directory to save the file to. Default is None, meaning that the data home will be set to "~/remote_datasets/bunny".
+        The directory to save the file to. Default is None, meaning that the data home will be set to "~/gudhi_data/bunny".
     accept_license : boolean
         Flag to specify if user accepts the file LICENSE and prevents from printing the corresponding license terms.
         Default is False.
@@ -174,7 +184,7 @@ def fetch_bunny(filename = "bunny.npy", dirname = None, accept_license = False):
     Returns
     -------
     points: array
-        Array of points stored in "bunny.npy".
+        Array of points.
     """
 
     file_url = "https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/bunny/bunny.npy"
@@ -182,13 +192,7 @@ def fetch_bunny(filename = "bunny.npy", dirname = None, accept_license = False):
     license_url = "https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/bunny/LICENSE"
     license_checksum = 'b763dbe1b2fc6015d05cbf7bcc686412a2eb100a1f2220296e3b4a644c69633a'
 
-    if dirname is None:
-        dirname = join(get_data_home(dirname), "bunny")
-        makedirs(dirname, exist_ok=True)
-    else:
-        dirname = get_data_home(dirname)
-
-    archive_path = join(dirname, filename)
+    archive_path, dirname = _get_archive_and_dir(dirname, filename, "bunny")
 
     if not exists(archive_path):
         license_path = _fetch_remote(license_url, "LICENSE", dirname, license_checksum)
diff --git a/src/python/test/test_remote_datasets.py b/src/python/test/test_remote_datasets.py
index cb53cb85..c44ac22b 100644
--- a/src/python/test/test_remote_datasets.py
+++ b/src/python/test/test_remote_datasets.py
@@ -22,7 +22,7 @@ def _check_dir_file_names(path_file_dw, filename, dirname):
     assert isfile(path_file_dw)
 
     names_dw = re.split(r' |/|\\', path_file_dw)
-    # Case where inner directories are created in "remote_datasets/"; e.g: "remote_datasets/bunny"
+    # Case where inner directories are created in "test_gudhi_data/"; e.g: "test_gudhi_data/bunny"
     if len(names_dw) >= 3:
         for i in range(len(names_dw)-1):
             assert re.split(r' |/|\\', dirname)[i] == names_dw[i]
@@ -31,7 +31,7 @@ def _check_dir_file_names(path_file_dw, filename, dirname):
         assert dirname == names_dw[0]
         assert filename == names_dw[1]
 
-def _check_fetch_output(url, filename, dirname = "remote_datasets", file_checksum = None):
+def _check_fetch_output(url, filename, dirname = "test_gudhi_data", file_checksum = None):
     makedirs(dirname, exist_ok=True)
     path_file_dw = remote._fetch_remote(url, filename, dirname, file_checksum)
     _check_dir_file_names(path_file_dw, filename, dirname)
@@ -41,9 +41,9 @@ def _get_bunny_license_print(accept_license = False):
     # Redirect stdout
     sys.stdout = capturedOutput
 
-    makedirs("remote_datasets/bunny", exist_ok=True)
+    makedirs("test_gudhi_data/bunny", exist_ok=True)
 
-    remote._fetch_remote("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/bunny/bunny.npy", "bunny.npy", "remote_datasets/bunny",
+    remote._fetch_remote("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/bunny/bunny.npy", "bunny.npy", "test_gudhi_data/bunny",
                  '13f7842ebb4b45370e50641ff28c88685703efa5faab14edf0bb7d113a965e1b', accept_license)
     # Reset redirect
     sys.stdout = sys.__stdout__
@@ -68,19 +68,21 @@ def test_fetch_remote_datasets():
 
     # Test printing existing LICENSE file when fetching bunny.npy with accept_license = False (default)
     # Fetch LICENSE file
-    makedirs("remote_datasets/bunny", exist_ok=True)
-    remote._fetch_remote("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/bunny/LICENSE", "LICENSE", "remote_datasets/bunny",
+    makedirs("test_gudhi_data/bunny", exist_ok=True)
+    remote._fetch_remote("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/bunny/LICENSE", "LICENSE", "test_gudhi_data/bunny",
                  'b763dbe1b2fc6015d05cbf7bcc686412a2eb100a1f2220296e3b4a644c69633a')
-    with open("remote_datasets/bunny/LICENSE") as f:
+    with open("test_gudhi_data/bunny/LICENSE") as f:
         assert f.read().rstrip("\n") == _get_bunny_license_print().getvalue().rstrip("\n")
 
     # Test not printing bunny.npy LICENSE when accept_license = True
     assert "" == _get_bunny_license_print(accept_license = True).getvalue()
 
-    # Remove "remote_datasets" directory and all its content
-    shutil.rmtree("remote_datasets")
+    # Remove "test_gudhi_data" directory and all its content
+    shutil.rmtree("test_gudhi_data")
 
 def test_fetch_remote_datasets_wrapped():
+    # Check if gudhi_data default dir exists already
+    to_be_removed = not isdir(expanduser("~/gudhi_data"))
     # Test fetch_spiral_2d and fetch_bunny wrapping functions (twice, to test case of already fetched files)
     for i in range(2):
         spiral_2d_arr = remote.fetch_spiral_2d()
@@ -90,29 +92,27 @@ def test_fetch_remote_datasets_wrapped():
         assert bunny_arr.shape == (35947, 3)
 
     # Check that default dir was created
-    assert isdir(expanduser("~/remote_datasets"))
+    assert isdir(expanduser("~/gudhi_data"))
 
     # Test fetch_spiral_2d and fetch_bunny wrapping functions with data directory different from default
-    spiral_2d_arr = remote.fetch_spiral_2d(dirname = "~/another_fetch_folder")
+    spiral_2d_arr = remote.fetch_spiral_2d(dirname = "./another_fetch_folder_for_test")
     assert spiral_2d_arr.shape == (114562, 2)
 
-    bunny_arr = remote.fetch_bunny(dirname = "~/another_fetch_folder")
+    bunny_arr = remote.fetch_bunny(dirname = "./another_fetch_folder_for_test")
     assert bunny_arr.shape == (35947, 3)
 
-    assert isdir(expanduser("~/another_fetch_folder"))
+    assert isdir(expanduser("./another_fetch_folder_for_test"))
 
     # Remove test folders
     del spiral_2d_arr
     del bunny_arr
-    shutil.rmtree(expanduser("~/remote_datasets"))
-    shutil.rmtree(expanduser("~/another_fetch_folder"))
-
-    assert not isdir(expanduser("~/remote_datasets"))
-    assert not isdir(expanduser("~/another_fetch_folder"))
+    if to_be_removed:
+        shutil.rmtree(expanduser("~/gudhi_data"))
+    shutil.rmtree(expanduser("./another_fetch_folder_for_test"))
 
 def test_data_home():
     # Test get_data_home and clear_data_home on new empty folder
-    empty_data_home = remote.get_data_home(data_home="empty_folder")
+    empty_data_home = remote.get_data_home(data_home="empty_folder_for_test")
     assert isdir(empty_data_home)
 
     remote.clear_data_home(data_home=empty_data_home)
-- 
cgit v1.2.3


From ef8284cce27a8f11947e7f076034aa2fd8b5a395 Mon Sep 17 00:00:00 2001
From: Hind-M <hind.montassif@gmail.com>
Date: Wed, 4 May 2022 15:27:34 +0200
Subject: Ask for file_path as parameter of remote fetching functions instead
 of both dirname and filename Modify remote fetching test

---
 src/python/gudhi/datasets/remote.py     | 106 +++++++++++++++-----------------
 src/python/test/test_remote_datasets.py |  94 ++++++++++------------------
 2 files changed, 83 insertions(+), 117 deletions(-)

(limited to 'src/python/test')

diff --git a/src/python/gudhi/datasets/remote.py b/src/python/gudhi/datasets/remote.py
index 8b3baef4..5b535911 100644
--- a/src/python/gudhi/datasets/remote.py
+++ b/src/python/gudhi/datasets/remote.py
@@ -7,7 +7,7 @@
 # Modification(s):
 #   - YYYY/MM Author: Description of the modification
 
-from os.path import join, exists, expanduser
+from os.path import join, split, exists, expanduser
 from os import makedirs, remove
 
 from urllib.request import urlretrieve
@@ -60,7 +60,7 @@ def _checksum_sha256(file_path):
     Parameters
     ----------
     file_path: string
-        Full path of the created file.
+        Full path of the created file including filename.
 
     Returns
     -------
@@ -77,7 +77,7 @@ def _checksum_sha256(file_path):
             sha256_hash.update(buffer)
     return sha256_hash.hexdigest()
 
-def _fetch_remote(url, filename, dirname, file_checksum = None, accept_license = False):
+def _fetch_remote(url, file_path, file_checksum = None):
     """
     Fetch the wanted dataset from the given url and save it in file_path.
 
@@ -85,21 +85,11 @@ def _fetch_remote(url, filename, dirname, file_checksum = None, accept_license =
     ----------
     url : string
         The url to fetch the dataset from.
-    filename : string
-        The name to give to downloaded file.
-    dirname : string
-        The directory to save the file to.
+    file_path : string
+        Full path of the downloaded file including filename.
     file_checksum : string
         The file checksum using sha256 to check against the one computed on the downloaded file.
         Default is 'None', which means the checksum is not checked.
-    accept_license : boolean
-        Flag to specify if user accepts the file LICENSE and prevents from printing the corresponding license terms.
-        Default is False.
-
-    Returns
-    -------
-    file_path: string
-        Full path of the created file.
 
     Raises
     ------
@@ -107,8 +97,6 @@ def _fetch_remote(url, filename, dirname, file_checksum = None, accept_license =
         If the computed SHA256 checksum of file does not match the one given by the user.
     """
 
-    file_path = join(dirname, filename)
-
     # Get the file
     urlretrieve(url, file_path)
 
@@ -121,36 +109,41 @@ def _fetch_remote(url, filename, dirname, file_checksum = None, accept_license =
                         "different from expected : {}."
                         "The file may be corrupted or the given url may be wrong !".format(file_path, checksum, file_checksum))
 
-    # Print license terms unless accept_license is set to True
-    if not accept_license:
-        license_file = join(dirname, "LICENSE")
-        if exists(license_file) and (file_path != license_file):
-            with open(license_file, 'r') as f:
-                print(f.read())
+def _get_archive_path(file_path, label):
+    """
+    Get archive path based on file_path given by user and label.
 
-    return file_path
+    Parameters
+    ----------
+    file_path: string
+        Full path of the file to get including filename, or None.
+    label: string
+        Label used along with 'data_home' to get archive path, in case 'file_path' is None.
 
-def _get_archive_and_dir(dirname, filename, label):
-    if dirname is None:
-        dirname = join(get_data_home(dirname), label)
+    Returns
+    -------
+        Full path of archive including filename.
+    """
+    if file_path is None:
+        archive_path = join(get_data_home(), label)
+        dirname = split(archive_path)[0]
         makedirs(dirname, exist_ok=True)
     else:
-        dirname = get_data_home(dirname)
-
-    archive_path = join(dirname, filename)
+        archive_path = file_path
+        dirname = split(archive_path)[0]
+        makedirs(dirname, exist_ok=True)
 
-    return archive_path, dirname
+    return archive_path
 
-def fetch_spiral_2d(filename = "spiral_2d.npy", dirname = None):
+def fetch_spiral_2d(file_path = None):
     """
     Fetch spiral_2d dataset remotely.
 
     Parameters
     ----------
-    filename : string
-        The name to give to downloaded file. Default is "spiral_2d.npy".
-    dirname : string
-        The directory to save the file to. Default is None, meaning that the downloaded file will be put in "~/gudhi_data/points/spiral_2d".
+    file_path : string
+        Full path of the downloaded file including filename.
+        Default is None, meaning that it's set to "data_home/points/spiral_2d/spiral_2d.npy".
 
     Returns
     -------
@@ -158,28 +151,25 @@ def fetch_spiral_2d(filename = "spiral_2d.npy", dirname = None):
         Array of shape (114562, 2).
     """
     file_url = "https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/spiral_2d/spiral_2d.npy"
-    file_checksum = '88312ffd6df2e2cb2bde9c0e1f962d7d644c6f58dc369c7b377b298dacdc4eaf'
+    file_checksum = '2226024da76c073dd2f24b884baefbfd14928b52296df41ad2d9b9dc170f2401'
 
-    archive_path, dirname = _get_archive_and_dir(dirname, filename, "points/spiral_2d")
+    archive_path = _get_archive_path(file_path, "points/spiral_2d/spiral_2d.npy")
 
     if not exists(archive_path):
-        file_path_pkl = _fetch_remote(file_url, filename, dirname, file_checksum)
+        _fetch_remote(file_url, archive_path, file_checksum)
 
-        return np.load(file_path_pkl, mmap_mode='r')
-    else:
-        return np.load(archive_path, mmap_mode='r')
+    return np.load(archive_path, mmap_mode='r')
 
-def fetch_bunny(filename = "bunny.npy", dirname = None, accept_license = False):
+def fetch_bunny(file_path = None, accept_license = False):
     """
     Fetch Stanford bunny dataset remotely and its LICENSE file.
     This dataset contains 35947 vertices.
 
     Parameters
     ----------
-    filename : string
-        The name to give to downloaded file. Default is "bunny.npy".
-    dirname : string
-        The directory to save the file to. Default is None, meaning that the downloaded files will be put in "~/gudhi_data/points/bunny".
+    file_path : string
+        Full path of the downloaded file including filename.
+        Default is None, meaning that it's set to "data_home/points/bunny/bunny.npy".
     accept_license : boolean
         Flag to specify if user accepts the file LICENSE and prevents from printing the corresponding license terms.
         Default is False.
@@ -191,16 +181,20 @@ def fetch_bunny(filename = "bunny.npy", dirname = None, accept_license = False):
     """
 
     file_url = "https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/bunny/bunny.npy"
-    file_checksum = '13f7842ebb4b45370e50641ff28c88685703efa5faab14edf0bb7d113a965e1b'
-    license_url = "https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/bunny/LICENSE"
+    file_checksum = 'f382482fd89df8d6444152dc8fd454444fe597581b193fd139725a85af4a6c6e'
+    license_url = "https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/bunny/bunny.LICENSE"
     license_checksum = 'b763dbe1b2fc6015d05cbf7bcc686412a2eb100a1f2220296e3b4a644c69633a'
 
-    archive_path, dirname = _get_archive_and_dir(dirname, filename, "points/bunny")
+    archive_path = _get_archive_path(file_path, "points/bunny/bunny.npy")
 
     if not exists(archive_path):
-        license_path = _fetch_remote(license_url, "LICENSE", dirname, license_checksum)
-        file_path_pkl = _fetch_remote(file_url, filename, dirname, file_checksum, accept_license)
-
-        return np.load(file_path_pkl, mmap_mode='r')
-    else:
-        return np.load(archive_path, mmap_mode='r')
+        _fetch_remote(file_url, archive_path, file_checksum)
+        license_path = join(split(archive_path)[0], "bunny.LICENSE")
+        _fetch_remote(license_url, license_path, license_checksum)
+        # Print license terms unless accept_license is set to True
+        if not accept_license:
+            if exists(license_path):
+                with open(license_path, 'r') as f:
+                    print(f.read())
+
+    return np.load(archive_path, mmap_mode='r')
diff --git a/src/python/test/test_remote_datasets.py b/src/python/test/test_remote_datasets.py
index c44ac22b..5d0d397d 100644
--- a/src/python/test/test_remote_datasets.py
+++ b/src/python/test/test_remote_datasets.py
@@ -9,76 +9,48 @@
 
 from gudhi.datasets import remote
 
-import re
 import shutil
 import io
 import sys
 import pytest
 
-from os.path import isfile, isdir, expanduser
-from os import makedirs
+from os.path import isdir, expanduser, exists
+from os import remove
 
-def _check_dir_file_names(path_file_dw, filename, dirname):
-    assert isfile(path_file_dw)
+def test_data_home():
+    # Test get_data_home and clear_data_home on new empty folder
+    empty_data_home = remote.get_data_home(data_home="empty_folder_for_test")
+    assert isdir(empty_data_home)
 
-    names_dw = re.split(r' |/|\\', path_file_dw)
-    # Case where inner directories are created in "test_gudhi_data/"; e.g: "test_gudhi_data/bunny"
-    if len(names_dw) >= 3:
-        for i in range(len(names_dw)-1):
-            assert re.split(r' |/|\\', dirname)[i] == names_dw[i]
-        assert filename == names_dw[i+1]
-    else:
-        assert dirname == names_dw[0]
-        assert filename == names_dw[1]
+    remote.clear_data_home(data_home=empty_data_home)
+    assert not isdir(empty_data_home)
 
-def _check_fetch_output(url, filename, dirname = "test_gudhi_data", file_checksum = None):
-    makedirs(dirname, exist_ok=True)
-    path_file_dw = remote._fetch_remote(url, filename, dirname, file_checksum)
-    _check_dir_file_names(path_file_dw, filename, dirname)
+def test_fetch_remote():
+    # Test fetch with a wrong checksum
+    with pytest.raises(OSError):
+        remote._fetch_remote("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/spiral_2d/spiral_2d.npy", "tmp_spiral_2d.npy", file_checksum = 'XXXXXXXXXX')
+    assert not exists("tmp_spiral_2d.npy")
 
 def _get_bunny_license_print(accept_license = False):
     capturedOutput = io.StringIO()
     # Redirect stdout
     sys.stdout = capturedOutput
 
-    makedirs("test_gudhi_data/bunny", exist_ok=True)
+    bunny_arr = remote.fetch_bunny("./tmp_for_test/bunny.npy", accept_license)
+    assert bunny_arr.shape == (35947, 3)
+    remove("./tmp_for_test/bunny.npy")
 
-    remote._fetch_remote("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/bunny/bunny.npy", "bunny.npy", "test_gudhi_data/bunny",
-                 '13f7842ebb4b45370e50641ff28c88685703efa5faab14edf0bb7d113a965e1b', accept_license)
     # Reset redirect
     sys.stdout = sys.__stdout__
     return capturedOutput
 
-def test_fetch_remote_datasets():
-    # Test fetch with a wrong checksum
-    with pytest.raises(OSError):
-        _check_fetch_output("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/spiral_2d/spiral_2d.npy", "spiral_2d.npy", file_checksum = 'XXXXXXXXXX')
-
-    # Test files download from given urls with checksums provided
-    _check_fetch_output("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/spiral_2d/spiral_2d.npy", "spiral_2d.npy",
-                                file_checksum = '88312ffd6df2e2cb2bde9c0e1f962d7d644c6f58dc369c7b377b298dacdc4eaf')
-
-    _check_fetch_output("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/sphere3D_pts_on_grid.off", "sphere3D_pts_on_grid.off",
-                                file_checksum = '32f96d2cafb1177f0dd5e0a019b6ff5658e14a619a7815ae55ad0fc5e8bd3f88')
-
-    # Test files download from given urls without checksums
-    _check_fetch_output("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/spiral_2d/spiral_2d.npy", "spiral_2d.npy")
-
-    _check_fetch_output("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/sphere3D_pts_on_grid.off", "sphere3D_pts_on_grid.off")
-
-    # Test printing existing LICENSE file when fetching bunny.npy with accept_license = False (default)
-    # Fetch LICENSE file
-    makedirs("test_gudhi_data/bunny", exist_ok=True)
-    remote._fetch_remote("https://raw.githubusercontent.com/GUDHI/gudhi-data/main/points/bunny/LICENSE", "LICENSE", "test_gudhi_data/bunny",
-                 'b763dbe1b2fc6015d05cbf7bcc686412a2eb100a1f2220296e3b4a644c69633a')
-    with open("test_gudhi_data/bunny/LICENSE") as f:
-        assert f.read().rstrip("\n") == _get_bunny_license_print().getvalue().rstrip("\n")
-
+def test_print_bunny_license():
     # Test not printing bunny.npy LICENSE when accept_license = True
     assert "" == _get_bunny_license_print(accept_license = True).getvalue()
-
-    # Remove "test_gudhi_data" directory and all its content
-    shutil.rmtree("test_gudhi_data")
+    # Test printing bunny.LICENSE file when fetching bunny.npy with accept_license = False (default)
+    with open("./tmp_for_test/bunny.LICENSE") as f:
+        assert f.read().rstrip("\n") == _get_bunny_license_print().getvalue().rstrip("\n")
+    shutil.rmtree("./tmp_for_test")
 
 def test_fetch_remote_datasets_wrapped():
     # Check if gudhi_data default dir exists already
@@ -93,27 +65,27 @@ def test_fetch_remote_datasets_wrapped():
 
     # Check that default dir was created
     assert isdir(expanduser("~/gudhi_data"))
+    # Check downloaded files
+    assert exists(expanduser("~/gudhi_data/points/spiral_2d/spiral_2d.npy"))
+    assert exists(expanduser("~/gudhi_data/points/bunny/bunny.npy"))
+    assert exists(expanduser("~/gudhi_data/points/bunny/bunny.LICENSE"))
 
     # Test fetch_spiral_2d and fetch_bunny wrapping functions with data directory different from default
-    spiral_2d_arr = remote.fetch_spiral_2d(dirname = "./another_fetch_folder_for_test")
+    spiral_2d_arr = remote.fetch_spiral_2d("./another_fetch_folder_for_test/spiral_2d.npy")
     assert spiral_2d_arr.shape == (114562, 2)
 
-    bunny_arr = remote.fetch_bunny(dirname = "./another_fetch_folder_for_test")
+    bunny_arr = remote.fetch_bunny("./another_fetch_folder_for_test/bunny.npy")
     assert bunny_arr.shape == (35947, 3)
 
-    assert isdir(expanduser("./another_fetch_folder_for_test"))
+    assert isdir("./another_fetch_folder_for_test")
+    # Check downloaded files
+    assert exists("./another_fetch_folder_for_test/spiral_2d.npy")
+    assert exists("./another_fetch_folder_for_test/bunny.npy")
+    assert exists("./another_fetch_folder_for_test/bunny.LICENSE")
 
     # Remove test folders
     del spiral_2d_arr
     del bunny_arr
     if to_be_removed:
         shutil.rmtree(expanduser("~/gudhi_data"))
-    shutil.rmtree(expanduser("./another_fetch_folder_for_test"))
-
-def test_data_home():
-    # Test get_data_home and clear_data_home on new empty folder
-    empty_data_home = remote.get_data_home(data_home="empty_folder_for_test")
-    assert isdir(empty_data_home)
-
-    remote.clear_data_home(data_home=empty_data_home)
-    assert not isdir(empty_data_home)
+    shutil.rmtree("./another_fetch_folder_for_test")
-- 
cgit v1.2.3


From 52d5b524403a43bfdc0b27a7feeec04e9c9c34c2 Mon Sep 17 00:00:00 2001
From: Hind-M <hind.montassif@gmail.com>
Date: Thu, 5 May 2022 17:43:12 +0200
Subject: Add GUDHI_DATA environment variable option

---
 src/python/gudhi/datasets/remote.py     | 16 +++++++++++-----
 src/python/test/test_remote_datasets.py | 13 ++++++++++++-
 2 files changed, 23 insertions(+), 6 deletions(-)

(limited to 'src/python/test')

diff --git a/src/python/gudhi/datasets/remote.py b/src/python/gudhi/datasets/remote.py
index 5b535911..eac8caf3 100644
--- a/src/python/gudhi/datasets/remote.py
+++ b/src/python/gudhi/datasets/remote.py
@@ -8,7 +8,7 @@
 #   - YYYY/MM Author: Description of the modification
 
 from os.path import join, split, exists, expanduser
-from os import makedirs, remove
+from os import makedirs, remove, environ
 
 from urllib.request import urlretrieve
 import hashlib
@@ -21,13 +21,16 @@ def get_data_home(data_home = None):
     Return the path of the remote datasets directory.
     This folder is used to store remotely fetched datasets.
     By default the datasets directory is set to a folder named 'gudhi_data' in the user home folder.
-    Alternatively, it can be set by giving an explicit folder path. The '~' symbol is expanded to the user home folder.
+    Alternatively, it can be set by the 'GUDHI_DATA' environment variable.
+    The '~' symbol is expanded to the user home folder.
     If the folder does not already exist, it is automatically created.
 
     Parameters
     ----------
     data_home : string
-        The path to remote datasets directory. Default is `None`, meaning that the data home directory will be set to "~/gudhi_data".
+        The path to remote datasets directory.
+        Default is `None`, meaning that the data home directory will be set to "~/gudhi_data",
+        if the 'GUDHI_DATA' environment variable does not exist.
 
     Returns
     -------
@@ -35,7 +38,7 @@ def get_data_home(data_home = None):
         The path to remote datasets directory.
     """
     if data_home is None:
-        data_home = join("~", "gudhi_data")
+        data_home = environ.get("GUDHI_DATA", join("~", "gudhi_data"))
     data_home = expanduser(data_home)
     makedirs(data_home, exist_ok=True)
     return data_home
@@ -48,7 +51,9 @@ def clear_data_home(data_home = None):
     Parameters
     ----------
     data_home : string, default is None.
-        The path to remote datasets directory. If `None`, the default directory to be removed is set to "~/gudhi_data".
+        The path to remote datasets directory.
+        If `None` and the 'GUDHI_DATA' environment variable does not exist,
+        the default directory to be removed is set to "~/gudhi_data".
     """
     data_home = get_data_home(data_home)
     shutil.rmtree(data_home)
@@ -170,6 +175,7 @@ def fetch_bunny(file_path = None, accept_license = False):
     file_path : string
         Full path of the downloaded file including filename.
         Default is None, meaning that it's set to "data_home/points/bunny/bunny.npy".
+        In this case, the LICENSE file would be downloaded as "data_home/points/bunny/bunny.LICENSE".
     accept_license : boolean
         Flag to specify if user accepts the file LICENSE and prevents from printing the corresponding license terms.
         Default is False.
diff --git a/src/python/test/test_remote_datasets.py b/src/python/test/test_remote_datasets.py
index 5d0d397d..af26d77c 100644
--- a/src/python/test/test_remote_datasets.py
+++ b/src/python/test/test_remote_datasets.py
@@ -15,7 +15,7 @@ import sys
 import pytest
 
 from os.path import isdir, expanduser, exists
-from os import remove
+from os import remove, environ
 
 def test_data_home():
     # Test get_data_home and clear_data_home on new empty folder
@@ -89,3 +89,14 @@ def test_fetch_remote_datasets_wrapped():
     if to_be_removed:
         shutil.rmtree(expanduser("~/gudhi_data"))
     shutil.rmtree("./another_fetch_folder_for_test")
+
+def test_gudhi_data_env():
+    # Set environment variable "GUDHI_DATA"
+    environ["GUDHI_DATA"] = "./test_folder_from_env_var"
+    bunny_arr = remote.fetch_bunny()
+    assert bunny_arr.shape == (35947, 3)
+    assert exists("./test_folder_from_env_var/points/bunny/bunny.npy")
+    assert exists("./test_folder_from_env_var/points/bunny/bunny.LICENSE")
+    # Remove test folder
+    del bunny_arr
+    shutil.rmtree("./test_folder_from_env_var")
-- 
cgit v1.2.3


From f344700ebee65de9ccc8799f2ec4e1c633ab864e Mon Sep 17 00:00:00 2001
From: Hind-M <hind.montassif@gmail.com>
Date: Thu, 5 May 2022 18:07:52 +0200
Subject: Remove default data home test (because of 'GUDHI_DATA' environment
 variable option)

---
 src/python/test/test_remote_datasets.py | 26 +++++---------------------
 1 file changed, 5 insertions(+), 21 deletions(-)

(limited to 'src/python/test')

diff --git a/src/python/test/test_remote_datasets.py b/src/python/test/test_remote_datasets.py
index af26d77c..6f569fd2 100644
--- a/src/python/test/test_remote_datasets.py
+++ b/src/python/test/test_remote_datasets.py
@@ -53,30 +53,16 @@ def test_print_bunny_license():
     shutil.rmtree("./tmp_for_test")
 
 def test_fetch_remote_datasets_wrapped():
-    # Check if gudhi_data default dir exists already
-    to_be_removed = not isdir(expanduser("~/gudhi_data"))
-    # Test fetch_spiral_2d and fetch_bunny wrapping functions (twice, to test case of already fetched files)
+    # Test fetch_spiral_2d and fetch_bunny wrapping functions with data directory different from default (twice, to test case of already fetched files)
+    # Default case is not tested because it would fail in case the user sets the 'GUDHI_DATA' environment variable locally
     for i in range(2):
-        spiral_2d_arr = remote.fetch_spiral_2d()
+        spiral_2d_arr = remote.fetch_spiral_2d("./another_fetch_folder_for_test/spiral_2d.npy")
         assert spiral_2d_arr.shape == (114562, 2)
 
-        bunny_arr = remote.fetch_bunny()
+        bunny_arr = remote.fetch_bunny("./another_fetch_folder_for_test/bunny.npy")
         assert bunny_arr.shape == (35947, 3)
 
-    # Check that default dir was created
-    assert isdir(expanduser("~/gudhi_data"))
-    # Check downloaded files
-    assert exists(expanduser("~/gudhi_data/points/spiral_2d/spiral_2d.npy"))
-    assert exists(expanduser("~/gudhi_data/points/bunny/bunny.npy"))
-    assert exists(expanduser("~/gudhi_data/points/bunny/bunny.LICENSE"))
-
-    # Test fetch_spiral_2d and fetch_bunny wrapping functions with data directory different from default
-    spiral_2d_arr = remote.fetch_spiral_2d("./another_fetch_folder_for_test/spiral_2d.npy")
-    assert spiral_2d_arr.shape == (114562, 2)
-
-    bunny_arr = remote.fetch_bunny("./another_fetch_folder_for_test/bunny.npy")
-    assert bunny_arr.shape == (35947, 3)
-
+    # Check that the directory was created
     assert isdir("./another_fetch_folder_for_test")
     # Check downloaded files
     assert exists("./another_fetch_folder_for_test/spiral_2d.npy")
@@ -86,8 +72,6 @@ def test_fetch_remote_datasets_wrapped():
     # Remove test folders
     del spiral_2d_arr
     del bunny_arr
-    if to_be_removed:
-        shutil.rmtree(expanduser("~/gudhi_data"))
     shutil.rmtree("./another_fetch_folder_for_test")
 
 def test_gudhi_data_env():
-- 
cgit v1.2.3


From a809771b6d7381d233656f7a0b02211559189bfe Mon Sep 17 00:00:00 2001
From: Hind-M <hind.montassif@gmail.com>
Date: Fri, 6 May 2022 09:52:26 +0200
Subject: Delete bunny array before removing the file

---
 src/python/test/test_remote_datasets.py | 1 +
 1 file changed, 1 insertion(+)

(limited to 'src/python/test')

diff --git a/src/python/test/test_remote_datasets.py b/src/python/test/test_remote_datasets.py
index 6f569fd2..cde9fa22 100644
--- a/src/python/test/test_remote_datasets.py
+++ b/src/python/test/test_remote_datasets.py
@@ -38,6 +38,7 @@ def _get_bunny_license_print(accept_license = False):
 
     bunny_arr = remote.fetch_bunny("./tmp_for_test/bunny.npy", accept_license)
     assert bunny_arr.shape == (35947, 3)
+    del bunny_arr
     remove("./tmp_for_test/bunny.npy")
 
     # Reset redirect
-- 
cgit v1.2.3


From ce34ee3e5c28c48d605f23332cfa3c10e471a047 Mon Sep 17 00:00:00 2001
From: Hind-M <hind.montassif@gmail.com>
Date: Tue, 24 May 2022 15:57:52 +0200
Subject: Make get_data_home function private

---
 src/python/doc/datasets.rst             | 2 --
 src/python/gudhi/datasets/remote.py     | 6 +++---
 src/python/test/test_remote_datasets.py | 4 ++--
 3 files changed, 5 insertions(+), 7 deletions(-)

(limited to 'src/python/test')

diff --git a/src/python/doc/datasets.rst b/src/python/doc/datasets.rst
index d2975533..8b0912c4 100644
--- a/src/python/doc/datasets.rst
+++ b/src/python/doc/datasets.rst
@@ -127,6 +127,4 @@ We provide some ready-to-use datasets that are not available by default when get
 
      2D spiral with 114562 vertices.
 
-.. autofunction:: gudhi.datasets.remote.get_data_home
-
 .. autofunction:: gudhi.datasets.remote.clear_data_home
diff --git a/src/python/gudhi/datasets/remote.py b/src/python/gudhi/datasets/remote.py
index d2ae2a75..7e6f647f 100644
--- a/src/python/gudhi/datasets/remote.py
+++ b/src/python/gudhi/datasets/remote.py
@@ -16,7 +16,7 @@ import shutil
 
 import numpy as np
 
-def get_data_home(data_home = None):
+def _get_data_home(data_home = None):
     """
     Return the path of the remote datasets directory.
     This folder is used to store remotely fetched datasets.
@@ -55,7 +55,7 @@ def clear_data_home(data_home = None):
         If `None` and the 'GUDHI_DATA' environment variable does not exist,
         the default directory to be removed is set to "~/gudhi_data".
     """
-    data_home = get_data_home(data_home)
+    data_home = _get_data_home(data_home)
     shutil.rmtree(data_home)
 
 def _checksum_sha256(file_path):
@@ -130,7 +130,7 @@ def _get_archive_path(file_path, label):
         Full path of archive including filename.
     """
     if file_path is None:
-        archive_path = join(get_data_home(), label)
+        archive_path = join(_get_data_home(), label)
         dirname = split(archive_path)[0]
         makedirs(dirname, exist_ok=True)
     else:
diff --git a/src/python/test/test_remote_datasets.py b/src/python/test/test_remote_datasets.py
index cde9fa22..e5d2de82 100644
--- a/src/python/test/test_remote_datasets.py
+++ b/src/python/test/test_remote_datasets.py
@@ -18,8 +18,8 @@ from os.path import isdir, expanduser, exists
 from os import remove, environ
 
 def test_data_home():
-    # Test get_data_home and clear_data_home on new empty folder
-    empty_data_home = remote.get_data_home(data_home="empty_folder_for_test")
+    # Test _get_data_home and clear_data_home on new empty folder
+    empty_data_home = remote._get_data_home(data_home="empty_folder_for_test")
     assert isdir(empty_data_home)
 
     remote.clear_data_home(data_home=empty_data_home)
-- 
cgit v1.2.3