fixing data_directory not actually changing download directory and en…

…forced spacy and thinc version requirements (#229)

Co-authored-by: Ethan Xia <[email protected]>
Co-authored-by: seanzhangkx8 <[email protected]>

convokit/util.py

@@ -6,7 +6,7 @@
 import warnings
 import zipfile
 from typing import Dict
-
+from .convokitConfig import ConvoKitConfig
 import requests
 
 
@@ -108,15 +108,16 @@ def download(
 
     custom_data_dir = data_dir
 
-    data_dir = os.path.expanduser("~/.convokit/")
-
+    config = ConvoKitConfig()
+    data_dir = config.data_directory
+    data_dir = os.path.expanduser(data_dir)
     # pkg_resources.resource_filename("convokit", "")
     if not os.path.exists(data_dir):
         os.mkdir(data_dir)
     if not os.path.exists(os.path.join(data_dir, "downloads")):
         os.mkdir(os.path.join(data_dir, "downloads"))
 
-    dataset_path = os.path.join(data_dir, "downloads", name)
+    dataset_path = os.path.join(data_dir, name)
 
     if custom_data_dir is not None:
         dataset_path = os.path.join(custom_data_dir, name)
@@ -192,7 +193,8 @@ def download_local(name: str, data_dir: str):
     :return: string path to local Corpus
     """
     custom_data_dir = data_dir
-    data_dir = os.path.expanduser("~/.convokit/")
+    config = ConvoKitConfig()
+    data_dir = config.data_directory
 
     # pkg_resources.resource_filename("convokit", "")
     if not os.path.exists(data_dir):

docs/source/troubleshooting.rst

@@ -12,7 +12,10 @@ Issues
 
 **Error associated with Numpy**
 
-ConvoKit currently requires Numpy 1.x.x, as Numpy 2.x is known to cause compatibility issues. Please verify your Numpy version. We are working on supporting Numpy 2.x and appreciate your understanding.
+Pre Spacy 3.8.2 is not compatible with numpy 2.0.0+ due to compatibility issues with thinc. Spacy 3.8.2 is compatible with numpy 2.0.0+ but currently requires thinc to be >=8.3.0, <8.4.0, so as a temporary solution ConvoKit now enforces spacy>=3.8.2, thinc >=8.3.0, <8.4.0. We will continue to keep an eye on spacy releases and update the requirements if there are new releases targeting this issue.
+For additional insight into the issue:
+`spaCy issue #13528 <https://github.com/explosion/spaCy/issues/13528>`_
+`thinc issue #939 <https://github.com/explosion/thinc/issues/939>`_
 
 -----------------------------
 

setup.py

@@ -44,7 +44,7 @@
         "matplotlib>=3.0.0",
         "pandas>=0.23.4",
         "msgpack-numpy>=0.4.3.2",
-        "spacy>=2.3.5",
+        "spacy>=3.8.2",
         "scipy>=1.1.0",
         "scikit-learn>=0.20.0",
         "nltk>=3.4",
@@ -56,6 +56,7 @@
         "pymongo>=4.0",
         "pyyaml>=5.4.1",
         "dnspython>=1.16.0",
+        "thinc>=8.3.0,<8.4.0",
     ],
     extras_require={
         "craft": ["torch>=0.12"],