cannot download en_ewt_model.zip in running demo/pipeline_demo.py after cloning on my local

[hayata-yamamoto]

Hi,
I have one trouble on demo script. After installing stanfordnlp from pip and cloning your repository on my local, I cannot run script demo/pipeline_demo.py to try your code.

Raised Error is below. Could you tell me solution of this problem.

Downloading models for: en_ewt
Download location: /Users/hayata.yamamoto/stanfordnlp_resources/en_ewt_models.zip
Traceback (most recent call last):
  File "demo/pipeline_demo.py", line 32, in <module>
    stanfordnlp.download(args.lang, args.models_dir, confirm_if_exists=True)
  File "/Users/hayata.yamamoto/anaconda3/envs/analytics/lib/python3.6/site-packages/stanfordnlp/utils/resources.py", line 134, in download
    download_ud_model(default_treebanks[download_label], resource_dir=resource_dir, confirm_if_exists=confirm_if_exists)
  File "/Users/hayata.yamamoto/anaconda3/envs/analytics/lib/python3.6/site-packages/stanfordnlp/utils/resources.py", line 101, in download_ud_model
    with open(download_file_path, 'wb') as f:
FileNotFoundError: [Errno 2] No such file or directory: '/Users/hayata.yamamoto/stanfordnlp_resources/en_ewt_models.zip'

[sebschu]

You'll either have to create the directory stanfordnlp_resources in your home directory (this is the default directory for models) or specify a target directory for the models using the -d command line argument of pipeline_demo.py.

[yuhaozhang]

This is indeed a caveat of the download module. One temporary solution is, as @sebschu pointed out, to create the target model directory manually (mkdir ${model_dir} in bash) before you run the demo script. We'll fix this issue quickly with a v0.1.0.post1 release.

[hayata-yamamoto]

@sebschu @yuhaozhang
Thanks to your quick answers, I could download!

We'll fix this issue quickly with a v0.1.0.post1 release.

Splendid! I look forward next update!
Thanks a lot! :)

[manning]

Can't it just be a 0.1.1 release, no?

[yuhaozhang]

@manning Yes. This is already in a v0.1.1 release on PyPI.