aboSamoor / polyglot

Multilingual text (NLP) processing toolkit
http://polyglot-nlp.com
Other
2.3k stars 337 forks source link

downloading models from behind an enterprise proxy, or manually installing them from disk #149

Open matanox opened 6 years ago

matanox commented 6 years ago

Hi,

I would like to install polyglot behind an enterprise firewall. In my case, like most enterprise cases, this would require disabling SSL in the tool/command that performs the download, because SSL is proxied by an on-prem proxy service for security reasons. The on-prem SSL proxy will verify the SSL certificate of the download servers, but the end-node on which polyglot is installed will fail to validate the certificate if it attempts to, unless configured not to try to.

This is typically solved in one of two ways:

  1. Bringing in the models on a flash drive, hacking them into the right places and manually configuring the tool/library to use them.
  2. Enabling proxy configuration and/or SSL de-configuration. Proxy configuration: configuring such that the download module/tool makes its requests to the on-prem proxy rather than try the regular Internet addresses. SSL relaxation: turning Certificate verification off or turning SSL altogether off.

The second option was just added to spaCy, as an example, but this is also enabled for things like pip install and such, it is needed for virtually any installer.

Are any of these already possible for polyglot?

Many thanks!!!