Bug: Capability to read METADATA from non pypi sources

[PhorstenkampFuzzy]

Before You Begin

Before proceeding, please make sure to follow these steps:

• I have checked for similar issues in the project's issue tracker.
• I have searched closed issues to see if a similar problem was reported
  before.

Issue Details

No license/authro/homepage detection from non pypi packages such as:

https://spacy.io/models/de#de_core_news_sm

Here a section from the METADATA file

Metadata-Version: 2.1
Name: de_core_news_sm
Version: 3.8.0
Summary: German pipeline optimized for CPU. Components: tok2vec, tagger, morphologizer, parser, lemmatizer (trainable_lemmatizer), senter, ner.
Home-page: https://explosion.ai
Author: Explosion
Author-email: [email protected]
License: MIT
License-File: LICENSE
License-File: LICENSES_SOURCES