Sentiment Analysis

[Castau]

This Pull Request is the solution to User Story #3

---

For Review

Check out this commit if you want to run this module during the review - it contains Rúnis preprocessing. Make sure you have everything from the setup file installed and downloadet. Then you should be able to just run it.

This is how each tweet ends up looking after the Sentiment Analysis:

---

Tasks

• [x] See guides and tutorials to better understand how to do a Sentiment Analysis
  • [x] Tutorials and guides (Digital Ocean, NLTK, Stack Abuse Scikit-Learn)
• [x] Get test data in the form that it will be received from the preprocessing
• [x] Get training data, for now use what is supplied in the Digital Ocean guide.
  • [ ] We might need other training data if we change the way the tweets come in
• [x] Rewrite the code from the guide to fit this project or write it from scratch in another file
• [x] Get an idea of the format for the output
• [x] Make sure this module works with the preprocessing module
• [x] Let model be trainable from outside

---

The result of the first edition looks like this:

[
  {
    'tweet': 'This tweet year', 
    'hashtags': [
      '#MyFirstTweet'
      ], 
    'people': [
      '@folketinget'
      ], 
    'urls': [
      'runivn.dk'
      ], 
    'author': 'Runi Vedel', 
    'date': '01/05/2020', 
    'sentiment_analysis': [
      {
        'verdict': 'Positive', 
        'positive_procent': 0.61, 
        'negative_procent': 0.39
      }
    ]
  }, 
  {
    'tweet': ......
  }
]

[Castau]

Nice and clean. :)

Would be cool if you could gitignore __pycache__, should be local only

Sure - I'll just fix that :)