blob: 02a8410bfc6983b8958e1a77046e2cea62b12956 (
plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
|
# ragequitte
Tech to evaluate:
* UI
* https://tabler.github.io/
* NLP
* https://github.com/orsinium/textdistance
* http://www.nltk.org/
* https://textblob.readthedocs.io/en/dev/index.html
* https://spacy.io/
* other
* https://prodi.gy/
SETUP
---
* Create a virtual Python environment `virtualenv venv -p python3`
* Enter it `source venv/bin/activate`
* Install dependencies `pip install -r requirements.txt`
* Set up a Redis server
* Copy and modify `config_example.py`
USAGE
---
* Enter the virtual environment
* Start the Discord log collector (optional): `PYTHONPATH="." python grabbers/discord_to_redis.py`
* Train a word2vec model: `PYTHONPATH="." python processors/redis_to_word2vec.py`
* Try the model, run an interactive Python shell with `python`:
* `>>> import gensim`
* `>>> import config`
* `>>> model = gensim.models.Word2Vec.load(config.model)`
* `>>> model.wv.most_similar_cosmul("what", topn=4)`
* `[('where', 0.8449756503105164), ('how', 0.8136355876922607), ('whatever', 0.8117746114730835), ('why', 0.8039281368255615)]`
* `>>> model.wv.similarity("league", "dota")`
* `0.6141897432617985`
* more examples here: https://radimrehurek.com/gensim/models/keyedvectors.html
|