Implementação da SM no Python 3

• Baixa versão do nlpnet

mkdir /srv/lbsociam-py2/terceiros
cd /srv/lbsociam-py2/terceiros
git clone https://github.com/eduardosan/nlpnet.git

• Executa o comando do python

cd /srv/lbsociam-py2/src/LBSociam
/srv/lbsociam-py2/bin/python setup.py develop

• Configura nltk

/srv/lbsociam-py2/bin/python
>>> import nltk
>>> nltk.download()
---------------------------------------------------------------------------
    d) Download   l) List    u) Update   c) Config   h) Help   q) Quit
---------------------------------------------------------------------------
Downloader> c
---------------------------------------------------------------------------
    s) Show Config   u) Set Server URL   d) Set Data Dir   m) Main Menu
---------------------------------------------------------------------------
Config> d
  New Directory> /srv/lbsociam-py2/nltk_data
---------------------------------------------------------------------------
    s) Show Config   u) Set Server URL   d) Set Data Dir   m) Main Menu
---------------------------------------------------------------------------
Config> m

---------------------------------------------------------------------------
    d) Download   l) List    u) Update   c) Config   h) Help   q) Quit
---------------------------------------------------------------------------
Downloader> d

Download which package (l=list; x=cancel)?
  Identifier> punkt
    Downloading package punkt to /srv/lbsociam/nltk_data...
      Unzipping tokenizers/punkt.zip.

---------------------------------------------------------------------------
    d) Download   l) List    u) Update   c) Config   h) Help   q) Quit
---------------------------------------------------------------------------
Downloader> q
True
>>> quit()

• Configura o diretório do nlpnet

mkdir /srv/lbsociam-py2/nlpnet_data
/srv/lbsociam-py2/bin/python
>>> import nlpnet
>>> nlpnet.set_data_dir('/srv/lbsociam-py2/nlpnet_data')

• Baixa os modelos

cd /srv/lbsociam-py2/nlpnet_data
wget http://nilc.icmc.usp.br/nlpnet/nlpnet-srl.zip
unzip nlpnet-srl.zip
rm nlpnet-srl.zip
mv srl/* .
rm -r srl

Roda o teste pra ver se está funcionando

export NLTK_DATA=/srv/lbsociam-py2/nltk_data
/srv/lbsociam-py2/bin/python
>>> import nlpnet
>>> nlpnet.set_data_dir('/srv/lbsociam-py2/nlpnet_data')
>>> tagger = nlpnet.SRLTagger()
>>> sent = tagger.tag(u'O rato roeu a roupa do rei de Roma.')[0]  
>>> sent.tokens
>>> sent.arg_structures

• Baixar o script de configuração

wget https://github.com/bear/python-twitter/raw/master/get_access_token.py
cd /srv/lbsociam-py2/src/LBSociam
/srv/lbsociam-py2/bin/pip install oauth2
/srv/lbsociam-py2/bin/python get_access_token.py

• Importa dados do twitter