开源bot工具Rasa学习---1

Rasa是一个不错的开源bot工具，全部基于python实现，主页是https://rasa-nlu.readthedocs.io/en/latest/index.html

本节是关于工具安装和初步运行的记录。

根据其文档进行安装，我选择的安装方式是：

git clone git@github.com:RasaHQ/rasa_nlu.git

cd rasa_nlu

pip install -r requirements.txt

python setup.py install

后端我选择的是“spaCy + sklearn”， 其中提示我需要安装sklearn-crfsuite。

然后当我安装这个工具时，又是各种与当前环境不匹配或找不到可用版本，最后通过以下方式解决了。

conda install -c conda-forge python-crfsuite=0.9.2

最后是在Ipython中针对自带的例子“ata\examples\rasa\demo-rasa.json”进行模型训练：

In [2]: run rasa_nlu/train.py -c config_spacy.json

INFO:rasa_nlu.utils.spacy_utils:Trying to load spacy model with name 'en'

INFO:rasa_nlu.components:Added 'nlp_spacy' to component cache. Key 'nlp_spacy-en'.

INFO:rasa_nlu.converters:Training data format at ./data/examples/rasa/demo-rasa.json is rasa_nlu

INFO:rasa_nlu.training_data:Training data stats:

- intent examples: 40 (4 distinct intents)

- found intents: 'affirm', 'goodbye', 'greet', 'restaurant_search'

- entity examples: 9 (2 distinct entities)

- found entities: 'cuisine', 'location'
INFO:rasa_nlu.model:Starting to train component nlp_spacy

INFO:rasa_nlu.model:Finished training component.

INFO:rasa_nlu.model:Starting to train component ner_crf

INFO:rasa_nlu.model:Finished training component.

INFO:rasa_nlu.model:Starting to train component ner_synonyms

INFO:rasa_nlu.model:Finished training component.

INFO:rasa_nlu.model:Starting to train component intent_featurizer_spacy

INFO:rasa_nlu.model:Finished training component.

INFO:rasa_nlu.model:Starting to train component intent_classifier_sklearn

Fitting 2 folds for each of 6 candidates, totalling 12 fits

[Parallel(n_jobs=1)]: Done 12 out of 12 | elapsed: 0.0s finished

INFO:rasa_nlu.model:Finished training component.

INFO:rasa_nlu.model:Successfully saved model into 'D:\Work\NLP\Task\Rasa\workspace\models\model_20170709-153926'

INFO:__main__:Finished training

针对以上得到的模型，进行预测的代码如下：

from rasa_nlu.model import Metadata, Interpreter

from rasa_nlu.config import RasaNLUConfig
metadata = Metadata.load('models/model_20170709-153926') # where model_directory points to the folder the model is persisted in

interpreter = Interpreter.load(metadata, RasaNLUConfig("config_spacy.json"))
interpreter.parse(u"I am looking for Chinese food")

Out[9]:

{'entities': [],

'intent': {'confidence': 0.7729154931115908, 'name': 'restaurant_search'},

'intent_ranking': [{'confidence': 0.7729154931115908,

'name': 'restaurant_search'},

{'confidence': 0.093570000080283558, 'name': 'goodbye'},

{'confidence': 0.093241162858677284, 'name': 'greet'},

{'confidence': 0.040273343949448648, 'name': 'affirm'}],

'text': 'I am looking for Chinese food'}