728x90
A guide to creating a chatbot with Rasa stack and Python.
2편에서는 앞서 만든 봇을 슬랙에 배포하는 실습을 진행해본다. 진행에 앞서 1편을 미리 읽어보고 넘어오길 권해드린다.
Rasa Stack 과 파이썬을 활용한 슬랙 챗봇 만들기 (1) A guide to creating a chatbot with Rasa stack and Python
Rasa Installations
1편과는 달리 여기서는 최신 버전의 Rasa Core 를 설치할 것이다. 아나콘다 등을 활용하여 가상환경을 만든 뒤 실습을 진행하길 권해드린다.
우분투 18.04 아나콘다 설치하기 How To Install the Anaconda Python Distribution on Ubuntu 18.04
아나콘다 가상환경을 생성한 후 활성화시킨다.
(base) founder@hilbert:~$ conda create --name rasabot python=3.6
Collecting package metadata: done
Solving environment: done
==> WARNING: A newer version of conda exists. <==
current version: 4.6.3
latest version: 4.6.4
Please update conda by running
$ conda update -n base -c defaults conda
## Package Plan ##
environment location: /home/founder/anaconda3/envs/rasabot
added / updated specs:
- python=3.6
The following packages will be downloaded:
package | build
---------------------------|-----------------
libedit-3.1.20181209 | hc058e9b_0 188 KB
pip-19.0.1 | py36_0 1.8 MB
python-3.6.8 | h0371630_0 34.4 MB
setuptools-40.8.0 | py36_0 647 KB
------------------------------------------------------------
Total: 37.1 MB
The following NEW packages will be INSTALLED:
ca-certificates pkgs/main/linux-64::ca-certificates-2019.1.23-0
certifi pkgs/main/linux-64::certifi-2018.11.29-py36_0
libedit pkgs/main/linux-64::libedit-3.1.20181209-hc058e9b_0
libffi pkgs/main/linux-64::libffi-3.2.1-hd88cf55_4
libgcc-ng pkgs/main/linux-64::libgcc-ng-8.2.0-hdf63c60_1
libstdcxx-ng pkgs/main/linux-64::libstdcxx-ng-8.2.0-hdf63c60_1
ncurses pkgs/main/linux-64::ncurses-6.1-he6710b0_1
openssl pkgs/main/linux-64::openssl-1.1.1a-h7b6447c_0
pip pkgs/main/linux-64::pip-19.0.1-py36_0
python pkgs/main/linux-64::python-3.6.8-h0371630_0
readline pkgs/main/linux-64::readline-7.0-h7b6447c_5
setuptools pkgs/main/linux-64::setuptools-40.8.0-py36_0
sqlite pkgs/main/linux-64::sqlite-3.26.0-h7b6447c_0
tk pkgs/main/linux-64::tk-8.6.8-hbc83047_0
wheel pkgs/main/linux-64::wheel-0.32.3-py36_0
xz pkgs/main/linux-64::xz-5.2.4-h14c3975_4
zlib pkgs/main/linux-64::zlib-1.2.11-h7b6447c_3
Proceed ([y]/n)? y
Downloading and Extracting Packages
libedit-3.1.20181209 | 188 KB | ##################################### | 100%
python-3.6.8 | 34.4 MB | ##################################### | 100%
setuptools-40.8.0 | 647 KB | ##################################### | 100%
pip-19.0.1 | 1.8 MB | ##################################### | 100%
Preparing transaction: done
Verifying transaction: done
Executing transaction: done
#
# To activate this environment, use
#
# $ conda activate rasabot
#
# To deactivate an active environment, use
#
# $ conda deactivate
(base) founder@hilbert:~$
(base) founder@hilbert:~$ source activate rasabot
(rasabot) founder@hilbert:~$(rasabot) founder@hilbert:~$ mkdir rasabot
(rasabot) founder@hilbert:~$ cd rasabot
(rasabot) founder@hilbert:~/rasabot$
최신 Rasa stack 을 설치한다.
Rasa NLU (https://rasa.com/docs/nlu/installation/)
(rasabot) founder@hilbert:~/rasabot$ python -m pip install rasa_nlu[spacy]
(rasabot) founder@hilbert:~/rasabot$ pip show rasa_nlu
Name: rasa-nlu
Version: 0.14.3
Summary: Rasa NLU a natural language parser for bots
Home-page: https://rasa.com
Author: Rasa Technologies GmbH
Author-email: hi@rasa.com
License: Apache 2.0
Location: /home/founder/anaconda3/envs/rasabot/lib/python3.6/site-packages
Requires: six, gevent, numpy, matplotlib, coloredlogs, requests, klein, future, typing, ruamel.yaml, cloudpickle, simplejson, jsonschema, tqdm, scikit-learn, boto3, packaging
Required-by:
Rasa Core (https://rasa.com/docs/core/installation/)
(rasabot) founder@hilbert:~/rasabot$ python -m pip install -U rasa_core(rasabot) founder@hilbert:~/rasabot$ pip show rasa_core
Name: rasa-core
Version: 0.13.2
Summary: Machine learning based dialogue engine for conversational software.
Home-page: https://rasa.com
Author: Rasa Technologies GmbH
Author-email: hi@rasa.com
License: Apache 2.0
Location: /home/founder/anaconda3/envs/rasabot/lib/python3.6/site-packages
Requires: pydot, slackclient, tensorflow, flask-cors, questionary, python-socketio, keras-preprocessing, rocketchat-API, scipy, jsonpickle, tqdm, python-dateutil, rasa-core-sdk, pykwalify, ruamel.yaml, colorhash, fakeredis, apscheduler, pymongo, flask, typing, redis, numpy, scikit-learn, requests, packaging, webexteamssdk, coloredlogs, networkx, mattermostwrapper, twilio, fbmessenger, pytz, rasa-nlu, jsonschema, pika, terminaltables, keras-applications, gevent, python-telegram-bot, flask-jwt-simple, colorclass
Required-by:(rasabot) founder@hilbert:~/rasabot$ python -m spacy download en_core_web_md
Collecting en_core_web_md==2.0.0 from https://github.com/explosion/spacy-models/releases/download/en_core_web_md-2.0.0/en_core_web_md-2.0.0.tar.gz#egg=en_core_web_md==2.0.0
Downloading https://github.com/explosion/spacy-models/releases/download/en_core_web_md-2.0.0/en_core_web_md-2.0.0.tar.gz (120.8MB)
100% |████████████████████████████████| 120.9MB 129.7MB/s
Installing collected packages: en-core-web-md
Running setup.py install for en-core-web-md ... done
Successfully installed en-core-web-md-2.0.0
Linking successful
/home/founder/anaconda3/envs/rasabot/lib/python3.6/site-packages/en_core_web_md
-->
/home/founder/anaconda3/envs/rasabot/lib/python3.6/site-packages/spacy/data/en_core_web_md
You can now load the model via spacy.load('en_core_web_md')
(rasabot) founder@hilbert:~/rasabot$ python -m spacy link en_core_web_md en --force;
Linking successful
/home/founder/anaconda3/envs/rasabot/lib/python3.6/site-packages/en_core_web_md
-->
/home/founder/anaconda3/envs/rasabot/lib/python3.6/site-packages/spacy/data/en
You can now load the model via spacy.load('en')앞선 생성한 아래 디렉토리에 복제했다.
/home/founder/rasabot
Settings
slack 설정
슬랙 앱을 생성을 통해 챗봇을 슬랙에 통합해보자. 슬랙앱 생성부터 진행하자.
1) https://api.slack.com/ 로그인 Workspace 및 Channels 생성했다고 가정하고 이하 진행한다. https://api.slack.com 으로 이동 Start Building 를 클릭한다.
2) 앱을 생성 앱 이름 및 소속될 Workspace 를 지정한다.
3) 좌측 메뉴 Features > Bot Users > Add a Bot User 클릭
4) 봇의 이름을 정하고 항상 온라인 상태로 설정 후, Add Bot User 클릭
5) Settings > Basic Information > Display Information 수정 (앱아이콘, 설명 등 수정) (그림 5)
6) Settings > Basic Information > install this app into our workplace > Install App to Workspace > Authorize
ngrok 설정
Ngrok 은 멀티플랫폼 터널링, 리버스 프록시 소프트웨어로 인터넷에서 로컬 구동 네트워크 서비스로의 안전한 터널링을 생성해준다. 간단히 말해서 인터넷에서 로컬 앱에 접근할 수 있도록 해준다는 것이다.
1) ngrok 사이트에서 로그인 후 적절한 파일을 다운로드받는다.
2) $ unzip /path/to/ngrok.zip 을 통해 파일압축을 해제한다.
3) $ ./ngrok <authtoken> 형식으로 계정에 접속한다.
4) ngrok 압축을 푼 디렉토리로 이동하여 $ ngrok <authtoken> 이라고 콘솔에 입력한다. 토큰 정보는 여기서 확인할 수 있다.
(rasabot) founder@hilbert:~/src$ ./ngrok authtoken 4xWEiEVWgkoJnSfyLEDgd_2bMEHLSdzQsa2dFU4LzDh
Authtoken saved to configuration file: /home/founder/.ngrok2/ngrok.yml
5) 다음과 같이 포트 번호를 지정해 서비스를 시작한다.
(rasabot) founder@hilbert:~/src$ ./ngrok http 5004
Session Status online
Account 홍 길 동 (Plan: Free)
Version 2.2.8
Region United States (us)
Web Interface http://127.0.0.1:4040
Forwarding http://dhidjid78.ngrok.io -> localhost:5004
Forwarding https://dhidjid78.ngrok.io -> localhost:5004
Connections ttl opn rt1 rt5 p50 p90
0 0 0.00 0.00 0.00 0.00
Deploying the Bot on Slack
Create a Python Script
Since we are done with all the requirements, it’s time to deploy our bot. For this, we will need to write a Python script called run_app.py, which will integrate our chatbot with the slack app that we created above. We will begin by creating a slack connector for our Rasa chatbot. We will use RasaNLU interpreter to load the NLU model directly from the python script.
We will train our model again to make sure everything is good and running.
Training the NLU Model
(rasabot) founder@hilbert:~/rasabot$ python nlu_model.py
Fitting 2 folds for each of 6 candidates, totalling 12 fits
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[Parallel(n_jobs=1)]: Done 12 out of 12 | elapsed: 0.2s finished
{'intent': {'name': 'mood_unhappy', 'confidence': 0.5305762770479238}, 'entities': [{'start': 40, 'end': 43, 'value': 'shibes', 'entity': 'group', 'confidence': 0.9894194765511979, 'extractor': 'ner_crf', 'processors': ['ner_synonyms']}], 'intent_ranking': [{'name': 'mood_unhappy', 'confidence': 0.5305762770479238}, {'name': 'goodbye', 'confidence': 0.1254236380112934}, {'name': 'mood_great', 'confidence': 0.10182979881087542}, {'name': 'inform', 'confidence': 0.09953166268980061}, {'name': 'greet', 'confidence': 0.07776521429451069}, {'name': 'mood_affirm', 'confidence': 0.04546210008854246}, {'name': 'mood_deny', 'confidence': 0.01941130905705334}], 'text': 'I am sad, plased send me a picture of a dog'}
(rasabot) founder@hilbert:~/rasabot$
Training the Rasa Core Model
The actions file that we created in Part 1, now needs to be run on a separate server. This is a change in the latest version of Rasa Core. Read the documentation for more details.
Start the custom action server
(rasabot) founder@hilbert:~/rasabot$ python -m rasa_core_sdk.endpoint --actions actions
2019-02-19 09:00:02 INFO __main__ - Starting action endpoint server...
2019-02-19 09:00:02 INFO rasa_core_sdk.executor - Registered function for 'action_retrieve_image'.
2019-02-19 09:00:02 INFO __main__ - Action endpoint is up and running. on ('0.0.0.0', 5055)\
터미널을 새로 열어 Rasa Core 모델을 훈련시킨다.
(rasabot) founder@hilbert:~/rasabot$ python dialogue_management_model.py
Processed Story Blocks: 100%|███| 12/12 [00:00<00:00, 3103.25it/s, # trackers=1]
Processed Story Blocks: 100%|███| 12/12 [00:00<00:00, 511.91it/s, # trackers=11]
Processed Story Blocks: 100%|███| 12/12 [00:00<00:00, 293.16it/s, # trackers=17]
Processed Story Blocks: 100%|███| 12/12 [00:00<00:00, 276.09it/s, # trackers=14]
Processed actions: 470it [00:00, 12279.86it/s, # examples=382]
_________________________________________________________________
Layer (type) Output Shape Param #
=================================================================
masking (Masking) (None, 5, 24) 0
_________________________________________________________________
lstm (LSTM) (None, 32) 7296
_________________________________________________________________
dense (Dense) (None, 15) 495
_________________________________________________________________
activation (Activation) (None, 15) 0
=================================================================
Total params: 7,791
Trainable params: 7,791
Non-trainable params: 0
_________________________________________________________________
Epoch 1/100
470/470 [==============================] - 1s 2ms/step - loss: 2.5890 - acc: 0.3043
Epoch 2/100
470/470 [==============================] - 0s 299us/step - loss: 2.3293 - acc: 0.4957
Epoch 3/100
470/470 [==============================] - 0s 302us/step - loss: 2.0503 - acc: 0.4979
Epoch 4/100
470/470 [==============================] - 0s 342us/step - loss: 1.8383 - acc: 0.4979
Epoch 5/100
470/470 [==============================] - 0s 315us/step - loss: 1.7796 - acc: 0.4979
Epoch 6/100
470/470 [==============================] - 0s 311us/step - loss: 1.7170 - acc: 0.4979
Epoch 7/100
470/470 [==============================] - 0s 351us/step - loss: 1.6711 - acc: 0.4979
Epoch 8/100
470/470 [==============================] - 0s 339us/step - loss: 1.6139 - acc: 0.4979
Epoch 9/100
470/470 [==============================] - 0s 351us/step - loss: 1.5845 - acc: 0.4979
Epoch 10/100
470/470 [==============================] - 0s 335us/step - loss: 1.5456 - acc: 0.4979
Epoch 11/100
470/470 [==============================] - 0s 333us/step - loss: 1.4948 - acc: 0.4979
Epoch 12/100
470/470 [==============================] - 0s 281us/step - loss: 1.4581 - acc: 0.4979
Epoch 13/100
470/470 [==============================] - 0s 325us/step - loss: 1.4139 - acc: 0.4979
Epoch 14/100
470/470 [==============================] - 0s 323us/step - loss: 1.3700 - acc: 0.5000
Epoch 15/100
470/470 [==============================] - 0s 376us/step - loss: 1.3357 - acc: 0.5043
Epoch 16/100
470/470 [==============================] - 0s 376us/step - loss: 1.3015 - acc: 0.5191
Epoch 17/100
470/470 [==============================] - 0s 207us/step - loss: 1.2591 - acc: 0.5404
Epoch 18/100
470/470 [==============================] - 0s 224us/step - loss: 1.2281 - acc: 0.5362
Epoch 19/100
470/470 [==============================] - 0s 209us/step - loss: 1.1905 - acc: 0.5617
Epoch 20/100
470/470 [==============================] - 0s 194us/step - loss: 1.1591 - acc: 0.5702
Epoch 21/100
470/470 [==============================] - 0s 186us/step - loss: 1.1229 - acc: 0.5830
Epoch 22/100
470/470 [==============================] - 0s 186us/step - loss: 1.0910 - acc: 0.5872
Epoch 23/100
470/470 [==============================] - 0s 218us/step - loss: 1.0473 - acc: 0.6191
Epoch 24/100
470/470 [==============================] - 0s 200us/step - loss: 1.0314 - acc: 0.6170
Epoch 25/100
470/470 [==============================] - 0s 203us/step - loss: 0.9916 - acc: 0.6426
Epoch 26/100
470/470 [==============================] - 0s 193us/step - loss: 0.9683 - acc: 0.6468
Epoch 27/100
470/470 [==============================] - 0s 196us/step - loss: 0.9296 - acc: 0.6745
Epoch 28/100
470/470 [==============================] - 0s 212us/step - loss: 0.9023 - acc: 0.6872
Epoch 29/100
470/470 [==============================] - 0s 212us/step - loss: 0.8729 - acc: 0.7149
Epoch 30/100
470/470 [==============================] - 0s 196us/step - loss: 0.8722 - acc: 0.6936
Epoch 31/100
470/470 [==============================] - 0s 191us/step - loss: 0.8239 - acc: 0.7340
Epoch 32/100
470/470 [==============================] - 0s 192us/step - loss: 0.8179 - acc: 0.7277
Epoch 33/100
470/470 [==============================] - 0s 195us/step - loss: 0.7904 - acc: 0.7532
Epoch 34/100
470/470 [==============================] - 0s 195us/step - loss: 0.7744 - acc: 0.7809
Epoch 35/100
470/470 [==============================] - 0s 190us/step - loss: 0.7499 - acc: 0.7532
Epoch 36/100
470/470 [==============================] - 0s 206us/step - loss: 0.7061 - acc: 0.7872
Epoch 37/100
470/470 [==============================] - 0s 215us/step - loss: 0.6891 - acc: 0.8085
Epoch 38/100
470/470 [==============================] - 0s 198us/step - loss: 0.7002 - acc: 0.7702
Epoch 39/100
470/470 [==============================] - 0s 208us/step - loss: 0.6630 - acc: 0.8000
Epoch 40/100
470/470 [==============================] - 0s 203us/step - loss: 0.6497 - acc: 0.8064
Epoch 41/100
470/470 [==============================] - 0s 197us/step - loss: 0.6276 - acc: 0.8021
Epoch 42/100
470/470 [==============================] - 0s 195us/step - loss: 0.6059 - acc: 0.8213
Epoch 43/100
470/470 [==============================] - 0s 207us/step - loss: 0.6136 - acc: 0.8106
Epoch 44/100
470/470 [==============================] - 0s 242us/step - loss: 0.5664 - acc: 0.8426
Epoch 45/100
470/470 [==============================] - 0s 196us/step - loss: 0.5551 - acc: 0.8489
Epoch 46/100
470/470 [==============================] - 0s 199us/step - loss: 0.5774 - acc: 0.8255
Epoch 47/100
470/470 [==============================] - 0s 206us/step - loss: 0.5429 - acc: 0.8319
Epoch 48/100
470/470 [==============================] - 0s 216us/step - loss: 0.5256 - acc: 0.8404
Epoch 49/100
470/470 [==============================] - 0s 235us/step - loss: 0.5075 - acc: 0.8574
Epoch 50/100
470/470 [==============================] - 0s 199us/step - loss: 0.5122 - acc: 0.8319
Epoch 51/100
470/470 [==============================] - 0s 196us/step - loss: 0.4940 - acc: 0.8404
Epoch 52/100
470/470 [==============================] - 0s 185us/step - loss: 0.4914 - acc: 0.8426
Epoch 53/100
470/470 [==============================] - 0s 189us/step - loss: 0.4816 - acc: 0.8383
Epoch 54/100
470/470 [==============================] - 0s 187us/step - loss: 0.4551 - acc: 0.8596
Epoch 55/100
470/470 [==============================] - 0s 186us/step - loss: 0.4444 - acc: 0.8617
Epoch 56/100
470/470 [==============================] - 0s 186us/step - loss: 0.4442 - acc: 0.8596
Epoch 57/100
470/470 [==============================] - 0s 192us/step - loss: 0.4378 - acc: 0.8660
Epoch 58/100
470/470 [==============================] - 0s 212us/step - loss: 0.4251 - acc: 0.8660
Epoch 59/100
470/470 [==============================] - 0s 201us/step - loss: 0.4228 - acc: 0.8638
Epoch 60/100
470/470 [==============================] - 0s 221us/step - loss: 0.4125 - acc: 0.8532
Epoch 61/100
470/470 [==============================] - 0s 191us/step - loss: 0.3825 - acc: 0.8596
Epoch 62/100
470/470 [==============================] - 0s 198us/step - loss: 0.3827 - acc: 0.8723
Epoch 63/100
470/470 [==============================] - 0s 203us/step - loss: 0.3824 - acc: 0.8681
Epoch 64/100
470/470 [==============================] - 0s 193us/step - loss: 0.3945 - acc: 0.8553
Epoch 65/100
470/470 [==============================] - 0s 188us/step - loss: 0.3741 - acc: 0.8532
Epoch 66/100
470/470 [==============================] - 0s 182us/step - loss: 0.3532 - acc: 0.8809
Epoch 67/100
470/470 [==============================] - 0s 200us/step - loss: 0.3518 - acc: 0.8787
Epoch 68/100
470/470 [==============================] - 0s 192us/step - loss: 0.3685 - acc: 0.8574
Epoch 69/100
470/470 [==============================] - 0s 201us/step - loss: 0.3623 - acc: 0.8596
Epoch 70/100
470/470 [==============================] - 0s 190us/step - loss: 0.3579 - acc: 0.8532
Epoch 71/100
470/470 [==============================] - 0s 218us/step - loss: 0.3455 - acc: 0.8638
Epoch 72/100
470/470 [==============================] - 0s 202us/step - loss: 0.3267 - acc: 0.8660
Epoch 73/100
470/470 [==============================] - 0s 191us/step - loss: 0.3212 - acc: 0.8872
Epoch 74/100
470/470 [==============================] - 0s 189us/step - loss: 0.3335 - acc: 0.8574
Epoch 75/100
470/470 [==============================] - 0s 196us/step - loss: 0.3117 - acc: 0.8766
Epoch 76/100
470/470 [==============================] - 0s 195us/step - loss: 0.3275 - acc: 0.8702
Epoch 77/100
470/470 [==============================] - 0s 197us/step - loss: 0.3208 - acc: 0.8745
Epoch 78/100
470/470 [==============================] - 0s 190us/step - loss: 0.3161 - acc: 0.8468
Epoch 79/100
470/470 [==============================] - 0s 193us/step - loss: 0.2953 - acc: 0.8787
Epoch 80/100
470/470 [==============================] - 0s 205us/step - loss: 0.3096 - acc: 0.8723
Epoch 81/100
470/470 [==============================] - 0s 188us/step - loss: 0.3079 - acc: 0.8532
Epoch 82/100
470/470 [==============================] - 0s 212us/step - loss: 0.3114 - acc: 0.8638
Epoch 83/100
470/470 [==============================] - 0s 216us/step - loss: 0.2918 - acc: 0.8723
Epoch 84/100
470/470 [==============================] - 0s 197us/step - loss: 0.2886 - acc: 0.8766
Epoch 85/100
470/470 [==============================] - 0s 186us/step - loss: 0.2954 - acc: 0.8745
Epoch 86/100
470/470 [==============================] - 0s 199us/step - loss: 0.2719 - acc: 0.8851
Epoch 87/100
470/470 [==============================] - 0s 197us/step - loss: 0.2785 - acc: 0.8787
Epoch 88/100
470/470 [==============================] - 0s 194us/step - loss: 0.2823 - acc: 0.8638
Epoch 89/100
470/470 [==============================] - 0s 194us/step - loss: 0.2840 - acc: 0.8723
Epoch 90/100
470/470 [==============================] - 0s 219us/step - loss: 0.2669 - acc: 0.8766
Epoch 91/100
470/470 [==============================] - 0s 191us/step - loss: 0.2617 - acc: 0.8894
Epoch 92/100
470/470 [==============================] - 0s 216us/step - loss: 0.2646 - acc: 0.8830
Epoch 93/100
470/470 [==============================] - 0s 191us/step - loss: 0.2688 - acc: 0.8830
Epoch 94/100
470/470 [==============================] - 0s 211us/step - loss: 0.2820 - acc: 0.8723
Epoch 95/100
470/470 [==============================] - 0s 204us/step - loss: 0.2520 - acc: 0.8915
Epoch 96/100
470/470 [==============================] - 0s 194us/step - loss: 0.2597 - acc: 0.8617
Epoch 97/100
470/470 [==============================] - 0s 199us/step - loss: 0.2509 - acc: 0.8830
Epoch 98/100
470/470 [==============================] - 0s 197us/step - loss: 0.2517 - acc: 0.8809
Epoch 99/100
470/470 [==============================] - 0s 193us/step - loss: 0.2446 - acc: 0.8809
Epoch 100/100
470/470 [==============================] - 0s 188us/step - loss: 0.2529 - acc: 0.8915
run_app.py 파일을 실행시켜 에이전트를 구동한다. 단, 실행 전에 해당 스크립트에 슬랙 토큰 정보가 정확히 들어가 있는지 확인한다.
OAuth & Permissons 의 2개의 토큰 정보 중 하단에 있는 Bot User OAuth Access Token 정보를 하단 스크립트 input_channel 정보에 입력한다.
input_channel = SlackInput('#your bot user authentication token')run_app.py
from rasa_core.channels.slack import SlackInput
from rasa_core.agent import Agent
from rasa_core.interpreter import RasaNLUInterpreter
import yaml
from rasa_core.utils import EndpointConfig
nlu_interpreter = RasaNLUInterpreter('./models/nlu/default/current')
action_endpoint = EndpointConfig(url="http://localhost:5055/webhook")
agent = Agent.load('./models/dialogue', interpreter = nlu_interpreter, action_endpoint = action_endpoint)
input_channel = SlackInput('xoxb-514185865477-514638817877-ZlAUhYSuoydYHkl0oCrUU7MC' #your bot user authentication token
)
agent.handle_channels([input_channel], 5004, serve_forever=True)
ngrok 를 5004번 포트로 구동시킨 후, ngrok_url 을 복사하여 Event Subscriptions > Request URL 란에 다음과 같은 형식으로 입력한다. 확인이 뜰 때까지 기다린다.
https://<your_ngrok_url>/webhooks/slack/webhook
마지막으로, 다음 2개의 Workplace events 를 받기로 한다.
- app_mention : 누군가가 자기 이름을 호출했을 경우 봇이 반응하도록 한다
- message_im : 봇에게 DM (direct messages)을 보낼 수 있게 한다.
그리고 다음과 같이 run_app.py 파일을 실행한다.
(rasabot) founder@hilbert:~/rasabot$ python run_app.py
Let’s Talk
1) custom actions 서버가 구동중인지 확인한다
2) ngrok 가 5004번 포트에서 구동중인지 확인한다
3) Slack 인터페이스에서 봇과 대화한다
엄청나게 힘든 작업처럼 들릴지 몰라도 한 단계씩 따라하다보면 어느 순간 Zoe 라는 멋진 챗봇을 완성할 수 있을 것이다.
아래와 같이 3개의 터미널이 돌아가고 있다.
[원문] https://towardsdatascience.com/building-a-conversational-chatbot-for-slack-using-rasa-and-python-part-2-ce7233f2e9e7
728x90
'프로그래밍 Programming' 카테고리의 다른 글
| 서프라이즈 라이브러리를 활용한 추천시스템 구축 및 검증 Building and Testing Recommender Systems With Surprise, Step-By-Step (0) | 2019.03.02 |
|---|---|
| 주피터랩 살펴보기 Jupyter Lab: Evolution of the Jupyter Notebook (0) | 2019.02.25 |
| Rasa Stack 과 파이썬을 활용한 슬랙 챗봇 만들기 (1) A guide to creating a chatbot with Rasa stack and Python (0) | 2019.02.22 |
| 핀터레스트 위젯 빌더 Pintereest Widget builder (0) | 2019.02.16 |
| 웹사이트, 블로그에 핀터레스트 핀 넣기 (0) | 2019.02.16 |