Merge branch 'master' into 32-camunda-reporting-input-format

diff --git a/service-analysis/Dockerfile b/service-analysis/Dockerfile
index 976abd116501890f71027063d4323dba04e06127..ad239ec90879ddb796656346ce9b6524ae2908eb 100644 (file)
--- a/service-analysis/Dockerfile
+++ b/service-analysis/Dockerfile
@@ -1,13 +1,11 @@
-FROM   python:3
+FROM   python:3.7-slim
 LABEL  maintainer="Sebastian Steiner"
 
-RUN            pip install flask
-RUN            pip install flask_restful
-RUN            pip install indicoio
-
 WORKDIR        /app
 COPY   . /app/
 
+RUN            pip install -r requirements.txt
+
 RUN            chmod a+x *.py
 
-CMD            ["python3", "./sentiment_analysis.py"]
+CMD            ["python3.7", "./sentiment_analysis.py"]

diff --git a/service-analysis/README.md b/service-analysis/README.md
index 59f472fe374df1d18402bd5c7aa85a9349462792..fb414facdeaefcf0091aa27e6bf9e16b4cbbaf37 100644 (file)
--- a/service-analysis/README.md
+++ b/service-analysis/README.md
@@ -2,8 +2,43 @@
 
 
 ## API
+Provides a REST-API to get the average sentiment of a given list of texts
+The service uses the the sentiment analysis service indico.
 
-You can post a json string to the REST API on http://localhost:8081/.
-The json string needs to contain name-value key 'text'.
+`GET`: `/` Displays the expected input for a POST
 
-The service uses the online sentiment analysis tool indico to get the sentiment for each text, calculates the average sentiment value and returns it.
+`POST`: `/` Calculates the average sentiment for a given list of texts
+- param: Term[] as Content-Tye: `application/json`
+example body of request:
+```json
+[
+       {
+               'text': 'text 1',
+       },
+       {       
+               'text': 'text 2',
+       },
+       {       
+               'text': 'text 3',
+       }
+]
+```
+JSON string may contain more value-key pairs than 'text', but 'text' is needed for the sentiment analysis. Everything else will be ignored.
+
+## API errors
+- `409`: input string does not meet specifications
+- `503`: indico sentiment analysis not available
+
+## run with docker
+- `docker build -t sentiment_analysis .`
+- `docker run -d -p YOUR_PORT:8081 sentiment_analysis`
+
+## run local
+### requirements
+- python 3.7
+- pip
+  -> indicoio
+  -> flask
+  -> flask_restfut
+### commands
+- `python3.7 sentiment_analysis.py`

diff --git a/service-analysis/requirements.txt b/service-analysis/requirements.txt
new file mode 100644 (file)
index 0000000..206be2f
--- /dev/null
+++ b/service-analysis/requirements.txt
@@ -0,0 +1,3 @@
+flask
+flask_restful
+indicoio

diff --git a/service-analysis/sentiment_analysis.py b/service-analysis/sentiment_analysis.py
index d9836af06928d75fa545db42a613e99748a288e6..933ce9c3114804a85a6f64487bfafd4cf5b703cc 100644 (file)
--- a/service-analysis/sentiment_analysis.py
+++ b/service-analysis/sentiment_analysis.py
@@ -2,30 +2,39 @@
 import json
 import indicoio
 
-from flask import Flask, request
-from flask_restful import Resource, Api
+from flask import Flask, request, jsonify
+from flask_restful import Resource, Api, output_json, abort
 
 app = Flask(__name__)
 api = Api(app)
 
 indicoio.config.api_key = '525f16078717a430f9dac17cdc9dbaa3'
 
+input_error_409 = 'Input must be a JSON string. JSON objects must contain contain key text.'
+service_error_503 = 'The sentiment analysis service is currently unavailable.'
+
 class Sentiment_Analysis(Resource):
        def get(self):
-               return "correct usage: curl -H 'content-type: application/json' -X POST http://localhost:8081 -d '$your JSON here$'"
+               return "POST a JSON string as content-type: application/json. JSON objects must contain key 'text'."
                
        def post(self):
-               tweets = request.json
+               if not request.json:
+                       return abort(409, message=input_error_409)
+               texts = request.json
+               text_array = []
                value = 0
-               for tweet in tweets:
-                       #possible performance improvement: batch sentiment: indicoio.sentiment($list_of_texts$)
-                       sentiment_value = indicoio.sentiment(tweet['text'])
-                       value += sentiment_value
-                       #returns number between 0 and 1. 
-                       #it is a probability representing the likelihood that the analyzed text is positive or negative
-                       #values > 0.5 indicate positive sentiment
-                       #values < 0.5 indicate negative sentiment
-               return value/len(tweets)
+               for text in texts:
+                       if 'text' not in text:
+                               return abort(409, message=input_error_409)
+                       text_array.append(text['text'])
+               try:
+                       for sentiment_value in indicoio.sentiment(text_array):
+                               value += sentiment_value
+               except:
+                       return make_error(503, message=service_error_503)
+               sentiment = value/len(text_array)
+               data = {'sentiment': sentiment}
+               return output_json(json.dumps(data), 200)
                                
 api.add_resource(Sentiment_Analysis, '/')