sentiment evaluation v0

lib/tagnews/senteval/__init__.py

@@ -0,0 +1,3 @@
+from . import eval
+
+__all__ = [eval]

lib/tagnews/senteval/eval.py

@@ -0,0 +1,82 @@
+from google.cloud import language
+from google.cloud.language import enums
+from google.cloud.language import types
+
+
+def process_google_result(text):
+    document = types.Document(content=text, type=enums.Document.Type.PLAIN_TEXT)
+    sentiment = client.analyze_entity_sentiment(document=document)
+
+    for entity in sentiment.entities:
+        clean_entity = "".join(filter(str.isalpha, entity)).lower()
+
+        if clean_entity in ["police", "officer", "cop", "officers", "pigs"]:
+
+            for mention in entity.mentions:
+                return mention.sentiment.score
+
+
+def clean_html_text(html_text):
+    return "".join(filter(str.isalpha, html_text)).lower()
+
+
+class SentimentGoogler:
+    def __init__(self):
+        self.client = self.connect_to_client()
+
+    def connect_to_client(self):
+        return language.LanguageServiceClient()
+
+    @staticmethod
+    def pre_process(html_text):
+        """
+        Parameters
+        ----------
+        html_text : str
+            Article text.
+
+        Returns
+        -------
+        words: str
+            lower case, just letters
+        """
+        words = "".join(filter(str.isalpha, html_text)).lower()
+        return words
+
+    def call_api(self, doc_text):
+        """
+        Parameters
+        ----------
+        doc_text : str
+            article text
+
+        Returns
+        -------
+        sentiment : json
+            google response call
+        """
+        cleaned_doc_text = self.pre_process(doc_text)
+        document = types.Document(
+            content=cleaned_doc_text, type=enums.Document.Type.PLAIN_TEXT
+        )
+        sentiment = self.client.analyze_entity_sentiment(document=document)
+
+        return sentiment
+
+    @staticmethod
+    def is_police_entity(sentiment_response):
+        possible_responses = [
+            "police",
+            "officer",
+            "cop",
+            "officers",
+            "pigs",
+            "policeofficer",
+        ]
+        for entity in sentiment_response.entities:
+            if clean_html_text(clean_entity) in possible_responses:
+                return entity
+            for mention in entity.mentions:
+                if clean_html_text(mention.text.content) in possible_responses:
+                    return entity
+                return False