Refactors get_birthplace to get_nationality.

processing/script/master.R

@@ -43,13 +43,13 @@ results <- pbapply(articles, 1, function(article) {
   awards_found <- get_awards(annotation)
   university_found <- get_university(annotation)
   birthdate_found <- get_birthdate(cleaned_text, annotation)
-  birthplace_found <- get_birthplace(cleaned_text, annotation)
+  nationality_found <- get_nationality(cleaned_text, annotation)
 
   ## Create Results
 
   data.frame(name = article[1],
              spouse = spouse_found,
-             birthplace = birthplace_found,
+             nationality = nationality_found,
              birthdate = birthdate_found,
              day_of_death = NA,
              place_of_death = NA,

processing/wikiproc/R/get_birthplace.R

-#!/usr/bin/env Rscript
-
-# Author: Lukas
-
-#' This script extracts Birthplace from physicist texts
-#' Try to search text for first occurrence of an NORP entity
-#' in a sentence containing born
-#'
-#' @export
-#' @param article Article in HTML-format
-#' @param annotations CNLP Annotations
-#' @return String with birthplace of the physicist|NA
-get_birthplace <- function(article) {
-  
-  # Try to extract birthplace from text as first NORP entity sentence
-  # - Get tokens
-  # - Get entities
-  # - Get word column of tokens
-  text <- clean_html(article)
-  annotations <- cnlp_annotate(text)
-  tokens_obj <- cnlp_get_token(annotations)
-  entities_obj <- cnlp_get_entity(annotations)
-  words <- tokens_obj$word
-  
-  # - Find position of "born" in tokens_obj
-  # - Get sid of sentence with "born"
-  # TODO catch case with no occurrence
-  born_occurrence <- which(words %in% "born")
-  if (length(born_occurrence == 0)) {
-    return(NA)
-  }
-  born_sid <- tokens_obj[born_occurrence[1], 2]
-  
-  # - Get NORPS-entities
-  norps <- entities_obj[entities_obj$entity_type == "NORP",]
-  
-  # - Assume birthplace is first norps entitiy in sentence with "born"
-  birthplace <- norps[norps$sid == born_sid,]$entity[1]
-  
-  ifelse(length(birthplace) >= 1, birthplace, NA)
-  
-}

processing/wikiproc/R/get_nationality.R

+#!/usr/bin/env Rscript
+
+# Author: Lukas
+
+#' This script extracts Nationality from physicist texts
+#' Try to search text for first occurrence of an NORP entity
+#'
+#' @export
+#' @param article Article in HTML-format
+#' @param annotations CNLP Annotations
+#' @return String with nationality of the physicist|NA
+get_nationality <- function(article, annotations) {
+  
+  # Try to extract nationality from text as first NORP entity sentence
+  # - Get entities
+  # - Get NORPs
+  # - Get first NORP
+  # - Return NA if length < 4 because some empty strings and garbage would be returned else
+  entities_obj <- cnlp_get_entity(annotations)
+  norps <- entities_obj[entities_obj$entity_type == "NORP",]
+  nationality <- norps$entity[1]
+  ifelse(nchar(nationality) >= 4, nationality, NA)
+}

processing/wikiproc/tests/testthat/test-get_birthplace.R → processing/wikiproc/tests/testthat/test-get_nationality.R

-context("test-get_birthplace")
+context("test-get_nationality")
 
-test_that("retrieving birthplace works", {
+test_that("retrieving nationality works", {
     filename_raw <- "article-4-raw.html"
     html <- readChar(filename_raw, file.info(filename_raw)$size)
-    expected <- ("Yaycı, Nakhchivan, Azerbaijan Democratic Republic")
-    actual <- get_birthplace(html)
+    expected <- ("Russian")
+    actual <- get_nationality(html)
 
     expect_equal(expected, actual)
 })

rasa/actions.py

@@ -10,7 +10,7 @@ Schema for Actions
 
 class ActionSearchInfo(Action):
     Starts Search Action for an Information, depending on
-    found intent (Intent 'birthplace' -> 'ActionSearchBirthplace')
+    found intent (Intent 'nationality' -> 'ActionSearchNationality')
     - Obtains slot value of recognized physicist entity
     - Iterates over 'data.tsv' to find demanded info for given physicist
     - Stores value of info in its slot
@@ -24,40 +24,40 @@ class ActionUtterInfo(Action):
 """
 
 #
-# Birthplace
+# nationality
 #
-class ActionSearchBirthplace(Action):
+class ActionSearchNationality(Action):
     def name(self):
-        return 'action_search_birthplace'
+        return 'action_search_nationality'
 
     def run(self, dispatcher, tracker, domain):
         import csv
         import re
         person = tracker.get_slot('physicist')
         name_regex = re.compile(person, re.IGNORECASE)
-        actual_birthplace = None
+        actual_nationality = None
         if os.path.isfile('data.tsv'):
             with open('data.tsv') as csvfile:
                 spamreader = csv.DictReader(csvfile, delimiter='\t')
                 for row in spamreader:
                     if name_regex.match(row['name']):
-                        actual_birthplace = row['birthplace']
+                        actual_nationality = row['nationality']
         else:
             dispatcher.utter_message("Error: Data not generated!")
             
-        return [SlotSet('birthplace', actual_birthplace
-            if actual_birthplace is not None and actual_birthplace is not ""
+        return [SlotSet('nationality', actual_nationality
+            if actual_nationality is not None and actual_nationality is not ""
             else "not known")]
         
 
-class ActionUtterBirthplace(Action):
+class ActionUtterNationality(Action):
     def name(self):
-        return 'action_utter_birthplace'
+        return 'action_utter_nationality'
 
     def run(self, dispatcher, tracker, domain):
         person = tracker.get_slot('physicist')
-        birthplace = tracker.get_slot('birthplace')
-        dispatcher.utter_message("Birthplace of {} is {}.".format(person, birthplace))
+        nationality = tracker.get_slot('nationality')
+        dispatcher.utter_message("Nationality of {} is {}.".format(person, nationality))
         return []
 
 

rasa/domain.yml

@@ -2,7 +2,7 @@ intents:
   - greet
   - goodbye
   - birthdate
-  - birthplace
+  - nationality
   - day_of_death
   - place_of_death
   - is_alive
@@ -16,8 +16,8 @@ intents:
 actions:
   - utter_greet
   - utter_goodbye
-  - action_search_birthplace
-  - action_utter_birthplace
+  - action_search_nationality
+  - action_utter_nationality
   - action_search_birthdate
   - action_utter_birthdate
   - action_search_day_of_death
@@ -45,7 +45,7 @@ entities:
 slots:
   physicist:
     type: text
-  birthplace:
+  nationality:
     type: unfeaturized
   birthdate:
     type: unfeaturized

rasa/nlu.md

@@ -6,11 +6,10 @@
 - bye
 - goodbye
 
-## intent:birthplace
-- where was [Albert Einstein](physicist) born
+## intent:nationality
+- what nation is [Albert Einstein](physicist) from
+- what nationality does [Albert Einstein](physicist) have
 - where is [Galileo Galilei](physicist) from
-- what is [Albert Einstein](physicist) birthplace
-- birthplace of [Galileo Galilei](physicist)
 
 ## intent:birthdate
 - when was [Albert Einstein](physicist) born

rasa/stories.md

@@ -7,9 +7,9 @@
   - utter_goodbye
 
 ## birthplace
-* birthplace{"physicist": "albert einstein"}
-  - action_search_birthplace
-  - action_utter_birthplace
+* nationality{"physicist": "albert einstein"}
+  - action_search_nationality
+  - action_utter_nationality
 
 ## birthdate
 * birthdate{"physicist": "albert einstein}