From b0ffb998467216c8a8b3e01fdcb5724043a9e579 Mon Sep 17 00:00:00 2001
From: heinpa <heipa@posteo.de>
Date: Tue, 20 Aug 2024 00:53:38 +0200
Subject: [PATCH] add common functionality for language components

---
 qanary_helpers/language_queries.py | 77 ++++++++++++++++++++++++++++++
 qanary_helpers/qanary_queries.py   | 27 ++++++-----
 setup.py                           |  2 +-
 3 files changed, 94 insertions(+), 12 deletions(-)
 create mode 100644 qanary_helpers/language_queries.py

diff --git a/qanary_helpers/language_queries.py b/qanary_helpers/language_queries.py
new file mode 100644
index 0000000..c9a5346
--- /dev/null
+++ b/qanary_helpers/language_queries.py
@@ -0,0 +1,77 @@
+from qanary_helpers.qanary_queries import select_from_triplestore
+
+
+class question_text_with_language:
+
+    def __init__(self, uri: str, text: str, lang: str):
+        self.uri = uri
+        self.text = text
+        self.lang = lang
+
+    def get_uri(self):
+        return self.uri
+
+    def get_text(self):
+        return self.text
+
+    def get_language(self):
+        return self.lang
+
+
+def get_texts_with_detected_language_in_triplestore(triplestore_endpoint: str, graph_uri: str, lang: str) -> list[question_text_with_language]:
+    source_texts = list()
+    sparql_find_ld = """
+        PREFIX qa: <http://www.wdaqua.eu/qa#>
+        PREFIX oa: <http://www.w3.org/ns/openannotation/core/>
+        PREFIX xsd: <http://www.w3.org/2001/XMLSchema#>
+
+        SELECT *
+        FROM <{graph}>
+        WHERE {{
+        ?annotationId a qa:AnnotationOfQuestionLanguage .
+        ?annotationId oa:hasTarget ?hasTarget ;
+          oa:hasBody ?hasBody ;
+          oa:annotatedBy ?annotatedBy ;
+          oa:annotatedAt ?annotatedAt .
+        FILTER(STR(?hasBody) = {lang})
+        }}
+    """.format(
+        graph = graph_uri,
+        lang=lang
+    )
+    results = select_from_triplestore(triplestore_endpoint, sparql_find_ld)
+    for result in results["results"]["bindings"]:
+        question_uri = result["hasTarget"]["value"]
+        question_text = get_question_text_from_uri(question_uri, triplestore_endpoint)
+        source_texts.append(question_text_with_language(uri=question_uri, text=question_text, lang=lang))
+
+    return source_texts
+
+
+def get_translated_texts_in_triplestore(triplestore_endpoint: str, graph_uri: str, lang: str) -> list[question_text_with_language]:
+    source_texts = list()
+    sparql_find_ld = """
+        PREFIX qa: <http://www.wdaqua.eu/qa#>
+        PREFIX oa: <http://www.w3.org/ns/openannotation/core/>
+
+        SELECT *
+        FROM <{graph}>
+        WHERE {{
+            ?annotationId a qa:AnnotationOfQuestionTranslation .
+            ?annotationId oa:hasTarget ?hasTarget ;
+                          oa:hasBody ?hasBody ;
+                          oa:annotatedBy ?annotatedBy ;
+                          oa:annotatedAt ?annotatedAt .
+            FILTER(lang(?hasBody) = {lang}).
+        }}
+    """.format(
+        graph = graph_uri,
+        lang=lang
+    )
+    results = select_from_triplestore(triplestore_endpoint, sparql_find_ld)
+    for result in results["results"]["bindings"]:
+        question_uri = result["hasTarget"]["value"]
+        question_text = result["hasBody"]["value"]
+        source_texts.append(question_text_with_language(question_uri, question_text, lang))
+
+    return source_texts
diff --git a/qanary_helpers/qanary_queries.py b/qanary_helpers/qanary_queries.py
index 5ceac7b..686cf75 100644
--- a/qanary_helpers/qanary_queries.py
+++ b/qanary_helpers/qanary_queries.py
@@ -5,6 +5,18 @@
 import re
 
 
+def get_text_question_from_uri(question_uri, triplestore_endpoint):
+    question_raw = question_uri + "/raw"
+    logging.info("found: questionURI={0}  questionURIraw={1}".format(
+        question_uri,
+        question_raw
+    ))
+    question_text = requests.get(question_raw.replace(
+        "localhost", urlparse(triplestore_endpoint).hostname)
+    )
+    return question_text.text
+
+
 def get_text_question_in_graph(triplestore_endpoint, graph):
     """
     Retrieves the questions from the triplestore returns an array
@@ -17,7 +29,7 @@ def get_text_question_in_graph(triplestore_endpoint, graph):
     query = """
         PREFIX rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#>
         SELECT DISTINCT ?questionURI
-        FROM <{uri}> 
+        FROM <{uri}>
         WHERE {{
             ?questionURI rdf:type <http://www.wdaqua.eu/qa#Question> .
         }}
@@ -26,16 +38,9 @@ def get_text_question_in_graph(triplestore_endpoint, graph):
     results = select_from_triplestore(triplestore_endpoint, query)
     for result in results["results"]["bindings"]:
         question_uri = result['questionURI']['value']
-        question_raw = question_uri + "/raw"
-        logging.info("found: questionURI={0}  questionURIraw={1}".format(
-            question_uri,
-            question_raw
-        ))
-        question_text = requests.get(question_raw.replace(
-            "localhost", urlparse(triplestore_endpoint).hostname)
-        )
-        logging.info("found question: \"{0}\"".format(question_text.text))
-        questions.append({"uri": question_uri, "text": question_text.text})
+        question_text = get_text_question_from_uri(question_uri, triplestore_endpoint)
+        logging.info("found question: \"{0}\"".format(question_text))
+        questions.append({"uri": question_uri, "text": question_text})
 
     return questions
 
diff --git a/setup.py b/setup.py
index 8cfefe6..a813e63 100644
--- a/setup.py
+++ b/setup.py
@@ -21,7 +21,7 @@ def read_requirements():
 
 setuptools.setup(
     name="qanary-helpers",
-    version="0.2.2",
+    version="0.3.0",
     author="Andreas Both, Aleksandr Perevalov",
     author_email="andreas.both@htwk-leipzig.de, aleksandr.perevalov@hs-anhalt.de",
     description="A package that helps to build Python components for the Qanary Question Answering framework",