From 1cd80037a5c3d83dc0eeb43a257bcb7e029bcf19 Mon Sep 17 00:00:00 2001
From: Osma Suominen <osma.suominen@helsinki.fi>
Date: Mon, 16 Sep 2024 13:53:23 +0300
Subject: [PATCH] adapt to annif.simplemma_util and newer Connexion

---
 annif/rest.py      | 16 +++++++++-------
 tests/test_rest.py |  2 +-
 2 files changed, 10 insertions(+), 8 deletions(-)

diff --git a/annif/rest.py b/annif/rest.py
index 84f8450d..c283a60b 100644
--- a/annif/rest.py
+++ b/annif/rest.py
@@ -7,12 +7,12 @@
 from typing import TYPE_CHECKING, Any
 
 import connexion
-from simplemma.langdetect import lang_detector
 
 import annif.registry
 from annif.corpus import Document, DocumentList, SubjectSet
 from annif.exception import AnnifException
 from annif.project import Access
+from annif.simplemma_util import get_language_detector
 
 if TYPE_CHECKING:
     from connexion.lifecycle import ConnexionResponse
@@ -83,7 +83,7 @@ def show_project(
     return project.dump(), 200, {"Content-Type": "application/json"}
 
 
-def detect_language(body):
+def detect_language(body: dict[str, Any]):
     """return scores for detected languages formatted according to Swagger spec"""
 
     text = body.get("text")
@@ -96,21 +96,23 @@ def detect_language(body):
             detail="no candidate languages given",
         )
 
-    scores = lang_detector(text, tuple(candidates))
-
-    if not scores:
+    detector = get_language_detector(tuple(candidates))
+    try:
+        proportions = detector.proportion_in_each_language(text)
+    except ValueError:
         return connexion.problem(
             status=400,
             title="Bad Request",
             detail="unsupported candidate languages",
         )
 
-    return {
+    result = {
         "results": [
             {"language": lang if lang != "unk" else None, "score": score}
-            for lang, score in scores
+            for lang, score in proportions.items()
         ]
     }
+    return result, 200, {"Content-Type": "application/json"}
 
 
 def _suggestion_to_dict(
diff --git a/tests/test_rest.py b/tests/test_rest.py
index 8e35aa0e..57adad64 100644
--- a/tests/test_rest.py
+++ b/tests/test_rest.py
@@ -66,7 +66,7 @@ def test_rest_detect_language_unknown(app):
     # an unknown language should return None
     with app.app_context():
         result = annif.rest.detect_language(
-            {"text": "example text", "candidates": ["fi", "sv"]}
+            {"text": "exampley texty", "candidates": ["fi", "sv"]}
         )
         assert {"language": None, "score": 1} in result["results"]