Add Jina integration tests for text/document embedders and ranker (#2841)

* Add Jina integration tests for text/document embedders and ranker

* small improvements

* fmt

---------

Co-authored-by: anakin87 <stefanofiorucci@gmail.com>

Author: kudos07

.github/workflows/jina.yml

@@ -10,7 +10,6 @@ on:
       - "integrations/jina/**"
       - "!integrations/jina/*.md"
       - ".github/workflows/jina.yml"
-
 defaults:
   run:
     working-directory: integrations/jina
@@ -20,6 +19,7 @@ concurrency:
   cancel-in-progress: true
 
 env:
+  JINA_API_KEY: ${{ secrets.JINA_API_KEY }}
   PYTHONUNBUFFERED: "1"
   FORCE_COLOR: "1"
 

integrations/jina/tests/test_document_embedder.py

@@ -2,6 +2,7 @@
 #
 # SPDX-License-Identifier: Apache-2.0
 import json
+import os
 from unittest.mock import patch
 
 import pytest
@@ -311,3 +312,27 @@ def test_run_with_v3(self):
             assert len(doc.embedding) == 3
             assert all(isinstance(x, float) for x in doc.embedding)
         assert metadata == {"model": model, "usage": {"prompt_tokens": 2 * 4, "total_tokens": 2 * 4}}
+
+    @pytest.mark.skipif(not os.environ.get("JINA_API_KEY", None), reason="JINA_API_KEY env var not set")
+    @pytest.mark.integration
+    def test_run_integration(self):
+        embedder = JinaDocumentEmbedder(model="jina-embeddings-v3", task="retrieval.passage")
+        docs = [
+            Document(content="Paris is the capital of France."),
+            Document(content="Berlin is the capital of Germany."),
+        ]
+
+        result = embedder.run(documents=docs)
+
+        assert "documents" in result
+        assert len(result["documents"]) == 2
+        for doc in result["documents"]:
+            assert isinstance(doc, Document)
+            assert isinstance(doc.embedding, list)
+            assert len(doc.embedding) > 0
+            assert all(isinstance(x, (int, float)) for x in doc.embedding)
+
+        assert "meta" in result
+        assert isinstance(result["meta"], dict)
+        assert "model" in result["meta"]
+        assert "usage" in result["meta"]

integrations/jina/tests/test_ranker.py

@@ -2,6 +2,7 @@
 #
 # SPDX-License-Identifier: Apache-2.0
 import json
+import os
 from unittest.mock import patch
 
 import pytest
@@ -145,3 +146,28 @@ def test_run_on_empty_docs(self):
 
         assert result["documents"] is not None
         assert not result["documents"]  # empty list
+
+    @pytest.mark.skipif(not os.environ.get("JINA_API_KEY", None), reason="JINA_API_KEY env var not set")
+    @pytest.mark.integration
+    def test_run_integration(self):
+        ranker = JinaRanker(model="jina-reranker-v1-base-en")
+        docs = [
+            Document(content="Paris is the capital of France."),
+            Document(content="Bananas are yellow fruits."),
+            Document(content="Berlin is the capital of Germany."),
+        ]
+
+        result = ranker.run(query="What is the capital of France?", documents=docs, top_k=2)
+
+        assert "documents" in result
+        ranked_docs = result["documents"]
+        assert len(ranked_docs) == 2
+        assert all(isinstance(doc, Document) for doc in ranked_docs)
+        assert all(doc.score is not None for doc in ranked_docs)
+        assert ranked_docs[0].score >= ranked_docs[1].score
+        assert "Paris" in ranked_docs[0].content
+
+        assert "meta" in result
+        assert isinstance(result["meta"], dict)
+        assert "model" in result["meta"]
+        assert "usage" in result["meta"]

integrations/jina/tests/test_reader_connector.py

@@ -122,7 +122,7 @@ def test_run_reader_mode(self):
         assert len(result) == 1
         document = result["documents"][0]
         assert isinstance(document, Document)
-        assert "This domain is for use in illustrative examples" in document.content
+        assert "This domain is for use in documentation examples" in document.content
         assert document.meta["title"] == "Example Domain"
         assert document.meta["url"] == "https://example.com/"
 
@@ -135,7 +135,7 @@ def test_run_search_mode(self):
         assert len(result) >= 1
         for doc in result["documents"]:
             assert isinstance(doc, Document)
-            assert doc.content
+            assert doc.content is not None
             assert "title" in doc.meta
             assert "url" in doc.meta
             assert "description" in doc.meta

integrations/jina/tests/test_text_embedder.py

@@ -2,6 +2,7 @@
 #
 # SPDX-License-Identifier: Apache-2.0
 import json
+import os
 from unittest.mock import patch
 
 import pytest
@@ -145,3 +146,19 @@ def test_with_v3(self):
             "model": "jina-embeddings-v3",
             "usage": {"prompt_tokens": 6, "total_tokens": 6},
         }
+
+    @pytest.mark.skipif(not os.environ.get("JINA_API_KEY", None), reason="JINA_API_KEY env var not set")
+    @pytest.mark.integration
+    def test_run_integration(self):
+        embedder = JinaTextEmbedder(task="retrieval.query")
+        result = embedder.run(text="What is the capital of France?")
+
+        assert "embedding" in result
+        assert isinstance(result["embedding"], list)
+        assert len(result["embedding"]) > 0
+        assert all(isinstance(x, (int, float)) for x in result["embedding"])
+
+        assert "meta" in result
+        assert isinstance(result["meta"], dict)
+        assert "model" in result["meta"]
+        assert "usage" in result["meta"]