Factor out build docs for building XML or JSON doc

Allows testing of logic to choose between XML and JSON

Author: mitchelljkotler

pysolr.py

@@ -916,6 +916,45 @@ def suggest_terms(self, fields, prefix, handler="terms", **kwargs):
         )
         return res
 
+    def _build_docs(self, docs, boost=None, fieldUpdates=None, commitWithin=None):
+        # if no boost needed use json multidocument api
+        #   The JSON API skips the XML conversion and speedup load from 15 to 20 times.
+        #   CPU Usage is drastically lower.
+        if boost is None:
+            solrapi = "JSON"
+            message = docs
+            # single doc convert to array of docs
+            if isinstance(message, dict):
+                # convert dict to list
+                message = [message]
+                # json array of docs
+            if isinstance(message, list):
+                # convert to string
+                cleaned_message = [
+                    self._build_json_doc(doc, fieldUpdates=fieldUpdates)
+                    for doc in message
+                ]
+                m = self.encoder.encode(cleaned_message).encode("utf-8")
+            else:
+                raise ValueError("wrong message type")
+        else:
+            solrapi = "XML"
+            message = ElementTree.Element("add")
+
+            if commitWithin:
+                message.set("commitWithin", commitWithin)
+
+            for doc in docs:
+                el = self._build_xml_doc(doc, boost=boost, fieldUpdates=fieldUpdates)
+                message.append(el)
+
+            # This returns a bytestring. Ugh.
+            m = ElementTree.tostring(message, encoding="utf-8")
+            # Convert back to Unicode please.
+            m = force_unicode(m)
+
+        return (solrapi, m, len(message))
+
     def _build_json_doc(self, doc, fieldUpdates=None):
         if fieldUpdates is None:
             cleaned_doc = {k: v for k, v in doc.items() if not self._is_null_value(v)}
@@ -928,7 +967,6 @@ def _build_json_doc(self, doc, fieldUpdates=None):
             }
 
         return cleaned_doc
-            
 
     def _build_xml_doc(self, doc, boost=None, fieldUpdates=None):
         doc_elem = ElementTree.Element("doc")
@@ -1035,43 +1073,13 @@ def add(
         """
         start_time = time.time()
         self.log.debug("Starting to build add request...")
-        solrapi = "XML"
-        # if no commands (no boost) needed use json multidocument api
-        #   The JSON API skips the XML conversion and speedup load from 15 to 20 times.
-        #   CPU Usage is drastically lower.
-        if boost is None:
-            solrapi = "JSON"
-            message = docs
-            # single doc convert to array of docs
-            if isinstance(message, dict):
-                # convert dict to list
-                message = [message]
-                # json array of docs
-            if isinstance(message, list):
-                # convert to string
-                cleaned_message = [self._build_json_doc(doc, fieldUpdates=fieldUpdates) for doc in message]
-                m = self.encoder.encode(cleaned_message).encode("utf-8")
-            else:
-                raise ValueError("wrong message type")
-        else:
-            message = ElementTree.Element("add")
-
-            if commitWithin:
-                message.set("commitWithin", commitWithin)
-
-            for doc in docs:
-                el = self._build_xml_doc(doc, boost=boost, fieldUpdates=fieldUpdates)
-                message.append(el)
-
-            # This returns a bytestring. Ugh.
-            m = ElementTree.tostring(message, encoding="utf-8")
-            # Convert back to Unicode please.
-            m = force_unicode(m)
-
+        solrapi, m, len_message = self._build_docs(
+            docs, boost, fieldUpdates, commitWithin
+        )
         end_time = time.time()
         self.log.debug(
             "Built add request of %s docs in %0.2f seconds.",
-            len(message),
+            len_message,
             end_time - start_time,
         )
         return self._update(

tests/test_client.py

@@ -772,6 +772,36 @@ def test_build_json_doc_matches_xml(self):
         self.assertNotIn("title", doc_json)
         self.assertIsNone(doc_xml.find("*[name='title']"))
 
+    def test__build_docs_plain(self):
+        docs = [{
+            "id": "doc_1",
+            "title": "",
+            "price": 12.59,
+            "popularity": 10
+        }]
+        solrapi, m, len_message = self.solr._build_docs(docs)
+        self.assertEqual(solrapi, "JSON")
+
+    def test__build_docs_boost(self):
+        docs = [{
+            "id": "doc_1",
+            "title": "",
+            "price": 12.59,
+            "popularity": 10
+        }]
+        solrapi, m, len_message = self.solr._build_docs(docs, boost={"title": 10.0})
+        self.assertEqual(solrapi, "XML")
+
+    def test__build_docs_field_updates(self):
+        docs = [{
+            "id": "doc_1",
+            "popularity": 10
+        }]
+        solrapi, m, len_message = self.solr._build_docs(
+            docs, fieldUpdates={"popularity": "inc"}
+        )
+        self.assertEqual(solrapi, "JSON")
+
     def test_add(self):
         self.assertEqual(len(self.solr.search("doc")), 3)
         self.assertEqual(len(self.solr.search("example")), 2)