Fix parsing different source formats (#3)

d3QUone · Nov 25, 2022 · c341a60 · c341a60
1 parent 80c648b
commit c341a60
Show file tree

Hide file tree

Showing 6 changed files with 47 additions and 10 deletions.
diff --git a/Makefile b/Makefile
@@ -2,4 +2,4 @@
 VERBOSE=1
 
 test:
-	poetry run nose2 --verbosity $(VERBOSE)
+	poetry run nose2 --verbosity $(VERBOSE) $(target)
diff --git a/aiosqs/__init__.py b/aiosqs/__init__.py
@@ -10,4 +10,4 @@
     SendMessageResponse,
 )
 
-VERSION = "1.0.1"
+VERSION = "1.0.2"
diff --git a/aiosqs/client.py b/aiosqs/client.py
@@ -138,7 +138,7 @@ async def request(self, params: Dict) -> Union[Dict, List, None]:
             self.logger.error(f"SQS API error: status_code={response.status}, body={response_body}")
             raise SQSClientBaseError
 
-        return parse_xml_result_response(action=params["Action"], body=response_body)
+        return parse_xml_result_response(action=params["Action"], body=response_body, logger=self.logger)
 
     async def get_queue_url(self, queue_name: str) -> GetQueueUrlResponse:
         params = {

diff --git a/aiosqs/parser.py b/aiosqs/parser.py
@@ -1,31 +1,50 @@
+from logging import getLogger
 from typing import Optional
 
 from lxml import etree
 
 from aiosqs.exceptions import ErrorData, SQSErrorResponse
+from aiosqs.types import LoggerType
+
+default_logger = getLogger(__name__)
+
+
+def el_text(el) -> Optional[str]:
+    if text := el.text:
+        return text.strip()
+    return None
 
 
 def collect_elements(root, xpath: str):
     multi_response = []
     for child in root.xpath(xpath):
         item = {}
         for elem in child:
-            value = elem.text.strip()
-            item[elem.tag] = value
+            item[elem.tag] = el_text(elem)
         if item:
             multi_response.append(item)
     return multi_response
 
 
 def find_request_id(root) -> Optional[str]:
     for child in root.xpath("./RequestId"):
-        if value := child.text.strip():
-            return value
+        if text := el_text(child):
+            return text
     return None
 
 
-def parse_xml_result_response(action: str, body: str):
-    root = etree.fromstring(body)
+def parse_xml_result_response(action: str, body: str, logger: Optional[LoggerType] = None):
+    logger = logger or default_logger
+    logger.debug("Message for %s: %s", action, body)
+
+    parser = etree.XMLParser(
+        remove_blank_text=True,
+        remove_comments=True,
+        remove_pis=True,
+        recover=False,
+    )
+    root = etree.fromstring(text=body, parser=parser)
+
     request_id = find_request_id(root=root)
 
     # Check for errors first
@@ -48,7 +67,7 @@ def parse_xml_result_response(action: str, body: str):
     # Response is a list of objects of the same type
     if len(elements) == 1 and len(elements[0]) == 1:
         key, value = list(elements[0].items())[0]
-        if value == "":
+        if not value:
             xpath = f"{xpath}/{key}"
             return collect_elements(root=root, xpath=xpath)
 

diff --git a/aiosqs/tests/fixtures/receive_message_ugly.xml b/aiosqs/tests/fixtures/receive_message_ugly.xml
@@ -0,0 +1 @@
+<ReceiveMessageResponse><ReceiveMessageResult><Message><MD5OfBody>cc12d89c4c6a34a4d8b78f970cd75534</MD5OfBody><Body>{&quot;external_id&quot;: 29, &quot;market&quot;: &quot;usdtrub&quot;, &quot;side&quot;: &quot;sell&quot;, &quot;amount&quot;: &quot;16.05&quot;}</Body><ReceiptHandle>1669388400-f96963bf-3f5f-40e8-8df2-9ce9ff90adf8</ReceiptHandle><MessageId>f96963bf-3f5f-40e8-8df2-9ce9ff90adf8</MessageId></Message></ReceiveMessageResult><ResponseMetadata><RequestId/></ResponseMetadata></ReceiveMessageResponse>
diff --git a/aiosqs/tests/test_receive_message.py b/aiosqs/tests/test_receive_message.py
@@ -32,6 +32,23 @@ def test_one_message(self):
             ],
         )
 
+    def test_parse_oneline_xml(self):
+        res = parse_xml_result_response(
+            action=self.action,
+            body=load_fixture("receive_message_ugly.xml"),
+        )
+        self.assertEqual(
+            res,
+            [
+                {
+                    "MD5OfBody": "cc12d89c4c6a34a4d8b78f970cd75534",
+                    "Body": '{"external_id": 29, "market": "usdtrub", "side": "sell", "amount": "16.05"}',
+                    "ReceiptHandle": "1669388400-f96963bf-3f5f-40e8-8df2-9ce9ff90adf8",
+                    "MessageId": "f96963bf-3f5f-40e8-8df2-9ce9ff90adf8",
+                },
+            ],
+        )
+
     def test_many_messages(self):
         res = parse_xml_result_response(
             action=self.action,