From 7b7aff4acfe4d1743672bc998d4945c5aa378d43 Mon Sep 17 00:00:00 2001
From: "Marcus Str." <marcus@marstr.online>
Date: Sun, 1 Dec 2024 15:55:34 +0100
Subject: [PATCH] Made sure that the XML parser gets valid XML as otherwise the
 check lead to crashes in the code that generates the orthos

---
 osmxml.py | 39 +++++++++++++++++++++++++--------------
 1 file changed, 25 insertions(+), 14 deletions(-)

diff --git a/osmxml.py b/osmxml.py
index 05e376e..f8ae5a5 100644
--- a/osmxml.py
+++ b/osmxml.py
@@ -13,16 +13,23 @@
 
 
 import xml.dom.minidom
+from pyexpat import ExpatError
+
 import requests
 import os
 from defines import *
 from log import *
+import time
 
 class mstr_osmxml:
     def __init__(self):
         #self._xmlfn = mstr_datafolder + "_cache/tile_" + str(lat) + "-" + str(v) + "_" + str(lng) + "-" + str(h) + ".xml"
         self._xmldata = None
         self._xmlcontent = ""
+        self._lat = 0
+        self._lng = 0
+        self._curB_lat = 0
+        self._curB_lng = 0
 
     
     # Adjust bbox for when this class should persost, but acquire data for a different bbox
@@ -58,7 +65,8 @@ class mstr_osmxml:
         mstr_msg("osmxml", "Acquiring OSM data for " + str(self._lat)+","+str(self._lng)+" - "+str(self._curB_lat)+","+str(self._curB_lng))
         
         # We will use our self-hosted API for this.
-        while self._xmlcontent == "":
+        parse = False
+        while parse == False:
             data = { 
                 "bbox": { 
                         "lat": str(self._lat),
@@ -73,21 +81,24 @@ class mstr_osmxml:
                     }
             r = requests.post(mstr_osm_endpoint, json=data)
 
-            self._xmlcontent = r.content
+            try:
+                # Attempt to parse the XML string
+                dom = xml.dom.minidom.parseString(r.content)
 
-            #if os.path.isfile(self._xmlfn):
-            #    os.remove(self._xmlfn)
-            #with open(self._xmlfn, 'wb') as textfile:
-            #    textfile.write(r.content)
+                # Check if the DOM object has a document element
+                if dom.documentElement:
+                    # Store the content in memory
+                    self._xmlcontent = r.content
+                    self._xmldata = xml.dom.minidom.parseString(self._xmlcontent)
+                    self._xmlcontent = ""  # Clear
+                    parse = True
 
-            # 1 second delay in case the request fails
-            if self._xmlcontent == "":
-            #if os.path.isfile(self._xmlfn) == False:
-                sleep(1)
-        
-        # Store the content in memory
-        self._xmldata = xml.dom.minidom.parseString(self._xmlcontent)
-        self._xmlcontent = "" # Clear
+            except ExpatError as e:
+                parse = False
+                time.sleep(1)
+            except Exception as e:
+                parse = False
+                time.sleep(1)
 
 
     # Get all nodes from the specified OSM file
-- 
2.30.2