Improvements to Google Parser

changes/375.added

@@ -0,0 +1 @@
+Added Google subject parser with support for multiple status types and notifications without end times

circuit_maintenance_parser/parsers/google.py

@@ -4,7 +4,7 @@
 import re
 from datetime import datetime
 
-from circuit_maintenance_parser.parser import CircuitImpact, Html, Impact, Status
+from circuit_maintenance_parser.parser import CircuitImpact, EmailSubjectParser, Html, Impact, Status
 
 # pylint: disable=too-many-nested-blocks, too-many-branches
 
@@ -18,7 +18,7 @@ def parse_html(self, soup):
         """Execute parsing."""
         data = {}
         data["circuits"] = []
-        data["status"] = Status.CONFIRMED
+        end_time_explicit = False
 
         for span in soup.find_all("span"):
             if span.string is None:
@@ -29,6 +29,7 @@ def parse_html(self, soup):
             elif span.string.strip() == "End Time:":
                 dt_str = span.next_sibling.string.strip()
                 data["end"] = self.dt2ts(datetime.strptime(dt_str, "%Y-%m-%d %H:%M:%S %z UTC"))
+                end_time_explicit = True
             elif span.string.strip() == "Peer ASN:":
                 data["account"] = span.parent.next_sibling.string.strip()
             elif span.string.strip() == "Google Neighbor Address(es):":
@@ -37,9 +38,42 @@ def parse_html(self, soup):
                 cid = googleaddr + "-" + span.parent.next_sibling.string.strip()
                 data["circuits"].append(CircuitImpact(circuit_id=cid, impact=Impact.OUTAGE))
 
-        summary = list(soup.find("div").find("div").strings)[-1].strip()
-        match = re.search(r" - Reference (.*)$", summary)
-        data["summary"] = summary
-        data["maintenance_id"] = match[1]
+        # Google sometimes send notifications without End Time specificed
+        if not end_time_explicit and data["start"]:
+            # Since start and end times cannot be equal, manufacturing end date by adding 1hr to start date
+            end_time_delta = 3600
+            data["end"] = data["start"] + end_time_delta
+
+        return [data]
+
+
+class SubjectParserGoogle1(EmailSubjectParser):
+    """Subject Parser for Google notifications."""
+
+    def parse_subject(self, subject):
+        """Parse the subject line."""
+        data = {}
+
+        # Example subject format - "[Scheduled] Google Planned Network Maintenance Notification - Reference PCR/123456"
+        # Group 1: Status (e.g., Scheduled, Completed, Canceled)
+        # Group 2: Maintenance ID (e.g., PCR/123456)
+        match = re.search(r"(\[\S+\]).*Reference\s+(\S+)", subject, re.IGNORECASE | re.DOTALL)
+        match_2 = re.search(r"\[\S+\]\s+(.*)", subject, re.IGNORECASE | re.DOTALL)
+
+        if match:
+            status_str = match.group(1).upper()
+            data["maintenance_id"] = match.group(2).strip()
+            if "COMPLETED" in status_str:
+                data["status"] = Status.COMPLETED
+            # To handle both Cancelled and Canceled spelling options just in case
+            elif "CANCEL" in status_str:
+                data["status"] = Status.CANCELLED
+            elif "SCHEDULED" in status_str:
+                data["status"] = Status.CONFIRMED
+            # If unable to match, we fallback to default confirmed
+            else:
+                data["status"] = Status.CONFIRMED
+        if match_2:
+            data["summary"] = match_2.group(1)
 
         return [data]

circuit_maintenance_parser/provider.py

@@ -24,7 +24,7 @@
 from circuit_maintenance_parser.parsers.crowncastle import HtmlParserCrownCastle1
 from circuit_maintenance_parser.parsers.equinix import HtmlParserEquinix, SubjectParserEquinix
 from circuit_maintenance_parser.parsers.globalcloudxchange import HtmlParserGcx1, SubjectParserGcx1
-from circuit_maintenance_parser.parsers.google import HtmlParserGoogle1
+from circuit_maintenance_parser.parsers.google import HtmlParserGoogle1, SubjectParserGoogle1
 from circuit_maintenance_parser.parsers.gtt import HtmlParserGTT1
 from circuit_maintenance_parser.parsers.hgc import HtmlParserHGC1, HtmlParserHGC2, SubjectParserHGC1
 from circuit_maintenance_parser.parsers.lumen import HtmlParserLumen1
@@ -364,7 +364,7 @@ class Google(GenericProvider):
 
     _processors: List[GenericProcessor] = PrivateAttr(
         [
-            CombinedProcessor(data_parsers=[EmailDateParser, HtmlParserGoogle1]),
+            CombinedProcessor(data_parsers=[EmailDateParser, HtmlParserGoogle1, SubjectParserGoogle1]),
         ]
     )
     _default_organizer = PrivateAttr("noc-noreply@google.com")

tests/unit/data/google/google1_parser_result.json → tests/unit/data/google/google1_html_parser_result.json

@@ -12,9 +12,6 @@
       }
     ],
     "end": 1701839100,
-    "maintenance_id": "PCR/123456",
-    "start": 1701809700,
-    "status": "CONFIRMED",
-    "summary": "Network Maintenance Notification - Reference PCR/123456"
+    "start": 1701809700
   }
 ]

tests/unit/data/google/google2_scheduled_subject

@@ -0,0 +1 @@
+[Scheduled] Google Planned Network Maintenance Notification - Reference PCR/123456

tests/unit/data/google/google2_scheduled_subject_parser_result.json

@@ -0,0 +1,7 @@
+[
+  {
+    "maintenance_id": "PCR/123456",
+    "status": "CONFIRMED",
+    "summary": "Google Planned Network Maintenance Notification - Reference PCR/123456"
+  }
+]

tests/unit/data/google/google3_completed_subject

@@ -0,0 +1 @@
+[Completed] Google Planned Network Maintenance Notification - Reference PCR/123456

tests/unit/data/google/google3_completed_subject_parser_result.json

@@ -0,0 +1,7 @@
+[
+  {
+    "maintenance_id": "PCR/123456",
+    "status": "COMPLETED",
+    "summary": "Google Planned Network Maintenance Notification - Reference PCR/123456"
+  }
+]

tests/unit/data/google/google4_canceled_subject

@@ -0,0 +1 @@
+[Canceled] Google Planned Network Maintenance Notification - Reference PCR/123456

tests/unit/data/google/google4_canceled_subject_parser_result.json

@@ -0,0 +1,7 @@
+[
+  {
+    "maintenance_id": "PCR/123456",
+    "status": "CANCELLED",
+    "summary": "Google Planned Network Maintenance Notification - Reference PCR/123456"
+  }
+]

tests/unit/data/google/google2_result.json → tests/unit/data/google/google5_scheduled_result.json

@@ -19,7 +19,7 @@
     "stamp": 1701823205,
     "start": 1701809700,
     "status": "CONFIRMED",
-    "summary": "Network Maintenance Notification - Reference PCR/123456",
+    "summary": "Google Planned Network Maintenance Notification - Reference PCR/123456",
     "uid": "0"
   }
 ]

tests/unit/data/google/google6_completed.eml

@@ -0,0 +1,152 @@
+Date: Tue, 15 Jul 2025 13:09:19 +0000
+Subject: [Completed] Google Planned Network Maintenance Notification -
+ Reference PCR/123456
+From: Google NOC <noc-noreply@google.com>
+To: sample@example.net
+Content-Type: multipart/alternative; boundary="000000000000ecbeea0639f77b7f"
+
+--000000000000ecbeea0639f77b7f
+Content-Type: text/plain; charset="UTF-8"; format=flowed; delsp=yes
+
+Google Network Maintenance Notification - Reference PCR/123456 - Completed
+
+
+This message was sent to: [sample@example.net]
+These addresses were selected from the Google ISP Portal. You can update  
+your preferences by navigating to "Configuration > Contacts" and adjusting  
+Email Subscriptions for each user. You can find additional details about  
+Contacts here.
+
+
+Details:
+Start Time: 2025-07-15 03:55:00 +0000 UTC
+End Time: 2025-07-15 14:05:00 +0000 UTC
+Duration: 10h10m0s
+
+
+Google ASN: 15169 Peer ASN: 65001
+
+Google Neighbor Address(es): 2001:db8::1 Peer Neighbor  
+Address(es): 2001:db8::
+
+Google Neighbor Address(es): 192.0.2.1 Peer Neighbor Address(es):  
+192.0.2.0
+
+
+Please note that the maintenance is completed, Google systems will  
+automatically re-route traffic back.
+
+Questions about this event can be sent to noc@google.com.
+
+Thank you for peering with Google!
+
+
+
+PRIVILEGE AND CONFIDENTIALITY NOTICE: The information in this e-mail  
+communication and any attached documents may be privileged, confidential or  
+otherwise protected from disclosure and is intended only for the use of the  
+designated recipient(s). If the reader is neither the intended recipient  
+nor an employee or agent thereof who is responsible for delivering it to  
+the intended recipient, you are hereby notified that any review,  
+dissemination, distribution, copying or other use of this communication is  
+strictly prohibited. If you have received this communication in error,  
+please immediately notify us by return e-mail and promptly delete the  
+original electronic e-mail communication and any attached documentation.  
+Receipt by anyone other than the intended recipient is not a waiver of any  
+attorney-client or work-product privilege.
+
+
+--000000000000ecbeea0639f77b7f
+Content-Type: text/html; charset="UTF-8"
+Content-Transfer-Encoding: quoted-printable
+
+
+<div style=3D"font-family: 'Open Sans', sans-serif;">
+  <div style=3D"font-weight: bold; font-size: larger;"><span style=3D"color=
+: rgb(51, 105, 232);">G</span><span style=3D"color: rgb(213, 15, 37);">o</s=
+pan><span style=3D"color: rgb(238, 178, 17);">o</span><span style=3D"color:=
+ rgb(51, 105, 232);">g</span><span style=3D"color: rgb(0, 153, 37);">l</spa=
+n><span style=3D"color: rgb(213, 15, 37);">e</span>  Network Maintenance No=
+tification - Reference PCR/123456 - <span style=3D"color: rgb(0, 153, 37);"=
+>Completed</span> </div>  <hr />
+ =20
+  <div style=3D"font-size: smaller;">
+  This message was sent to: [sample@example.net]<b=
+r />
+  These addresses were selected from the <a href=3D"http://peering.google.c=
+om/">Google ISP Portal</a>. You can update your preferences by navigating t=
+o "Configuration > Contacts" and adjusting Email Subscriptions for each use=
+r. You can find additional details about Contacts <a href=3D"https://suppor=
+t.google.com/interconnect/answer/7658597?hl=3Den&ref_topic=3D7650153">here<=
+/a>.
+  </div>
+
+  <br />
+  <br />
+ =20
+  <span style=3D"font-weight: bold;">Details:</span><br />
+  <span style=3D"font-weight: bold;">Start Time:</span> 2025-07-15 03:55:00=
+ &#43;0000 UTC<br />
+  <span style=3D"font-weight: bold;">End Time:</span> 2025-07-15 14:05:00 &=
+#43;0000 UTC<br />
+  <span style=3D"font-weight: bold;">Duration:</span> 10h10m0s<br />
+  <table>
+    <tr>
+      <td><span style=3D"font-weight: bold;">Google ASN:</span></td><td>151=
+69</td>
+      <td><span style=3D"font-weight: bold;">Peer ASN:</span></td><td>65001=
+</td>
+    </tr>
+   =20
+    <tr style=3D"vertical-align: top;">
+      <td><span style=3D"font-weight: bold;">Google Neighbor Address(es):</=
+span></td><td>2001:db8::1</td>
+      <td><span style=3D"font-weight: bold;">Peer Neighbor Address(es):</sp=
+an></td><td>2001:db8::</td>
+    </tr>
+   =20
+    <tr style=3D"vertical-align: top;">
+      <td><span style=3D"font-weight: bold;">Google Neighbor Address(es):</=
+span></td><td>192.0.2.1</td>
+      <td><span style=3D"font-weight: bold;">Peer Neighbor Address(es):</sp=
+an></td><td>192.0.2.0</td>
+    </tr>
+   =20
+  </table>
+
+  <p>
+  Please note that the maintenance is completed, Google systems will automa=
+tically re-route traffic back.
+  </p>
+  <p>
+  Questions about this event can be sent to noc@google.com.
+  </p>
+  <p>
+  Thank you for peering with <span style=3D"color: rgb(51, 105, 232);">G</s=
+pan><span style=3D"color: rgb(213, 15, 37);">o</span><span style=3D"color: =
+rgb(238, 178, 17);">o</span><span style=3D"color: rgb(51, 105, 232);">g</sp=
+an><span style=3D"color: rgb(0, 153, 37);">l</span><span style=3D"color: rg=
+b(213, 15, 37);">e</span>!<br />
+  </p>
+  <hr />
+  <br />
+  <div style=3D"font-size: smaller; color: silver;">
+    PRIVILEGE AND CONFIDENTIALITY NOTICE: The information in this
+    e-mail communication and any attached documents may be privileged,
+    confidential or otherwise protected from disclosure and is intended onl=
+y
+    for the use of the designated recipient(s). If the reader is neither th=
+e
+    intended recipient nor an employee or agent thereof who is responsible
+    for delivering it to the intended recipient, you are hereby notified
+    that any review, dissemination, distribution, copying or other use of
+    this communication is strictly prohibited. If you have received this
+    communication in error, please immediately notify us by return e-mail
+    and promptly delete the original electronic e-mail communication and
+    any attached documentation. Receipt by anyone other than the intended
+    recipient is not a waiver of any attorney-client or work-product
+    privilege.
+  </div>
+</div>
+
+--000000000000ecbeea0639f77b7f--

tests/unit/data/google/google6_completed_result.json

@@ -0,0 +1,25 @@
+[
+  {
+    "account": "65001",
+    "circuits": [
+      {
+        "circuit_id": "2001:db8::1-2001:db8::",
+        "impact": "OUTAGE"
+      },
+      {
+        "circuit_id": "192.0.2.1-192.0.2.0",
+        "impact": "OUTAGE"
+      }
+    ],
+    "end": 1752588300,
+    "maintenance_id": "PCR/123456",
+    "organizer": "noc-noreply@google.com",
+    "provider": "google",
+    "sequence": 1,
+    "stamp": 1752584959,
+    "start": 1752551700,
+    "status": "COMPLETED",
+    "summary": "Google Planned Network Maintenance Notification - Reference PCR/123456",
+    "uid": "0"
+  }
+]