Tata parser (#297)

* add tata parser * add tata parser unit tests * fix bs4.element import-untyped and _parse_time method return-value * correct black code formatting * ignore pylint c0104 disallowed-name in circuit_maintenance_parser/parsers/tata.py * add tata parser in README.md * avoid computing lower() every time
networktocode · Nov 29, 2024 · ffd24f5 · ffd24f5
1 parent 82efe55
commit ffd24f5
Show file tree

Hide file tree

Showing 36 changed files with 1,418 additions and 0 deletions.
diff --git a/README.md b/README.md
@@ -84,6 +84,7 @@ By default, there is a `GenericProvider` that supports a `SimpleProcessor` using
 - Netflix (AS2906 only)
 - Seaborn
 - Sparkle
+- Tata
 - Telstra (\*)
 - Turkcell
 - Verizon

diff --git a/circuit_maintenance_parser/__init__.py b/circuit_maintenance_parser/__init__.py
@@ -28,6 +28,7 @@
     PacketFabric,
     Seaborn,
     Sparkle,
+    Tata,
     Telia,
     Telstra,
     Turkcell,
@@ -59,6 +60,7 @@
     PacketFabric,
     Seaborn,
     Sparkle,
+    Tata,
     Telia,
     Telstra,
     Turkcell,

diff --git a/circuit_maintenance_parser/parsers/tata.py b/circuit_maintenance_parser/parsers/tata.py
@@ -0,0 +1,71 @@
+# pylint: disable=disallowed-name
+"""Circuit maintenance parser for Tata Email notifications."""
+from typing import List, Dict, Any
+from datetime import datetime
+
+from bs4.element import ResultSet  # type: ignore
+from circuit_maintenance_parser.output import Impact, Status
+from circuit_maintenance_parser.parser import Html, EmailSubjectParser
+
+
+class HtmlParserTata(Html):
+    """Custom Parser for HTML portion of Tata circuit maintenance notifications."""
+
+    def parse_html(self, soup: ResultSet) -> List[Dict]:
+        """Parse Tata circuit maintenance email."""
+        prev: str = ""
+        data: Dict[str, Any] = {
+            "account": "N/A",
+            "circuits": [],
+            "organizer": soup.select("a[href^=mailto]")[0].text.strip(),
+        }
+        for span in soup.find_all("span"):
+            curr = span.text.strip()
+            if curr != prev:
+                prev_lower = prev.lower()
+                if prev_lower == "ticket reference - tcl":
+                    data["maintenance_id"] = curr
+                elif prev_lower == "service id":
+                    for circuit in curr.split(","):
+                        data["circuits"].append(
+                            {
+                                "circuit_id": circuit.strip(),
+                                "impact": Impact.OUTAGE,
+                            }
+                        )
+                elif prev_lower in ("activity window (gmt)", "revised activity window (gmt)"):
+                    start_end = curr.split("to")
+                    data["start"] = self._parse_time(start_end[0])
+                    data["end"] = self._parse_time(start_end[1])
+                elif "extended up to time window" in prev_lower:
+                    if "gmt" in curr.lower():
+                        data["end"] = self._parse_time(curr)
+            prev = span.text.strip()
+
+        return [data]
+
+    @staticmethod
+    def _parse_time(string: str) -> int:
+        """Convert YYYY-MM-DD HH:MM:SS GMT to epoch."""
+        return int((datetime.strptime(string.strip(), "%Y-%m-%d %H:%M:%S GMT") - datetime(1970, 1, 1)).total_seconds())
+
+
+class SubjectParserTata(EmailSubjectParser):
+    """Custom Parser for Email subject of Tata circuit maintenance notifications."""
+
+    def parse_subject(self, subject: str) -> List[Dict]:
+        """Parse Tata Email subject for summary and status."""
+        data: Dict[str, Any] = {"summary": subject.strip().replace("\n", "")}
+        subject_lower = subject.lower()
+        if "completion" in subject_lower:
+            data["status"] = Status.COMPLETED
+        elif "reschedule" in subject_lower or "extension" in subject_lower:
+            data["status"] = Status.RE_SCHEDULED
+        elif "reminder" in subject_lower:
+            data["status"] = Status.CONFIRMED
+        elif "cancellation" in subject_lower:
+            data["status"] = Status.CANCELLED
+        else:
+            data["status"] = Status.CONFIRMED
+
+        return [data]
diff --git a/circuit_maintenance_parser/provider.py b/circuit_maintenance_parser/provider.py
@@ -37,6 +37,7 @@
     SubjectParserSeaborn2,
 )
 from circuit_maintenance_parser.parsers.sparkle import HtmlParserSparkle1
+from circuit_maintenance_parser.parsers.tata import HtmlParserTata, SubjectParserTata
 from circuit_maintenance_parser.parsers.telstra import HtmlParserTelstra1, HtmlParserTelstra2
 from circuit_maintenance_parser.parsers.turkcell import HtmlParserTurkcell1
 from circuit_maintenance_parser.parsers.verizon import HtmlParserVerizon1
@@ -428,6 +429,19 @@ class Sparkle(GenericProvider):
     _default_organizer = PrivateAttr("[email protected]")
 
 
+class Tata(GenericProvider):
+    """Tata provider custom class."""
+
+    _include_filter = PrivateAttr({EMAIL_HEADER_SUBJECT: ["Planned Work Notification"]})
+
+    _processors: List[GenericProcessor] = PrivateAttr(
+        [
+            CombinedProcessor(data_parsers=[HtmlParserTata, SubjectParserTata, EmailDateParser]),
+        ]
+    )
+    _default_organizer = PrivateAttr("[email protected]")
+
+
 class Telia(Arelion):
     """Telia provider custom class."""
 

diff --git a/tests/unit/data/tata/tata_body.html b/tests/unit/data/tata/tata_body.html
@@ -0,0 +1,102 @@
+<html><head>
+    <meta http-equiv="Content-Type" content="text/html; charset=utf-8">
+    </head>
+    <body>
+    <div style="background-color:#FFEB9C; width:100%; padding:2pt; font-size:10pt; line-height:12pt; font-family:'Calibri'; color:Black; text-align: left;">
+    <span style="color:#9C6500; font-weight:bold;">CAUTION:</span> This email has been sent from an external source. Do not click links, open attachments, or provide sensitive business information unless you can verify the sender’s legitimacy.
+    </div>
+    <br>
+    <div>
+    <p></p>
+    <p></p>
+    <p></p>
+    <p><span style="color: #000000; font-size: 10pt; font-family: tahoma, arial, helvetica, sans-serif;">Dear&nbsp;Customer,</span></p>
+    <p>&nbsp;</p>
+    <p><span style="color: #000000; font-size: 10pt; font-family: tahoma, arial, helvetica, sans-serif;">Tata Communication is in receipt of maintenance activity. The details of maintenance activity are as below. Your below mentioned service would experience an
+     outage of “8 Days 23 Hours 58 Minutes” during the activity window.</span></p>
+    <p>&nbsp;</p>
+    <table style="height: 42px; border-color: #000000;" border="0" width="623" cellspacing="0">
+    <tbody>
+    <tr style="background-color: #4b85c5;">
+    <td style="border: none; text-align: left; padding-left: 5px;"><span style="font-size: 11pt; color: #ffffff; font-family: helvetica;"><strong>TATA COMMUNICATIONS</strong></span></td>
+    <td style="border: 0.5px #bbbbbb;">&nbsp;<img style="align: baseline;" title="" src="cid:[email protected]" alt="" width="44" height="38" align="right" border="" hspace="10" vspace=""></td>
+    </tr>
+    </tbody>
+    </table>
+    <table style="height: 38px; border-color: #000000;" border="1" width="620" cellspacing="0">
+    <tbody>
+    <tr>
+    <td style="background-color: #4b85c5;" colspan="2"><span style="color: #ffffff; font-family: arial, helvetica, sans-serif; font-size: 12pt;">&nbsp;Planned Work Notification</span></td>
+    </tr>
+    <tr>
+    <td><span style="font-family: arial, helvetica, sans-serif; font-size: 10pt;"><strong><span style="color: black;">Ticket Reference - TCL</span></strong></span></td>
+    <td><span style="font-family: arial, helvetica, sans-serif; font-size: 10pt;">CHGP0123456</span></td>
+    </tr>
+    <tr>
+    <td><span style="font-family: arial, helvetica, sans-serif; font-size: 10pt;"><strong><span style="color: black;">Service ID</span></strong></span></td>
+    <td><span style="font-family: arial, helvetica, sans-serif; font-size: 10pt;">A012345678901234567, B012345678901234567</span></td>
+    </tr>
+    <tr>
+    <td><span style="font-family: arial, helvetica, sans-serif; font-size: 10pt;"><strong><span style="color: black;">NIMS ID</span></strong></span></td>
+    <td><span style="font-family: arial, helvetica, sans-serif; font-size: 10pt;">F0123-45678-901, F0123-45678-901</span></td>
+    </tr>
+    <tr>
+    <td><span style="font-family: arial, helvetica, sans-serif; font-size: 10pt;"><strong><strong><span style="color: black;">Maintenance Type</span></strong></strong></span></td>
+    <td><span style="font-family: arial, helvetica, sans-serif; font-size: 10pt;">Normal</span></td>
+    </tr>
+    <tr>
+    <td><span style="font-family: arial, helvetica, sans-serif; font-size: 10pt;"><strong><strong><span style="color: black;">Activity Status</span></strong></strong></span></td>
+    <td><span style="font-family: arial, helvetica, sans-serif; font-size: 10pt;">Scheduled</span></td>
+    </tr>
+    <tr>
+    <td><span style="font-family: arial, helvetica, sans-serif; font-size: 10pt;"><strong><strong><span style="color: black;">Execution Owner</span></strong></strong></span></td>
+    <td><span style="font-family: arial, helvetica, sans-serif; font-size: 10pt;">International Cables</span></td>
+    </tr>
+    <tr>
+    <td><span style="font-family: arial, helvetica, sans-serif; font-size: 10pt;"><strong><strong><span style="color: black;">Location of activity</span></strong></strong></span></td>
+    <td><span style="font-family: arial, helvetica, sans-serif; font-size: 10pt;">Mumbai</span></td>
+    </tr>
+    <tr>
+    <td style="background-color: #4b85c5;" colspan="2">&nbsp;</td>
+    </tr>
+    <tr>
+    <td><span style="font-family: arial, helvetica, sans-serif; font-size: 10pt;"><strong><strong><span style="color: black;">Activity Window (IST)</span></strong></strong></span></td>
+    <td><span style="font-family: arial, helvetica, sans-serif; font-size: 10pt;">2024-12-22 05:31:00 IST to 2024-12-31 05:29:00 IST</span></td>
+    </tr>
+    <tr>
+    <td><span style="font-family: arial, helvetica, sans-serif; font-size: 10pt;"><strong><strong><span style="color: black;">Activity Window (GMT)</span></strong></strong></span></td>
+    <td><span style="font-family: arial, helvetica, sans-serif; font-size: 10pt;">2024-12-22 00:01:00 GMT to 2024-12-30 23:59:00 GMT</span></td>
+    </tr>
+    <tr>
+    <td><span style="font-family: arial, helvetica, sans-serif; font-size: 10pt;"><strong><strong><span style="color: black;">Expected Impact Duration(DD:HH:MM) :</span></strong></strong></span></td>
+    <td><span style="font-family: arial, helvetica, sans-serif; font-size: 10pt;">8 Days 23 Hours 58 Minutes</span></td>
+    </tr>
+    <tr>
+    <td style="background-color: #4b85c5;" colspan="2">&nbsp;</td>
+    </tr>
+    <tr>
+    <td><span style="font-family: arial, helvetica, sans-serif; font-size: 10pt;"><strong><span style="color: black;">Activity Description</span></strong></span></td>
+    <td><span style="font-family: arial, helvetica, sans-serif; font-size: 10pt;">Planned activity for shunt fault repair of Seg 11 near Mumbai in TGNEA network.</span></td>
+    </tr>
+    <tr>
+    <td colspan="2">
+    <p><span style="font-family: arial, helvetica, sans-serif; font-size: 10pt;">We request you to reschedule sensitive operations at your end accordingly.<br>
+    We apologize for any inconvenience due to this event and resulting downtime.<br>
+    If you have any queries with respect to this activity, feel free to contact us on the coordinates mentioned below:<br>
+    Mail ID :<a href="mailto:[email protected]">[email protected]</a><br>
+    Contact Number : + 91 20 6614 4444 &amp; Toll Free no: 1-8002660660<br>
+    We look forward to your co-operation and a long term synergic association.<br>
+    Best Regards,<br>
+    Customer Service<br>
+    TATA COMMUNICATIONS</span></p>
+    </td>
+    </tr>
+    </tbody>
+    </table>
+    <p>&nbsp;</p>
+    <div>&nbsp;</div>
+    <div style="display:inline">Ref:MSGTCL000142929801</div>
+    </div>
+    </body>
+    </html>
+
diff --git a/tests/unit/data/tata/tata_body_result.json b/tests/unit/data/tata/tata_body_result.json
@@ -0,0 +1,19 @@
+[
+    {
+        "account": "N/A",
+        "circuits": [
+            {
+                "circuit_id": "A012345678901234567",
+                "impact": "OUTAGE"
+            },
+            {
+                "circuit_id": "B012345678901234567",
+                "impact": "OUTAGE"
+            }
+        ],
+        "end": 1735603140,
+        "maintenance_id": "CHGP0123456",
+        "organizer": "[email protected]",
+        "start": 1734825660
+    }
+]
diff --git a/tests/unit/data/tata/tata_cancellation_body.html b/tests/unit/data/tata/tata_cancellation_body.html
@@ -0,0 +1,103 @@
+<html><head>
+    <meta http-equiv="Content-Type" content="text/html; charset=utf-8">
+    </head>
+    <body>
+    <div style="background-color:#FFEB9C; width:100%; padding:2pt; font-size:10pt; line-height:12pt; font-family:'Calibri'; color:Black; text-align: left;">
+    <span style="color:#9C6500; font-weight:bold;">CAUTION:</span> This email has been sent from an external source. Do not click links, open attachments, or provide sensitive business information unless you can verify the sender’s legitimacy.
+    </div>
+    <br>
+    <div>
+    <p></p>
+    <p></p>
+    <p></p>
+    <p><span style="color: black; font-family: tahoma, arial, helvetica, sans-serif; font-size: 10pt;">Dear Customer,</span></p>
+    <p><span style="color: black; font-family: tahoma, arial, helvetica, sans-serif; font-size: 10pt;">Please note that this activity has been cancelled.
+    </span></p>
+    <p><span style="color: black; font-family: tahoma, arial, helvetica, sans-serif; font-size: 10pt;">The details for the same have been mentioned below:</span><strong><span style="color: black;">&nbsp;
+    </span></strong></p>
+    <table style="height: 42px; border-color: #000000;" border="0" width="623" cellspacing="0">
+    <tbody>
+    <tr style="background-color: #4b85c5;">
+    <td style="border: none; text-align: left; padding-left: 5px;"><span style="font-size: 11pt; color: #ffffff; font-family: helvetica;"><strong>TATA COMMUNICATIONS</strong></span></td>
+    <td style="border: 0.5px #bbbbbb;">&nbsp;<img style="align: baseline;" title="" src="cid:[email protected]" alt="" width="44" height="38" align="right" border="" hspace="10" vspace=""></td>
+    </tr>
+    </tbody>
+    </table>
+    <table style="height: 38px; border-color: #000000;" border="1" width="620" cellspacing="0">
+    <tbody>
+    <tr>
+    <td style="background-color: #4b85c5;" colspan="2"><span style="color: #ffffff; font-family: arial, helvetica, sans-serif; font-size: 12pt;">&nbsp;Planned Work Notification</span></td>
+    </tr>
+    <tr>
+    <td><span style="font-family: arial, helvetica, sans-serif; font-size: 10pt;"><strong><span style="color: black;">Ticket Reference - TCL</span></strong></span></td>
+    <td><span style="font-family: arial, helvetica, sans-serif; font-size: 10pt;">CHGP0123456</span></td>
+    </tr>
+    <tr>
+    <td><span style="font-family: arial, helvetica, sans-serif; font-size: 10pt;"><strong><span style="color: black;">Service ID</span></strong></span></td>
+    <td><span style="font-family: arial, helvetica, sans-serif; font-size: 10pt;">01234567890123456789</span></td>
+    </tr>
+    <tr>
+    <td><span style="font-family: arial, helvetica, sans-serif; font-size: 10pt;"><strong><span style="color: black;">NIMS ID</span></strong></span></td>
+    <td><span style="font-family: arial, helvetica, sans-serif; font-size: 10pt;"></span></td>
+    </tr>
+    <tr>
+    <td><span style="font-family: arial, helvetica, sans-serif; font-size: 10pt;"><strong><strong><span style="color: black;">Maintenance Type</span></strong></strong></span></td>
+    <td><span style="font-family: arial, helvetica, sans-serif; font-size: 10pt;">Normal</span></td>
+    </tr>
+    <tr>
+    <td><span style="font-family: arial, helvetica, sans-serif; font-size: 10pt;"><strong><strong><span style="color: black;">Activity Status</span></strong></strong></span></td>
+    <td><span style="font-family: arial, helvetica, sans-serif; font-size: 10pt;">Canceled</span></td>
+    </tr>
+    <tr>
+    <td><span style="font-family: arial, helvetica, sans-serif; font-size: 10pt;"><strong><strong><span style="color: black;">Execution Owner</span></strong></strong></span></td>
+    <td><span style="font-family: arial, helvetica, sans-serif; font-size: 10pt;">3rd party service Provider</span></td>
+    </tr>
+    <tr>
+    <td><span style="font-family: arial, helvetica, sans-serif; font-size: 10pt;"><strong><strong><span style="color: black;">Location of activity</span></strong></strong></span></td>
+    <td><span style="font-family: arial, helvetica, sans-serif; font-size: 10pt;">LONDON</span></td>
+    </tr>
+    <tr>
+    <td style="background-color: #4b85c5;" colspan="2">&nbsp;</td>
+    </tr>
+    <tr>
+    <td><span style="font-family: arial, helvetica, sans-serif; font-size: 10pt;"><strong><strong><span style="color: black;">Activity Window (IST)</span></strong></strong></span></td>
+    <td><span style="font-family: arial, helvetica, sans-serif; font-size: 10pt;">2024-11-21 04:30:00 IST to 2024-11-22 10:30:00 IST</span></td>
+    </tr>
+    <tr>
+    <td><span style="font-family: arial, helvetica, sans-serif; font-size: 10pt;"><strong><strong><span style="color: black;">Activity Window (GMT)</span></strong></strong></span></td>
+    <td><span style="font-family: arial, helvetica, sans-serif; font-size: 10pt;">2024-11-20 23:00:00 GMT to 2024-11-22 05:00:00 GMT</span></td>
+    </tr>
+    <tr>
+    <td><span style="font-family: arial, helvetica, sans-serif; font-size: 10pt;"><strong><strong><span style="color: black;">Expected Impact Duration(DD:HH:MM) :</span></strong></strong></span></td>
+    <td><span style="font-family: arial, helvetica, sans-serif; font-size: 10pt;">6 Hours</span></td>
+    </tr>
+    <tr>
+    <td style="background-color: #4b85c5;" colspan="2">&nbsp;</td>
+    </tr>
+    <tr>
+    <td><span style="font-family: arial, helvetica, sans-serif; font-size: 10pt;"><strong><span style="color: black;">Activity Description</span></strong></span></td>
+    <td><span style="font-family: arial, helvetica, sans-serif; font-size: 10pt;">&nbsp;Third party service provider will execute network maintenance activity for Network redevelopment.<br>
+    Services will be impacted for 6 hrs on each Days.</span></td>
+    </tr>
+    <tr>
+    <td colspan="2">
+    <p><span style="font-family: arial, helvetica, sans-serif; font-size: 10pt;">We request you to reschedule sensitive operations at your end accordingly.<br>
+    We apologize for any inconvenience due to this event and resulting downtime.<br>
+    If you have any queries with respect to this activity, feel free to contact us on the coordinates mentioned below:<br>
+    Mail ID :<a href="mailto:[email protected]">[email protected]</a><br>
+    Contact Number : + 91 20 6614 4444 &amp; Toll Free no: 1-8002660660<br>
+    We look forward to your co-operation and a long term synergic association.<br>
+    Best Regards,<br>
+    Customer Service<br>
+    TATA COMMUNICATIONS</span></p>
+    </td>
+    </tr>
+    </tbody>
+    </table>
+    <p>&nbsp;</p>
+    <div>&nbsp;</div>
+    <div style="display:inline">Ref:MSGTCL000143017654</div>
+    </div>
+    </body>
+    </html>
+
diff --git a/tests/unit/data/tata/tata_cancellation_body_result.json b/tests/unit/data/tata/tata_cancellation_body_result.json
@@ -0,0 +1,15 @@
+[
+    {
+        "account": "N/A",
+        "circuits": [
+            {
+                "circuit_id": "01234567890123456789",
+                "impact": "OUTAGE"
+            }
+        ],
+        "end": 1732251600,
+        "maintenance_id": "CHGP0123456",
+        "organizer": "[email protected]",
+        "start": 1732143600
+    }
+]
diff --git a/tests/unit/data/tata/tata_cancellation_subject.txt b/tests/unit/data/tata/tata_cancellation_subject.txt
@@ -0,0 +1 @@
+Cancellation Notification /TCL TT: "CHGP0123456" / "Normal" / "Canceled" / Planned Work Notification
diff --git a/tests/unit/data/tata/tata_cancellation_subject_result.json b/tests/unit/data/tata/tata_cancellation_subject_result.json
@@ -0,0 +1,6 @@
+[
+    {
+        "status": "CANCELLED",
+        "summary": "Cancellation Notification /TCL TT: \"CHGP0123456\" / \"Normal\" / \"Canceled\" / Planned Work Notification"
+    }
+]
-Original file line number
+Diff line change
@@ Expand Up @@
     - Netflix (AS2906 only)
     - Seaborn
     - Sparkle
+    - Tata
     - Telstra (\*)
     - Turkcell
     - Verizon
@@ Expand Down @@
Original file line number	Diff line number	Diff line change
		@@ -0,0 +1 @@
		Cancellation Notification /TCL TT: "CHGP0123456" / "Normal" / "Canceled" / Planned Work Notification