disable starttls if tls is enabled

update node & actual

.pre-commit-config.yaml

@@ -1,14 +1,14 @@
 # .pre-commit-config.yaml
 repos:
   - repo: https://github.com/astral-sh/ruff-pre-commit
-    rev: v0.1.5  # Check for the latest version
+    rev: v0.15.12  # Check for the latest version
     hooks:
       - id: ruff
         args: ["check", "--select", "I", "--fix"]
       - id: ruff-format
 
   - repo: https://github.com/pre-commit/pre-commit-hooks
-    rev: v4.5.0  # Check for the latest version
+    rev: v6.0.0  # Check for the latest version
     hooks:
       - id: check-yaml
       - id: end-of-file-fixer
@@ -16,6 +16,6 @@ repos:
       - id: requirements-txt-fixer
 
   - repo: https://github.com/PyCQA/flake8
-    rev: 6.0.0  # Check for the latest version
+    rev: 7.3.0  # Check for the latest version
     hooks:
       - id: flake8

Dockerfile

@@ -5,7 +5,7 @@ RUN apk add py3-uv
 RUN uv build
 
 # Use an official Node.js runtime as a base
-FROM node:20-alpine AS runtime
+FROM node:25-alpine AS runtime
 
 # Create app directory
 WORKDIR /app

actual_imap_poll/model.py

@@ -1,5 +1,5 @@
+import datetime
 from dataclasses import dataclass
-from datetime import date
 from decimal import Decimal
 from uuid import UUID
 
@@ -12,7 +12,7 @@ class Transaction:
     """
 
     account: UUID
-    date: date
+    date: datetime.date
     amount: Decimal  # Note: decimal dollars, JS shim will convert to cents as described in the API
     payee: str  # imported_payee in API
     notes: str

actual_imap_poll/parsers.py

@@ -1,16 +1,19 @@
+import logging
 import re
 from abc import ABC, abstractmethod
+from collections.abc import Sequence
 from datetime import date, datetime
 from decimal import Decimal
 from email.message import EmailMessage, Message
-from logging import info, warning
+from typing import Any
-from typing import Any, Optional, Sequence
 from uuid import UUID
 
 from bs4 import BeautifulSoup
 
 from .model import Transaction
 
+log = logging.getLogger("parse")
+
 
 def parse_email_time(s: str) -> datetime:
     return datetime.strptime(s, "%a, %d %b %Y %H:%M:%S %z")
@@ -28,10 +31,10 @@ class TransactionParser(ABC):
         Returns:
             bool: True if the message matches the parser's criteria, False otherwise.
         """
-        pass
+        ...
 
     @abstractmethod
-    def extract(self, msg: EmailMessage) -> Optional[Transaction]:
+    def extract(self, msg: EmailMessage) -> Transaction | None:
         """
         Extracts transaction details from the given email message.
 
@@ -44,7 +47,7 @@ class TransactionParser(ABC):
         Raises:
             TransactionParsingFailed: If the message cannot be parsed successfully.
         """
-        pass
+        ...
 
     @staticmethod
     def get_content(
@@ -58,6 +61,18 @@ class TransactionParser(ABC):
             raise TransactionParsingFailed("No content of message found")
         return content
 
+    @staticmethod
+    def strip_html(msg: EmailMessage) -> str:
+        body = msg.get_body(preferencelist=("html", "plain"))
+        if body is None:
+            raise TransactionParsingFailed("No HTML body of message found")
+        content = body.get_content()
+        if content is None:
+            raise TransactionParsingFailed("No content of message found")
+
+        soup = BeautifulSoup(content, "html.parser")
+        return soup.get_text()
+
 
 class TransactionParsingFailed(Exception):
     pass
@@ -65,7 +80,7 @@ class TransactionParsingFailed(Exception):
 
 class RogersBankParser(TransactionParser):
     EXTRACT_RE = re.compile(
-        r"Attempt of \$([0-9,]+\.\d{2}) was made on ([A-z]{3} \d{1,2}, \d{4})[^<]*at ([^<]+) in ([^<]+)."  # noqa: E501
+        r"\$([0-9,]+\.\d{2}) on ((Jan|Feb|Mar|Apr|May|Jun|Jul|Aug|Sep|Oct|Nov|Dec) [0-9]{2}, [0-9]{4}) on your credit card ending in ([0-9]{4}). Details: ([^.]+)"  # noqa: E501
     )
 
     def __init__(self, account_id: UUID):
@@ -77,18 +92,17 @@ class RogersBankParser(TransactionParser):
             and msg["Subject"] == "Purchase amount alert"
         )
 
-    def extract(self, msg: EmailMessage) -> Optional[Transaction]:
+    def extract(self, msg: EmailMessage) -> Transaction | None:
-        content = self.get_content(msg)
+        content = self.strip_html(msg)
         matches = self.EXTRACT_RE.search(content)
         if matches is None:
             raise TransactionParsingFailed("No matches for extraction RE")
         amount = Decimal(matches[1].replace(",", "")) * -1
 
         date_raw = matches[2]
-        payee = matches[3]
+        payee = matches[5]
-        location = matches[4]
 
-        if "Rebate" == location and "CashBack" in payee:
+        if "CashBack" in payee:
             amount = amount * -1
 
         date = datetime.strptime(date_raw, "%b %d, %Y").date()
@@ -97,7 +111,7 @@ class RogersBankParser(TransactionParser):
             date=date,
             amount=amount,
             payee=payee,
-            notes=f"in {location} (via email)",
+            notes="(via email)",
             imported_id=msg["Message-ID"],
         )
 
@@ -116,7 +130,7 @@ class MBNAParser(TransactionParser):
             and msg["Subject"] == "MBNA - Transaction Alert"
         )
 
-    def extract(self, msg: EmailMessage) -> Optional[Transaction]:
+    def extract(self, msg: EmailMessage) -> Transaction | None:
         content = self.get_content(msg)
         matches = self.EXTRACT_RE.search(content)
         if matches is None:
@@ -136,8 +150,8 @@ class MBNAParser(TransactionParser):
 
 class BMOParser(TransactionParser):
     EXTRACT_RE = re.compile(
-        r"We want to let you know that a (withdrawal|deposit) of\s+\$([0-9,]+\.\d{2})\s+has been made (?:to|from) your account ending\s+in\s+(\d{3})",  # noqa: E501
+        r"There was a (withdrawal|deposit).*Amount:\s*\$([0-9,]+\.\d{2}).*Account:\s*Ending in ([0-9]+)",  # noqa: E501
-        flags=re.MULTILINE,
+        flags=re.DOTALL,
     )
 
     def __init__(self, account_map: dict[int, UUID]):
@@ -146,10 +160,9 @@ class BMOParser(TransactionParser):
     def match(self, msg: Message) -> bool:
         return msg["From"] == "bmoalerts@bmo.com"
 
-    def extract(self, msg: EmailMessage) -> Optional[Transaction]:
+    def extract(self, msg: EmailMessage) -> Transaction | None:
-        content = self.get_content(msg)
+        content = self.strip_html(msg)
-        soup = BeautifulSoup(content, "html.parser")
+        matches = self.EXTRACT_RE.search(content)
-        matches = self.EXTRACT_RE.search(soup.get_text())
         if matches is None:
             raise TransactionParsingFailed("No matches for extraction RE")
 
@@ -159,7 +172,9 @@ class BMOParser(TransactionParser):
         date = parse_email_time(msg["Date"]).date()
         account_ref = int(matches[3])
         if account_ref not in self._account_map:
-            warning("Account %s not in account map, skipping transaction", account_ref)
+            log.warning(
+                "Account %s not in account map, skipping transaction", account_ref
+            )
             return None
         account_id = self._account_map[account_ref]
         return Transaction(
@@ -178,11 +193,16 @@ class CIBCParser(TransactionParser):
         flags=re.MULTILINE,
     )
 
+    PURCHASE_EXTRACT_RE = re.compile(
+        r"made a purchase with your CIBC.*ending in (?P<account>\d{4}) for \$(?P<amount>\d+\.\d{2}) at (?P<payee>.*?).<br",  # noqa: E501
+        re.MULTILINE,
+    )
+
     def __init__(self, account_map: dict[int, UUID]):
         self._account_map = account_map
 
     def match(self, msg: Message) -> bool:
-        return msg["From"] == "CIBC Banking <mailbox.noreply@cibc.com>"
+        return msg["From"] == "CIBC Banking <Mailbox.noreply@cibc.com>"
 
     def extract_payment(self, msg: EmailMessage):
         content = self.get_content(msg)
@@ -196,7 +216,9 @@ class CIBCParser(TransactionParser):
         date = parse_email_time(msg["Date"]).date()
 
         if account_ref not in self._account_map:
-            warning("Account %s not in account map, skipping transaction", account_ref)
+            log.warning(
+                "Account %s not in account map, skipping transaction", account_ref
+            )
             return None
         account_id = self._account_map[account_ref]
 
@@ -209,10 +231,39 @@ class CIBCParser(TransactionParser):
             imported_id=msg["Message-ID"],
         )
 
-    def extract(self, msg: EmailMessage) -> Optional[Transaction]:
+    def extract_purchase(self, msg: EmailMessage) -> Transaction | None:
+        content = self.get_content(msg)
+
+        matches = self.PURCHASE_EXTRACT_RE.search(content)
+        if matches is None:
+            raise TransactionParsingFailed("no matches for extraction RE")
+
+        amount = Decimal(matches["amount"].replace(",", "")) * -1
+        date = parse_email_time(msg["Date"]).date()
+        account_ref = int(matches["account"])
+
+        if account_ref not in self._account_map:
+            log.warning(
+                "Account %s not in account map, skipping transaction", account_ref
+            )
+            return None
+        account_id = self._account_map[account_ref]
+
+        return Transaction(
+            account=account_id,
+            date=date,
+            amount=amount,
+            payee=matches["payee"],
+            notes="via email",
+            imported_id=msg["Message-ID"],
+        )
+
+    def extract(self, msg: EmailMessage) -> Transaction | None:
         match msg["Subject"]:
             case "New payment to your credit card":
                 return self.extract_payment(msg)
+            case "New purchase on your credit card":
+                return self.extract_purchase(msg)
         return None
 
 
@@ -228,18 +279,18 @@ class ScotiaBankParser(TransactionParser):
     def match(self, msg: Message) -> bool:
         return msg["From"] == "Scotia InfoAlerts <infoalerts@scotiabank.com>"
 
-    def extract_transaction(self, msg: EmailMessage) -> Optional[Transaction]:
+    def extract_transaction(self, msg: EmailMessage) -> Transaction | None:
         content = self.get_content(msg)
 
         matches = self.PAYMENT_EXTRACT_RE.search(content)
         if matches is None:
             raise TransactionParsingFailed("no matches for extraction RE")
 
-        amount = Decimal(matches["amount"].replace(",", ""))
+        amount = Decimal(matches["amount"].replace(",", "")) * -1
         date = parse_email_time(msg["Date"]).date()
 
         if matches["account"] not in self._account_map:
-            warning(
+            log.warning(
                 "Account %s not in account map, skipping transaction",
                 matches["account"],
             )
@@ -257,12 +308,51 @@ class ScotiaBankParser(TransactionParser):
             imported_id=msg["Message-ID"],
         )
 
-    def extract(self, msg: EmailMessage) -> Optional[Transaction]:
+    def extract(self, msg: EmailMessage) -> Transaction | None:
         match msg["Subject"]:
             case (
                 "Authorization on your credit account"
                 | "Authorization without credit card present"
             ):
                 return self.extract_transaction(msg)
-        info("Subject `%s` didn't match any extractors", msg["Subject"])
+        log.info("Subject `%s` didn't match any extractors", msg["Subject"])
+        return None
+
+
+class NeoFinancialParser(TransactionParser):
+    PURCHASE_EXTRACT_RE = re.compile(
+        r"your purchase of \$(?P<amount>[0-9]+\.[0-9]{2}) at (?P<payee>.+) on (?P<date>(January|February|March|April|May|June|July|August|September|October|November|December) \d{1,2}(st|nd|rd|th)?, \d{4})",  # noqa: E501
+        re.MULTILINE,
+    )
+
+    def __init__(self, account_id: UUID):
+        self.account_id = account_id
+
+    def match(self, msg: Message) -> bool:
+        return msg["From"] == "Neo Financial <info@neofinancial.com>" and msg[
+            "Subject"
+        ].startswith("You earned")
+
+    def extract_purchase(self, msg: EmailMessage) -> Transaction | None:
+        content = self.get_content(msg, preferencelist=["plain"])
+        matches = self.PURCHASE_EXTRACT_RE.search(content)
+        if matches is None:
+            log.debug(f"No match on\n\n{content}")
+            raise TransactionParsingFailed("no matches for purchase extraction RE")
+
+        amount = Decimal(matches["amount"].replace(",", "")) * -1
+        # Use date from email header, it seems to match the body and includes TZ info
+        date = parse_email_time(msg["Date"]).date()
+        return Transaction(
+            account=self.account_id,
+            date=date,
+            amount=amount,
+            payee=matches["payee"],
+            notes="via email",
+            imported_id=msg["Message-ID"],
+        )
+
+    def extract(self, msg: EmailMessage) -> Transaction | None:
+        if msg["Subject"].startswith("You earned"):
+            return self.extract_purchase(msg)
         return None

actual_imap_poll/watcher.py

@@ -1,12 +1,11 @@
-#!/usr/bin/env python
 import asyncio
 import email.policy
+import importlib.metadata
 import importlib.util
 import logging
 import os
 import ssl
-from imaplib import IMAP4
+from imaplib import IMAP4, IMAP4_SSL
-from logging import debug, error, info, warning
 from os import getenv
 
 from pydantic import Field, SecretStr
@@ -19,6 +18,8 @@ CONFIG_PATH = getenv("CONFIG_PATH", "/data/config.py")
 ACTUAL_PATH = "./cli.js"
 TIMEOUT = 30
 
+log = logging.getLogger("watch")
+
 
 async def ticker(interval: float):
     while True:
@@ -42,7 +43,8 @@ class AppConfig(BaseSettings):
     imap_port: int = Field(default=143)
     imap_mailbox: str = Field(default="INBOX")
     imap_interval: int = Field(default=300)
-    imap_starttls: bool = Field(default=True)
+    imap_starttls: bool = Field(default_factory=lambda data: not data["imap_tls"])
+    imap_tls: bool = Field(default_factory=lambda data: data["imap_port"] == 993)
 
     model_config = {"env_prefix": ""}
 
@@ -62,7 +64,7 @@ class App:
             + f' -d "{t.date}"'
             + (f' -n "{t.notes}"' if t.notes else "")
         )
-        debug("Actual command: %s", cmd)
+        log.debug("Actual command: %s", cmd)
         proc = await asyncio.create_subprocess_shell(
             cmd=cmd,
             env=os.environ,
@@ -71,12 +73,12 @@ class App:
         )
         stdout, _stderr = await proc.communicate()
         if proc.returncode != 0:
-            error("Submitting to actual failed: %s", stdout)
+            log.error("Submitting to actual failed: %s", stdout)
 
     async def process_message(self, msg_b: bytes):
-        debug("parsing message")
+        log.debug("parsing message")
         msg = email.message_from_bytes(msg_b, policy=email.policy.default)
-        info(
+        log.info(
             "Found message from %s to %s subject %s",
             msg.get("From", "<unknown>"),
             msg.get("To", "<unknown>"),
@@ -84,24 +86,27 @@ class App:
         )
 
         for parser in self._parsers:
-            debug("Running parser %s", type(parser).__name__)
+            log.debug("Running parser %s", type(parser).__name__)
             try:
                 if parser.match(msg):
-                    info("Parser %s claimed message", type(parser).__name__)
+                    log.info("Parser %s claimed message", type(parser).__name__)
                     trans = parser.extract(msg)
                     if trans is not None:
-                        info("Submitting transaction to Actual: %s", trans)
+                        log.info("Submitting transaction to Actual: %s", trans)
                         await self.submit_transaction(trans)
                     else:
-                        warning("Parser %s returned None", type(parser).__name__)
+                        log.warning("Parser %s returned None", type(parser).__name__)
             except TransactionParsingFailed as e:
-                warning("Unable to parse message %s", e)
+                log.warning("Unable to parse message %s", e)
-            except Exception as e:
+            except Exception as e:  # noqa: BLE001 - trap all errors in the parsers
-                warning("Unexpected exception %s", e)
+                log.warning("Unexpected exception %s", e)
 
     async def poll_imap(self):
-        info("polling mailbox")
+        log.info("polling mailbox")
-        with IMAP4(self._config.imap_server, self._config.imap_port, TIMEOUT) as M:
+        imap_class = IMAP4_SSL if self._config.imap_tls else IMAP4
+        with imap_class(
+            self._config.imap_server, self._config.imap_port, timeout=TIMEOUT
+        ) as M:
             if self._config.imap_starttls:
                 context = ssl.create_default_context()
                 M.starttls(context)
@@ -110,16 +115,16 @@ class App:
             _status, m_set = M.search(None, "UNSEEN")
 
             for msg_id in m_set[0].split():
-                debug("Retrieving msg id %s", msg_id)
+                log.debug("Retrieving msg id %s", msg_id)
                 _status, msg = M.fetch(msg_id, "(RFC822)")
                 if _status != "OK" or msg[0] is None:
-                    error("Unable to fetch message %s", msg_id)
+                    log.error("Unable to fetch message %s", msg_id)
                     continue
                 msg_body = msg[0][1]
                 if isinstance(msg_body, int):
-                    error("Unable to fetch message %s", msg_id)
+                    log.error("Unable to fetch message %s", msg_id)
                     continue
-                debug("Processing message %s", msg_id)
+                log.debug("Processing message %s", msg_id)
                 await self.process_message(msg_body)
 
     async def run(self):
@@ -129,6 +134,11 @@ class App:
 
 def main():
     logging.basicConfig(level=getenv("LOG_LEVEL", "INFO"))
+    log.info(
+        "actual-imap-parser %s starting up",
+        importlib.metadata.version("actual-imap-poll"),
+    )
     app_config = AppConfig()
+    log.debug("Config: %s", app_config)
     app = App(app_config, CONFIG_PATH)
     asyncio.run(app.run())

package.json

@@ -14,7 +14,7 @@
   "license": "ISC",
   "type": "commonjs",
   "dependencies": {
-    "@actual-app/api": "^25.11.0",
+    "@actual-app/api": "^26.5.2",
     "yargs": "^17.7.2"
   }
 }

pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "actual-imap-poll"
-version = "0.1.1"
+version = "0.2.1"
 description = "Poll an IMAP mailbox looking for transactions to submit to Actual Budget"
 authors = [
     {name = "Keenan Tims",email = "ktims@gotroot.ca"}