Merge pull request #82 from smkent/selenium

Perform Safeway account sign in using undetected-chromedriver

Dockerfile

@@ -1,4 +1,6 @@
-FROM python:3-alpine
+FROM busybox
+
+FROM python:3
 ARG POETRY_DYNAMIC_VERSIONING_BYPASS="0.0.0"
 ENV CRON_SCHEDULE "5 2 * * *"
 ENV SMTPHOST=
@@ -7,12 +9,25 @@ ENV SAFEWAY_ACCOUNT_PASSWORD=
 ENV SAFEWAY_ACCOUNT_MAIL_FROM=
 ENV SAFEWAY_ACCOUNT_MAIL_TO=
 ENV SAFEWAY_ACCOUNTS_FILE=
+ENV DEBUG_DIR="/debug"
+
+RUN DEBIAN_FRONTEND=noninteractive && \
+    wget -q -O - https://dl-ssl.google.com/linux/linux_signing_key.pub > /usr/share/keyrings/chrome.pub && \
+    echo 'deb [arch=amd64 signed-by=/usr/share/keyrings/chrome.pub] http://dl.google.com/linux/chrome/deb/ stable main' > /etc/apt/sources.list.d/google-chrome.list && \
+    apt update -y && \
+    apt install -y google-chrome-stable
+RUN apt install -y tini
+
+# Install busybox utilities using static binary from official image
+COPY --from=busybox /bin/busybox /bin/busybox
+RUN for target in /usr/sbin/sendmail /usr/sbin/crond /usr/bin/crontab; do \
+    ln -svf /bin/busybox ${target}; \
+    done
 
-RUN apk add --no-cache tini
 COPY docker/entrypoint /
 
 COPY . /python-build
 RUN python3 -m pip install /python-build && rm -rf /python-build
 
-ENTRYPOINT ["/sbin/tini", "--"]
+ENTRYPOINT ["/usr/bin/tini", "--"]
 CMD ["/entrypoint"]

README.md

@@ -10,6 +10,17 @@
 and attempt to select all of the "Safeway for U" electronic coupons on the site
 so they don't have to each be clicked manually.
 
+## Design notes
+
+Safeway's sign in page is protected by a web application firewall (WAF).
+safeway-coupons performs authentication using a headless instance of Google
+Chrome. Authentication may fail based on your IP's reputation, either by
+presenting a CAPTCHA or denying sign in attempts altogether. safeway-coupons
+currently does not have support for prompting the user to solve CAPTCHAs.
+
+Once a signed in session is established, coupon clipping is performed using HTTP
+requests via [requests][requests].
+
 ## Installation and usage with Docker
 
 A Docker container is provided which runs safeway-coupons with cron. The cron
@@ -65,18 +76,23 @@ docker-compose logs -f
 
 ## Installation from PyPI
 
+### Prerequisites
+
+* Google Chrome (for authentication performed via Selenium).
+* Optional: `sendmail` (for email support)
+
+### Installation
+
 [safeway-coupons is available on PyPI][pypi]:
 
 ```console
 pip install safeway-coupons
 ```
 
-`sendmail` is needed for email support.
+### Usage
 
 For best results, run this program once a day or so with a cron daemon.
 
-### Usage
-
 For full usage options, run
 
 ```console
@@ -163,3 +179,4 @@ Created from [smkent/cookie-python][cookie-python] using
 [poetry]: https://python-poetry.org/docs/#installation
 [pypi]: https://pypi.org/project/safeway-coupons/
 [repo]: https://github.com/smkent/safeway-coupons
+[requests]: https://requests.readthedocs.io/en/latest/

docker/entrypoint

@@ -21,11 +21,15 @@ else
     env | grep -vie 'password' | grep -e '^SAFEWAY_' -e '^SMTPHOST='
 fi
 
-mkdir /etc/cron.d
+if [ -d "${DEBUG_DIR}" ]; then
+    args="${args} --debug-dir ${DEBUG_DIR}"
+fi
+
+mkdir -vp /etc/cron.d /var/spool/cron/crontabs
 (
     echo "${CRON_SCHEDULE?} safeway-coupons ${args} >/proc/1/fd/1 2>/proc/1/fd/2"
 ) > /var/spool/cron/crontabs/root
 
-crontab -l
+busybox crontab -l
 
-exec crond -f -d 8
+exec busybox crond -f

pyproject.toml

@@ -21,6 +21,9 @@ classifiers = [
 python = "^3.8"
 dataclasses-json = "*"
 requests = "*"
+selenium = "^4.10"
+webdriver-manager = "^3.8.6"
+undetected-chromedriver = "^3.5.0"
 
 [tool.poetry.dev-dependencies]
 bandit = {extras = ["toml"], version = "*"}

safeway_coupons/app.py

@@ -1,6 +1,7 @@
 import argparse
 import sys
 from http.client import HTTPConnection
+from pathlib import Path
 
 from .config import Config
 from .safeway import SafewayCoupons
@@ -19,6 +20,17 @@ def _parse_args() -> argparse.Namespace:
             "accounts information"
         ),
     )
+    arg_parser.add_argument(
+        "-D",
+        "--debug-dir",
+        dest="debug_dir",
+        metavar="directory",
+        default=".",
+        help=(
+            "Destination directory for debug output files, "
+            "such as browser screenshots (default: %(default)s)"
+        ),
+    )
     arg_parser.add_argument(
         "-d",
         "--debug",
@@ -79,6 +91,7 @@ def main() -> None:
     sc = SafewayCoupons(
         send_email=args.send_email,
         debug_level=args.debug_level,
+        debug_dir=Path(args.debug_dir) if args.debug_dir else None,
         sleep_level=args.sleep_level,
         dry_run=args.dry_run,
         max_clip_count=args.max_clip_count,

safeway_coupons/client.py

@@ -1,5 +1,6 @@
 import json
 import random
+from pathlib import Path
 from typing import List, Optional
 
 import requests
@@ -12,8 +13,8 @@
 
 
 class SafewayClient(BaseSession):
-    def __init__(self, account: Account) -> None:
-        self.session = LoginSession(account)
+    def __init__(self, account: Account, debug_dir: Optional[Path]) -> None:
+        self.session = LoginSession(account, debug_dir)
         self.requests.headers.update(
             {
                 "Authorization": f"Bearer {self.session.access_token}",

safeway_coupons/email.py

@@ -1,7 +1,9 @@
 import collections
+import mimetypes
 import os
 import subprocess
-from email.mime.text import MIMEText
+from email.message import EmailMessage
+from pathlib import Path
 from typing import List, Optional
 
 from .accounts import Account
@@ -15,6 +17,7 @@ def _send_email(
     mail_message: List[str],
     debug_level: int,
     send_email: bool,
+    attachments: Optional[List[Path]] = None,
 ) -> None:
     mail_message_str = os.linesep.join(mail_message)
     if debug_level >= 1:
@@ -27,16 +30,26 @@ def _send_email(
         print("<<<<<<")
     if not send_email:
         return
-    email_data = MIMEText(mail_message_str)
-    email_data["To"] = account.mail_to
-    email_data["From"] = account.mail_from
+    msg = EmailMessage()
+    msg["To"] = account.mail_to
+    msg["From"] = account.mail_from
     if subject:
-        email_data["Subject"] = subject
+        msg["Subject"] = subject
+    msg.set_content(mail_message_str)
+    for attachment in attachments or []:
+        mt = mimetypes.guess_type(attachment.name)[0]
+        main, sub = mt.split("/", 1) if mt else ("application", "octet-stream")
+        msg.add_attachment(
+            attachment.read_bytes(),
+            filename=attachment.name,
+            maintype=main,
+            subtype=sub,
+        )
     p = subprocess.Popen(
         ["/usr/sbin/sendmail", "-f", account.mail_to, "-t"],
         stdin=subprocess.PIPE,
     )
-    p.communicate(bytes(email_data.as_string(), "UTF-8"))
+    p.communicate(bytes(msg.as_string(), "UTF-8"))
 
 
 def email_clip_results(
@@ -77,4 +90,11 @@ def email_error(
         mail_message += ["Clipped coupons:", ""]
         for offer in error.clipped_offers:
             mail_message += str(offer)
-    _send_email(account, mail_subject, mail_message, debug_level, send_email)
+    _send_email(
+        account,
+        mail_subject,
+        mail_message,
+        debug_level,
+        send_email,
+        attachments=getattr(error, "attachments", None),
+    )

safeway_coupons/errors.py

@@ -1,4 +1,5 @@
 from dataclasses import dataclass
+from pathlib import Path
 from typing import List, Optional
 
 import requests
@@ -15,6 +16,7 @@ class Error(Exception):
 @dataclass
 class AuthenticationFailure(Error):
     account: Account
+    attachments: Optional[List[Path]] = None
 
     def __str__(self) -> str:
         return f"Authentication Failure ({self.exception})"

safeway_coupons/safeway.py

@@ -1,4 +1,5 @@
-from typing import List
+from pathlib import Path
+from typing import List, Optional
 
 from .accounts import Account
 from .client import SafewayClient
@@ -15,24 +16,26 @@ def __init__(
         self,
         send_email: bool = True,
         debug_level: int = 0,
+        debug_dir: Optional[Path] = None,
         sleep_level: int = 0,
         dry_run: bool = False,
         max_clip_count: int = 0,
         max_clip_errors: int = CLIP_ERROR_MAX,
     ) -> None:
         self.send_email = send_email
         self.debug_level = debug_level
+        self.debug_dir = debug_dir
         self.sleep_level = sleep_level
         self.dry_run = dry_run
         self.max_clip_count = max_clip_count
         self.max_clip_errors = max_clip_errors
 
     def clip_for_account(self, account: Account) -> None:
         print(f"Clipping coupons for Safeway account {account.username}")
-        swy = SafewayClient(account)
-        clipped_offers: List[Offer] = []
-        clip_errors: List[ClipError] = []
         try:
+            swy = SafewayClient(account, self.debug_dir)
+            clipped_offers: List[Offer] = []
+            clip_errors: List[ClipError] = []
             offers = swy.get_offers()
             unclipped_offers = [
                 o for o in offers if o.status == OfferStatus.Unclipped