Merge pull request #1 from dmdhrumilmistry/implement-mass-check

implement async email checker

.gitignore

@@ -184,3 +184,7 @@ poetry.toml
 
 # LSP config files
 pyrightconfig.json
+
+# breach-check data
+*.txt
+output_*.json

breach_check/__main__.py

@@ -0,0 +1,33 @@
+from argparse import ArgumentParser
+from breach_check.breach import BreachChecker
+from breach_check.logger import console
+from breach_check.utils import generate_unique_filename, extract_emails, write_json_file
+from asyncio import run
+from sys import exit
+
+
+def main():
+    parser = ArgumentParser('breach-check')
+    parser.add_argument('-i', '--input', dest='input_file',
+                        help='input file containing emails on each line', type=str, required=True)
+    parser.add_argument('-o', '--output', dest='output_file',
+                        help='output json file path', required=False, default=generate_unique_filename(), type=str)
+
+    args = parser.parse_args()
+
+    emails = extract_emails(args.input_file)
+
+    if not emails:
+        exit(-1)
+
+    results = run(
+        BreachChecker().mass_check(emails=emails)
+    )
+
+    if not write_json_file(args.output_file, results):
+        console.print('Results:')
+        console.print(results)
+
+
+if __name__ == '__main__':
+    main()

breach_check/breach.py

@@ -1,7 +1,7 @@
 from aiohttp.client_exceptions import ClientProxyConnectionError
-from asyncio import run
+from asyncio import run, ensure_future, gather
 from breach_check.http import AsyncRequests
-from breach_check.logger import logger
+from breach_check.logger import logger, console
 from json import loads as json_loads
 from rich.progress import Progress, TaskID
 from re import compile
@@ -14,6 +14,9 @@ def __init__(self, rate_limit: int | None = None, delay: float | None = None, he
                 'User-Agent': 'Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/119.0.0.0 Safari/537.36'
             }
 
+        self.progress = Progress(console=console)
+        self.progress_task_id: TaskID | None = None
+
         self._http_client = AsyncRequests(
             rate_limit=rate_limit,
             delay=delay,
@@ -28,12 +31,42 @@ async def mass_check(self, emails: list[str] | None = None):
         if not emails or len(emails) == 0:
             return []
 
-    async def check(self, email: str | None = None) -> list[dict]:
+        self.progress.start()
+        self.progress_task_id = self.progress.add_task(
+            '[orange] Checking for Breaches:',
+            total=len(emails)
+        )
+
+        tasks = []
+        for email in emails:
+            tasks.append(
+                ensure_future(
+                    self.check(email)
+                )
+            )
+
+        try:
+            results = await gather(*tasks)
+
+            self.progress.stop()
+            return results
+        except Exception as e:
+            logger.error(
+                f'[*] Exception occurred while gathering results: {e}',
+                stack_info=True
+            )
+            return []
+
+    async def check(self, email: str | None = None) -> dict:
         if not email:
             logger.warning('email param cannot be None')
-            return []
+            return {}
 
-        res_data = []
+        res_data = {
+            'email': email,
+            'breaches': [],
+            'total': None
+        }
         try:
             email_validator = compile(r"^[^@\s']+@[^@\s']+\.[^@\s']+$")
             if email_validator.match(email):
@@ -51,22 +84,33 @@ async def check(self, email: str | None = None) -> list[dict]:
                 is_success = res_body.get('success', False)
 
                 if status_code == 200 and is_success:
-                    res_data = res_body.get('breaches')
+                    breaches = res_body.get('breaches', [])
+                    total = res_body.get('total', -1)
+                    res_data['breaches'] = breaches
+                    res_data['total'] = total
 
                 elif status_code == 429:
                     logger.warning('Rate Limited')
+
                 else:
                     logger.error(f'Failed with status code: {status_code}')
                     logger.error(response, res_body)
+
             else:
                 logger.warning(f'{email} is not a valid email')
+
+            # advance progress bar
+            if self.progress_task_id != None:
+                self.progress.update(self.progress_task_id,
+                                     advance=1, refresh=True)
+            else:
+                logger.error('No Progress Bar Task Found!')
+
+            if self.progress and self.progress.finished:
+                self.progress.stop()
+
             return res_data
         except ConnectionRefusedError:
             logger.error('Connection Failed! Server refused Connection!!')
         except ClientProxyConnectionError as e:
             logger.error(f'Proxy Connection Error: {e}')
-
-
-if __name__ == '__main__':
-    result = run(BreachChecker().check(email='admin@example.com'))
-    logger.info(result)

breach_check/http.py

@@ -1,10 +1,11 @@
-from aiohttp import ClientSession, ClientResponse, TCPConnector
+from aiohttp import ClientSession, TCPConnector
 from os import name as os_name
 
 
 import asyncio
 import aiohttp.resolver
 
+
 aiohttp.resolver.DefaultResolver = aiohttp.resolver.AsyncResolver
 if os_name == 'nt':
     asyncio.set_event_loop_policy(asyncio.WindowsSelectorEventLoopPolicy())

breach_check/utils.py

@@ -0,0 +1,43 @@
+from breach_check.logger import logger
+from datetime import datetime
+from json import dumps as json_dumps, JSONDecodeError
+from os.path import isfile
+
+
+def generate_unique_filename():
+    current_time = datetime.now()
+    timestamp = current_time.strftime(
+        "%Y%m%d%H%M%S")  # YearMonthDayHourMinuteSecond
+    unique_filename = f"output_{timestamp}.json"
+    return unique_filename
+
+
+def extract_emails(file_path: str) -> list[str] | None:
+    if not isfile(file_path):
+        logger.error(f'Input File with Emails Not Found: {file_path}')
+        return
+
+    with open(file_path, 'r') as f:
+        emails = [email.strip() for email in f.read().splitlines()]
+
+    return emails
+
+
+def write_json_file(file_path: str, json_data) -> bool:
+    if isfile(file_path):
+        logger.warning(f'{file_path} data will be overwritten')
+
+    try:
+        json_data = json_dumps(json_data)
+    except JSONDecodeError:
+        logger.error('Invalid JSON Data')
+        return False
+    except Exception as e:
+        logger.error(f'Exception: {e}')
+        return False
+
+    with open(file_path, 'w') as f:
+        f.write(json_data)
+        logger.info(f'data written to {file_path} successfully')
+
+    return True

pyproject.toml

@@ -1,7 +1,7 @@
 [tool.poetry]
 name = "breach-check"
 version = "0.1.0"
-description = "check for dark web breaches for provided emails"
+description = "check for data breaches for provided emails"
 authors = ["Dhrumil Mistry <56185972+dmdhrumilmistry@users.noreply.github.com>"]
 license = "MIT"
 readme = "README.md"
@@ -14,6 +14,16 @@ rich = "^13.7.0"
 [tool.poetry.group.dev.dependencies]
 pytest = "^7.4.3"
 
+[tool.poetry.urls]
+"Home" = "https://github.com/dmdhrumilmistry/breach-check"
+"Bug Tracker" = "https://github.com/dmdhrumilmistry/breach-check/issues"
+"Support" = "https://github.com/sponsors/dmdhrumilmistry/"
+"PayPal" = "https://paypal.me/dmdhrumilmistry"
+
+[tool.poetry.scripts]
+breach-check = "breach_check.__main__:main"
+
 [build-system]
 requires = ["poetry-core"]
 build-backend = "poetry.core.masonry.api"
+