lacuscore 1.9.5__tar.gz → 1.9.6__tar.gz

{lacuscore-1.9.5 → lacuscore-1.9.6}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: lacuscore
-Version: 1.9.5
+Version: 1.9.6
 Summary: Core of Lacus, usable as a module
 Home-page: https://github.com/ail-project/LacusCore
 License: BSD-3-Clause
@@ -29,7 +29,7 @@ Requires-Dist: async-timeout (>=4.0.3,<5.0.0) ; python_version < "3.11"
 Requires-Dist: defang (>=0.5.3,<0.6.0)
 Requires-Dist: dnspython (>=2.6.1,<3.0.0)
 Requires-Dist: playwrightcapture[recaptcha] (>=1.24.11,<2.0.0)
-Requires-Dist: redis[hiredis] (>=5.0.4,<6.0.0)
+Requires-Dist: redis[hiredis] (>=5.0.6,<6.0.0)
 Requires-Dist: requests (>=2.32.3,<3.0.0)
 Requires-Dist: ua-parser (>=0.18.0,<0.19.0)
 Project-URL: Documentation, https://lacuscore.readthedocs.io/en/latest/

{lacuscore-1.9.5 → lacuscore-1.9.6}/lacuscore/lacuscore.py

@@ -531,6 +531,27 @@ class LacusCore():
                     browser_engine = 'firefox'
                 else:
                     browser_engine = 'webkit'
+
+            cookies: list[dict[str, Any]] = []
+            if to_capture.get('cookies') and to_capture['cookies'] is not None:
+                # In order to properly pass the cookies to playwright,
+                # each of then must have a name, a value and either a domain + path or a URL
+                # Name and value are mandatory, and we cannot auto-fill them.
+                # If the cookie doesn't have a domain + path OR a URL, we fill the domain
+                # with the hostname of the URL we try to capture and the path with "/"
+                for cookie in to_capture['cookies']:
+                    if len(cookie) == 1:
+                        # we have a cookie in the format key: value
+                        name, value = cookie.popitem()
+                        cookie = {'name': name, 'value': value}
+                    if 'name' not in cookie or 'value' not in cookie:
+                        logger.warning(f'Invalid cookie: {cookie}')
+                        continue
+                    if 'domain' not in cookie and 'url' not in cookie:
+                        cookie['domain'] = splitted_url.hostname
+                        cookie['path'] = '/'
+                    cookies.append(cookie)
+
             try:
                 logger.debug(f'Capturing {url}')
                 stats_pipeline.sadd(f'stats:{today}:captures', url)
@@ -543,7 +564,7 @@ class LacusCore():
                         uuid=uuid) as capture:
                     # required by Mypy: https://github.com/python/mypy/issues/3004
                     capture.headers = to_capture.get('headers')  # type: ignore[assignment]
-                    capture.cookies = to_capture.get('cookies')  # type: ignore[assignment]
+                    capture.cookies = cookies  # type: ignore[assignment]
                     capture.viewport = to_capture.get('viewport')  # type: ignore[assignment]
                     capture.user_agent = to_capture.get('user_agent')  # type: ignore[assignment]
                     capture.http_credentials = to_capture.get('http_credentials')  # type: ignore[assignment]
@@ -583,7 +604,7 @@ class LacusCore():
             except RetryCapture as e:
                 raise e
             except PlaywrightCaptureException as e:
-                logger.exception(f'Invalid parameters for the capture of {url} - {e}')
+                logger.warning(f'Invalid parameters for the capture of {url} - {e}')
                 result = {'error': f'Invalid parameters for the capture of {url} - {e}'}
                 raise CaptureError(f'Invalid parameters for the capture of {url} - {e}')
             except asyncio.CancelledError:

{lacuscore-1.9.5 → lacuscore-1.9.6}/pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "lacuscore"
-version = "1.9.5"
+version = "1.9.6"
 description = "Core of Lacus, usable as a module"
 authors = ["Raphaël Vinot <raphael.vinot@circl.lu>"]
 license = "BSD-3-Clause"
@@ -37,7 +37,7 @@ Sphinx = [
 playwrightcapture = {extras = ["recaptcha"], version = "^1.24.11"}
 defang = "^0.5.3"
 ua-parser = "^0.18.0"
-redis = {version = "^5.0.4", extras = ["hiredis"]}
+redis = {version = "^5.0.6", extras = ["hiredis"]}
 dnspython = "^2.6.1"
 async-timeout = {version = "^4.0.3", python = "<3.11"}
 
@@ -47,14 +47,14 @@ docs = ["Sphinx"]
 [tool.poetry.group.dev.dependencies]
 types-redis = {version = "^4.6.0.20240425"}
 mypy = "^1.10.0"
-types-requests = "^2.32.0.20240523"
+types-requests = "^2.32.0.20240602"
 types-beautifulsoup4 = "^4.12.0.20240511"
 ipython = [
     {version = "<8.13.0", python = "<3.9"},
     {version = "^8.18.0", python = ">=3.9"},
     {version = "^8.19.0", python = ">=3.10"}
 ]
-pytest = "^8.2.1"
+pytest = "^8.2.2"
 
 [build-system]
 requires = ["poetry_core"]