lacuscore 1.8.9__py3-none-any.whl → 1.9.0__py3-none-any.whl

lacuscore/lacuscore.py

@@ -134,6 +134,7 @@ class CaptureSettings(TypedDict, total=False):
     viewport: dict[str, int] | None
     referer: str | None
     with_favicon: bool
+    allow_tracking: bool
     force: bool
     recapture_interval: int
     priority: int
@@ -213,6 +214,7 @@ class LacusCore():
                 referer: str | None=None,
                 rendered_hostname_only: bool=True,
                 with_favicon: bool=False,
+                allow_tracking: bool=False,
                 force: bool=False,
                 recapture_interval: int=300,
                 priority: int=0,
@@ -240,6 +242,7 @@ class LacusCore():
                 referer: str | None=None,
                 rendered_hostname_only: bool=True,
                 with_favicon: bool=False,
+                allow_tracking: bool=False,
                 force: bool=False,
                 recapture_interval: int=300,
                 priority: int=0,
@@ -269,6 +272,7 @@ class LacusCore():
         :param referer: The referer URL for the capture
         :param rendered_hostname_only: If depth > 0: only capture URLs with the same hostname as the rendered page
         :param with_favicon: If True, PlaywrightCapture will attempt to get the potential favicons for the rendered URL. It is a dirty trick, see this issue for details: https://github.com/Lookyloo/PlaywrightCapture/issues/45
+        :param allow_tracking: If True, PlaywrightCapture will attempt to click through the cookie banners. It is totally dependent on the framework used on the website.
         :param force: Force recapture, even if the same one was already done within the recapture_interval
         :param recapture_interval: The time the enqueued settings are kept in memory to avoid duplicates
         :param priority: The priority of the capture
@@ -324,6 +328,8 @@ class LacusCore():
                 to_enqueue['referer'] = referer
             if with_favicon:
                 to_enqueue['with_favicon'] = with_favicon
+            if allow_tracking:
+                to_enqueue['allow_tracking'] = allow_tracking
 
         hash_query = hashlib.sha512(pickle.dumps(to_enqueue)).hexdigest()
         if not force:
@@ -471,7 +477,7 @@ class LacusCore():
                             'document', 'browser', 'device_name', 'user_agent', 'proxy',
                             'general_timeout_in_sec', 'cookies', 'headers', 'http_credentials',
                             'viewport', 'referer', 'geolocation', 'timezone_id', 'locale',
-                            'color_scheme', 'with_favicon']
+                            'color_scheme', 'with_favicon', 'allow_tracking']
             result: CaptureResponse = {}
             to_capture: CaptureSettings = {}
             document_as_bytes = b''
@@ -496,7 +502,7 @@ class LacusCore():
                     elif k in ['general_timeout_in_sec', 'depth']:
                         # int
                         to_capture[k] = int(v)  # type: ignore[literal-required]
-                    elif k in ['rendered_hostname_only', 'with_favicon']:
+                    elif k in ['rendered_hostname_only', 'with_favicon', 'allow_tracking']:
                         # bool
                         to_capture[k] = bool(int(v))  # type: ignore[literal-required]
                     elif k == 'document':
@@ -633,6 +639,7 @@ class LacusCore():
                             depth=to_capture.get('depth', 0),
                             rendered_hostname_only=to_capture.get('rendered_hostname_only', True),
                             with_favicon=to_capture.get('with_favicon', False),
+                            allow_tracking=to_capture.get('allow_tracking', False),
                             max_depth_capture_time=self.max_capture_time)
                     result = cast(CaptureResponse, playwright_result)
                     if 'error' in result and 'error_name' in result:

{lacuscore-1.8.9.dist-info → lacuscore-1.9.0.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: lacuscore
-Version: 1.8.9
+Version: 1.9.0
 Summary: Core of Lacus, usable as a module
 Home-page: https://github.com/ail-project/LacusCore
 License: BSD-3-Clause
@@ -28,7 +28,7 @@ Requires-Dist: Sphinx (>=7.2,<8.0) ; (python_version >= "3.9") and (extra == "do
 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] (>=1.23.10,<2.0.0)
+Requires-Dist: playwrightcapture[recaptcha] (>=1.24.0,<2.0.0)
 Requires-Dist: redis[hiredis] (>=5.0.3,<6.0.0)
 Requires-Dist: requests (>=2.31.0,<3.0.0)
 Requires-Dist: ua-parser (>=0.18.0,<0.19.0)

lacuscore-1.9.0.dist-info/RECORD

@@ -0,0 +1,8 @@
+lacuscore/__init__.py,sha256=10JsUmjLD9xQIsjTDYT5nGowsJVe9C-FHNkaUFuAUtU,321
+lacuscore/lacus_monitoring.py,sha256=UOfE_1-_rhVeKJXQ_m9XxYkr7VwyQnA6iK-x_tcXJfo,2775
+lacuscore/lacuscore.py,sha256=UhcFsqQ00U5lv35Y2ctLE5ybryRuru3QJjlzLKSal9w,44397
+lacuscore/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+lacuscore-1.9.0.dist-info/LICENSE,sha256=4C4hLYrIkUD96Ggk-y_Go1Qf7PBZrEm9PSeTGe2nd4s,1516
+lacuscore-1.9.0.dist-info/METADATA,sha256=SHOpneBAkbhPpDqvqfeqFylCaJlR1wQjLgD5QI1g4uo,2628
+lacuscore-1.9.0.dist-info/WHEEL,sha256=sP946D7jFCHeNz5Iq4fL4Lu-PrWrFsgfLXbbkciIZwg,88
+lacuscore-1.9.0.dist-info/RECORD,,

lacuscore-1.8.9.dist-info/RECORD

@@ -1,8 +0,0 @@
-lacuscore/__init__.py,sha256=10JsUmjLD9xQIsjTDYT5nGowsJVe9C-FHNkaUFuAUtU,321
-lacuscore/lacus_monitoring.py,sha256=UOfE_1-_rhVeKJXQ_m9XxYkr7VwyQnA6iK-x_tcXJfo,2775
-lacuscore/lacuscore.py,sha256=iI1Cz9gnhQo8dw7gPII_vCsGHB_wzv-_9cpB3jRDizw,43901
-lacuscore/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-lacuscore-1.8.9.dist-info/LICENSE,sha256=4C4hLYrIkUD96Ggk-y_Go1Qf7PBZrEm9PSeTGe2nd4s,1516
-lacuscore-1.8.9.dist-info/METADATA,sha256=L1ITEAiio8MB6NUSn-Yi9tAL6BaYMnz-VVKFg_2IX-A,2629
-lacuscore-1.8.9.dist-info/WHEEL,sha256=sP946D7jFCHeNz5Iq4fL4Lu-PrWrFsgfLXbbkciIZwg,88
-lacuscore-1.8.9.dist-info/RECORD,,