pylookyloo 1.30.0__tar.gz → 1.31.1__tar.gz

{pylookyloo-1.30.0 → pylookyloo-1.31.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: pylookyloo
-Version: 1.30.0
+Version: 1.31.1
 Summary: Python CLI and module for Lookyloo
 License: GPL-2.0-or-later
 Author: Raphaël Vinot
@@ -24,8 +24,8 @@ Classifier: Topic :: Security
 Provides-Extra: docs
 Provides-Extra: examples
 Requires-Dist: Sphinx (>=8.2.3) ; (python_version >= "3.11") and (extra == "docs")
-Requires-Dist: pylacus (>=1.15.0) ; extra == "examples"
-Requires-Dist: requests (>=2.32.3)
+Requires-Dist: pylacus (>=1.15.1) ; extra == "examples"
+Requires-Dist: requests (>=2.32.4)
 Project-URL: Documentation, https://pylookyloo.readthedocs.io/en/latest/
 Project-URL: Repository, https://github.com/lookyloo/PyLookyloo
 Project-URL: issues, https://github.com/lookyloo/PyLookyloo/issues

{pylookyloo-1.30.0 → pylookyloo-1.31.1}/pylookyloo/api.py

@@ -7,6 +7,7 @@ import base64
 import warnings
 
 from datetime import datetime
+from hashlib import sha512
 from importlib.metadata import version
 from io import BytesIO, StringIO
 from typing import Any, TypedDict, overload, Literal
@@ -89,7 +90,7 @@ class CompareSettings(TypedDict, total=False):
 class Lookyloo():
 
     def __init__(self, root_url: str='https://lookyloo.circl.lu/', useragent: str | None=None,
-                 *, proxies: dict[str, str] | None=None):
+                 *, proxies: dict[str, str] | None=None, verify: bool | str=True) -> None:
         '''Query a specific lookyloo instance.
 
         :param root_url: URL of the instance to query.
@@ -109,6 +110,7 @@ class Lookyloo():
             self.session.proxies.update(proxies)
         retries = Retry(total=5, backoff_factor=0.1, status_forcelist=[500, 502, 503, 504])
         self.session.mount('http://', HTTPAdapter(max_retries=retries))
+        self.session.verify = verify
 
     @property
     def is_up(self) -> bool:
@@ -391,6 +393,14 @@ class Lookyloo():
         r = self.session.post(urljoin(self.root_url, str(PurePosixPath('admin', tree_uuid, 'remove'))))
         return r.json()
 
+    def get_favicons(self, capture_uuid: str) -> dict[str, Any]:
+        '''Returns the potential favicons of the capture.
+
+        :param capture_uuid: UUID of the capture
+        '''
+        r = self.session.get(urljoin(self.root_url, str(PurePosixPath('json', capture_uuid, 'favicons'))))
+        return r.json()
+
     def get_redirects(self, capture_uuid: str) -> dict[str, Any]:
         '''Returns the initial redirects.
 
@@ -415,6 +425,14 @@ class Lookyloo():
         r = self.session.get(urljoin(self.root_url, str(PurePosixPath('json', capture_uuid, 'hostnames'))))
         return r.json()
 
+    def get_ips(self, capture_uuid: str) -> dict[str, Any]:
+        '''Returns all the IPs seen during the capture.
+
+        :param capture_uuid: UUID of the capture
+        '''
+        r = self.session.get(urljoin(self.root_url, str(PurePosixPath('json', capture_uuid, 'ips'))))
+        return r.json()
+
     def get_screenshot(self, capture_uuid: str) -> BytesIO:
         '''Returns the screenshot.
 
@@ -475,35 +493,78 @@ class Lookyloo():
         r = self.session.get(urljoin(self.root_url, str(PurePosixPath('bin', capture_uuid, 'export'))))
         return BytesIO(r.content)
 
-    def get_hash_occurrences(self, h: str) -> dict[str, Any]:
-        '''Returns the base 64 body related the the hash, and a list of all the captures containing that hash.
+    def get_hash_occurrences(self, h: str, *, with_urls_occurrences: bool=False, cached_captures_only: bool=True, limit: int=20, offset: int=0) -> dict[str, Any]:
+        '''Returns the base64 body related the the hash, and a list of all the captures containing that hash.
 
         :param h: sha512 to search
+        :param with_urls_occurrences: If true, add details about the URLs from the URL nodes in the tree.
+        :param cached_captures_only: If False, Lookyloo will attempt to re-cache the missing captures. It might take some time.
+        :param limit: The max amount of entries to return.
+        :param offset: The offset to start from, useful for pagination.
         '''
-        r = self.session.get(urljoin(self.root_url, str(PurePosixPath('json', 'hash_info', h))))
+        r = self.session.post(urljoin(self.root_url, str(PurePosixPath('json', 'hash_info'))),
+                              json={'body_hash': h, 'with_urls_occurrences': with_urls_occurrences,
+                                    'cached_captures_only': cached_captures_only, 'limit': limit, 'offset': offset})
         return r.json()
 
-    def get_url_occurrences(self, url: str, limit: int=20, cached_captures_only: bool=True) -> dict[str, Any]:
+    def get_url_occurrences(self, url: str, *, with_urls_occurrences: bool=False, cached_captures_only: bool=True, limit: int=20, offset: int=0) -> dict[str, Any]:
         '''Returns all the captures contining the URL
 
         :param url: URL to lookup
-        :param limit: The max amount of entries to return.
+        :param with_urls_occurrences: If true, add details about the URLs from the URL nodes in the tree.
         :param cached_captures_only: If False, Lookyloo will attempt to re-cache the missing captures. It might take some time.
+        :param limit: The max amount of entries to return.
+        :param offset: The offset to start from, useful for pagination.
         '''
         r = self.session.post(urljoin(self.root_url, str(PurePosixPath('json', 'url_info'))),
-                              json={'url': url, 'limit': limit})
+                              json={'url': url, 'with_urls_occurrences': with_urls_occurrences,
+                                    'cached_captures_only': cached_captures_only,
+                                    'limit': limit, 'offset': offset})
         return r.json()
 
-    def get_hostname_occurrences(self, hostname: str, with_urls_occurrences: bool=False, limit: int=20, cached_captures_only: bool=True) -> dict[str, Any]:
-        '''Returns all the captures contining the hostname. It will be pretty slow on very common domains.
+    def get_hostname_occurrences(self, hostname: str, *, with_urls_occurrences: bool=False, cached_captures_only: bool=True, limit: int=20, offset: int=0) -> dict[str, Any]:
+        '''Returns all the captures contining the hostname.
 
         :param hostname: Hostname to lookup
         :param with_urls_occurrences: If true, add details about the related URLs.
-        :param limit: The max amount of entries to return.
         :param cached_captures_only: If False, Lookyloo will attempt to re-cache the missing captures. It might take some time.
+        :param limit: The max amount of entries to return.
+        :param offset: The offset to start from, useful for pagination.
         '''
         r = self.session.post(urljoin(self.root_url, str(PurePosixPath('json', 'hostname_info'))),
-                              json={'hostname': hostname, 'with_urls_occurrences': with_urls_occurrences, 'limit': limit})
+                              json={'hostname': hostname, 'with_urls_occurrences': with_urls_occurrences,
+                                    'cached_captures_only': cached_captures_only,
+                                    'limit': limit, 'offset': offset})
+        return r.json()
+
+    def get_ip_occurrences(self, ip: str, *, with_urls_occurrences: bool=False, cached_captures_only: bool=True, limit: int=20, offset: int=0) -> dict[str, Any]:
+        '''Returns all the captures containing the IP address.
+
+        :param ip: IP to lookup
+        :param with_urls_occurrences: If true, add details about the related URLs.
+        :param cached_captures_only: If False, Lookyloo will attempt to re-cache the missing captures. It might take some time.
+        :param limit: The max amount of entries to return.
+        :param offset: The offset to start from, useful for pagination.
+        '''
+        r = self.session.post(urljoin(self.root_url, str(PurePosixPath('json', 'ip_info'))),
+                              json={'ip': ip, 'with_urls_occurrences': with_urls_occurrences,
+                                    'cached_captures_only': cached_captures_only,
+                                    'limit': limit})
+        return r.json()
+
+    def get_favicon_occurrences(self, favicon: str | BytesIO, *, cached_captures_only: bool=True, limit: int=20, offset: int=0) -> dict[str, Any]:
+        '''Returns all the captures containing the favicon.
+
+        :param favicon: Favicon to lookup. Either the hash, or the file in a BytesIO (hash will be generated on the fly)
+        :param cached_captures_only: If False, Lookyloo will attempt to re-cache the missing captures. It might take some time.
+        :param limit: The max amount of entries to return.
+        :param offset: The offset to start from, useful for pagination.
+        '''
+        if isinstance(favicon, BytesIO):
+            favicon = sha512(favicon.read()).hexdigest()
+        r = self.session.post(urljoin(self.root_url, str(PurePosixPath('json', 'favicon_info'))),
+                              json={'favicon': favicon, 'cached_captures_only': cached_captures_only,
+                                    'limit': limit, 'offset': offset})
         return r.json()
 
     def get_stats(self) -> dict[str, Any]:

{pylookyloo-1.30.0 → pylookyloo-1.31.1}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "pylookyloo"
-version = "1.30.0"
+version = "1.31.1"
 description = "Python CLI and module for Lookyloo"
 authors = [
     {name="Raphaël Vinot", email="raphael.vinot@circl.lu"}
@@ -12,7 +12,7 @@ requires-python = ">=3.9"
 dynamic = [ "classifiers" ]
 
 dependencies = [
-    "requests (>=2.32.3)"
+    "requests (>=2.32.4)"
 ]
 
 [project.urls]
@@ -37,12 +37,12 @@ lookyloo = 'pylookyloo:main'
 
 [project.optional-dependencies]
 docs = ["Sphinx (>=8.2.3) ; python_version >= \"3.11\""]
-examples = ["pylacus (>=1.15.0)"]
+examples = ["pylacus (>=1.15.1)"]
 
 [tool.poetry.group.dev.dependencies]
-mypy = "^1.15.0"
-types-requests = "^2.32.0.20250515"
-pytest = "^8.3.5"
+mypy = "^1.16.1"
+types-requests = "^2.32.4.20250611"
+pytest = "^8.4.1"
 
 [build-system]
 requires = ["poetry-core>=2.0"]