har2tree 1.32.0__tar.gz → 1.34.2__tar.gz

{har2tree-1.32.0 → har2tree-1.34.2}/PKG-INFO

@@ -1,36 +1,35 @@
-Metadata-Version: 2.3
+Metadata-Version: 2.4
 Name: har2tree
-Version: 1.32.0
+Version: 1.34.2
 Summary: HTTP Archive (HAR) to ETE Toolkit generator
-License: BSD-3-Clause
+License-Expression: BSD-3-Clause
+License-File: LICENSE
 Author: Raphaël Vinot
 Author-email: raphael.vinot@circl.lu
-Requires-Python: >=3.9,<4.0
+Requires-Python: >=3.10,<3.15
 Classifier: Intended Audience :: Information Technology
 Classifier: Intended Audience :: Science/Research
 Classifier: Intended Audience :: Telecommunications Industry
-Classifier: License :: OSI Approved :: BSD License
 Classifier: Operating System :: POSIX :: Linux
 Classifier: Programming Language :: Python :: 3
-Classifier: Programming Language :: Python :: 3.9
 Classifier: Programming Language :: Python :: 3.10
 Classifier: Programming Language :: Python :: 3.11
 Classifier: Programming Language :: Python :: 3.12
 Classifier: Programming Language :: Python :: 3.13
+Classifier: Programming Language :: Python :: 3.14
 Classifier: Topic :: Internet
 Classifier: Topic :: Security
 Provides-Extra: docs
 Requires-Dist: Sphinx (>=8.2.3) ; (python_version >= "3.11") and (extra == "docs")
-Requires-Dist: beautifulsoup4[charset-normalizer,lxml] (>=4.13.5)
+Requires-Dist: beautifulsoup4[charset-normalizer,lxml] (>=4.14.2)
 Requires-Dist: ete3 (>=3.1.3)
 Requires-Dist: filetype (>=1.2.0)
 Requires-Dist: json-stream (>=2.3.3,<3.0.0)
-Requires-Dist: legacy-cgi (>=2.6.3) ; python_version >= "3.13,<4.0"
+Requires-Dist: legacy-cgi (>=2.6.4) ; python_version >= "3.13" and python_version < "4.0"
 Requires-Dist: multipart (>=1.3.0,<2.0.0)
-Requires-Dist: numpy (<2.1) ; python_version < "3.10"
-Requires-Dist: numpy (<2.3) ; python_version < "3.11"
-Requires-Dist: numpy (>=2.3.2) ; python_version >= "3.11"
-Requires-Dist: publicsuffixlist (>=1.0.2.20250824)
+Requires-Dist: numpy (>=2.2,<2.3) ; python_version < "3.11"
+Requires-Dist: numpy (>=2.3.4) ; python_version >= "3.11" and python_version < "4.0"
+Requires-Dist: publicsuffixlist (>=1.0.2.20251115)
 Requires-Dist: requests-toolbelt (>=1.0.0,<2.0.0)
 Requires-Dist: six (>=1.17.0) ; extra == "docs"
 Requires-Dist: tinycss2 (>=1.4.0)

{har2tree-1.32.0 → har2tree-1.34.2}/har2tree/har2tree.py

@@ -13,7 +13,9 @@ from functools import wraps, lru_cache
 from io import BytesIO
 from operator import itemgetter
 from pathlib import Path
-from typing import Any, Callable
+from typing import Any, TypedDict
+from collections.abc import Iterator
+from collections.abc import Callable
 from urllib.parse import unquote_plus, urlparse
 
 from .helper import rebuild_url, Har2TreeError, Har2TreeLogAdapter
@@ -58,7 +60,8 @@ def trace_make_subtree_fallback(method: Callable[..., None]) -> Callable[..., No
 
 def trace_make_subtree(method: Callable[..., None]) -> Callable[..., None]:
     @wraps(method)
-    def _impl(self: Any, root: URLNode, nodes_to_attach: list[URLNode] | None=None, dev_debug: bool=False) -> None:
+    def _impl(self: Any, root: URLNode, nodes_to_attach: list[URLNode] | None=None,
+              dev_debug: bool=False, fallback: bool=False) -> None:
         if dev_debug_mode:
             __load_debug_files()
             if dev_debug_url and root.name == dev_debug_url or nodes_to_attach is not None and any(True for u in nodes_to_attach if u.name == dev_debug_url):
@@ -67,7 +70,7 @@ def trace_make_subtree(method: Callable[..., None]) -> Callable[..., None]:
             elif dev_debug_hostname and root.hostname == dev_debug_hostname or nodes_to_attach is not None and any(True for u in nodes_to_attach if u.hostname == dev_debug_hostname):
                 root.logger.warning(f'Debugging Hostname: {dev_debug_hostname}.')
                 dev_debug = True
-        return method(self, root, nodes_to_attach, dev_debug)
+        return method(self, root, nodes_to_attach, dev_debug, fallback)
     return _impl
 
 
@@ -84,6 +87,15 @@ def __load_debug_files() -> None:
             dev_debug_hostname = f.read().strip()
 
 
+# NOTE: Copy from PlaywrightCapture to avoid extra dep
+class FramesResponse(TypedDict, total=False):
+
+    name: str
+    url: str
+    content: str | None
+    children: list[FramesResponse] | None
+
+
 class HarFile():
 
     def __init__(self, harfile: Path, capture_uuid: str):
@@ -118,7 +130,11 @@ class HarFile():
                 last_redirect = unquote_plus(_lr.read())
             self.final_redirect: str = last_redirect
             if not self._search_final_redirect():
-                self.logger.warning(f'Final redirect URL from address bar not in tree: {last_redirect}')
+                if last_redirect.startswith('chrome') or last_redirect.startswith('about'):
+                    # the capture failed.
+                    pass
+                else:
+                    self.logger.info(f'Final redirect URL from address bar not in tree: {last_redirect}')
         else:
             self.logger.debug('No last_redirect file available.')
             self.final_redirect = ''
@@ -131,6 +147,14 @@ class HarFile():
             self.logger.debug('No cookies file available.')
             self.cookies = []
 
+        framesfile = self.path.parent / f'{root_name}.frames.json'
+        if framesfile.is_file():
+            with framesfile.open() as c:
+                self.frames: FramesResponse = json.load(c)
+        else:
+            self.logger.debug('No frames file available.')
+            self.frames = {}
+
         dlfile = self.path.parent / f'{root_name}.data'
         dlfilename = self.path.parent / f'{root_name}.data.filename'
         self.downloaded_file: BytesIO | None
@@ -309,8 +333,9 @@ class Har2Tree:
         self.pages_root: dict[str, str] = {}
 
         self.all_redirects: list[str] = []
-        self.all_referer: dict[str, list[str]] = defaultdict(list)
-        self.all_initiator_url: dict[str, list[str]] = defaultdict(list)
+        # 2025-11-16: make values of referers and initiators sets because there will be duplicates
+        self.all_referer: dict[str, set[str]] = defaultdict(set)
+        self.all_initiator_url: dict[str, set[str]] = defaultdict(set)
         self._load_url_entries()
 
         # Generate cookies lookup tables
@@ -401,6 +426,62 @@ class Har2Tree:
 
         self.url_tree = self._nodes_list.pop(0)
 
+    def _load_iframes(self, current: URLNode, frames: FramesResponse) -> None:
+        if not frames.get('content') or frames['content'] is None:
+            # NOTE: debug stuff, no content makes it pretty useless.
+            if frames.get('url'):
+                if frames['url'] == "about:blank":
+                    self.logger.info('Got a frame to about:blank with no content.')
+                else:
+                    u = unquote_plus(frames['url'])
+                    self.logger.warning(f'Got a url ({u}) for the frame, but no content')
+            else:
+                self.logger.info('Got a frame, but no content.')
+            return
+
+        if (frames.get('url')
+                and not (frames['url'] in ['about:blank']  # not loading anything, same as empty
+                         or frames['url'].startswith('data')  # base64 encoded content
+                         or frames['url'].startswith('chrome-error')  # not in the HAR/tree
+                         or frames['url'].startswith('blob')  # blobs aren't URLs
+                         )):
+            u = unquote_plus(frames['url'])
+            possible_child_name = {u, u.split('#', 1)[0]}
+            # this url should be in a node directly attached to that one
+            # we need to find that node
+            for child in current.traverse():
+                if child.name in possible_child_name:
+                    self.logger.debug(f'Found URL "{u}".')
+                    # Found the node, adding the content
+                    if not hasattr(child, 'rendered_frame'):
+                        child.rendered_frame = []
+                    child.rendered_frame.append(BytesIO(frames['content'].encode()))
+                    # and mark the node as iframe
+                    child.add_feature('iframe', True)
+                    # if there are children, use that node as parent and call the current method recursvely
+                    if f_children := frames.get('children'):
+                        for f_child in f_children:
+                            self._load_iframes(child, f_child)
+                    break
+            else:
+                # Couldn'd find the node Oo
+                to_print = ', '.join(possible_child_name)
+                children_to_print = ', '.join([child.name for child in current.traverse()])
+                self.logger.warning(f'Unable to find "{to_print}" in the children of "{current.name}" - {children_to_print}')
+        else:
+            self.logger.debug(f'"{current.name}" contains an iFrame.')
+            # No URL, this frame is directly in the parent frame.
+            if not hasattr(current, 'rendered_frame'):
+                current.rendered_frame = []
+            current.rendered_frame.append(BytesIO(frames['content'].encode()))
+            self.logger.debug(f'"{current.name}" has {len(current.rendered_frame)} iFrames.')
+            # and mark the node as iframe
+            current.add_feature('iframe', True)
+            # if there are children, use that node as parent and call the current method recursvely
+            if f_children := frames.get('children'):
+                for f_child in f_children:
+                    self._load_iframes(current, f_child)
+
     @property
     def initial_referer(self) -> str | None:
         '''The referer passed to the first URL in the tree'''
@@ -523,7 +604,7 @@ class Har2Tree:
 
             if hasattr(n, 'initiator_url'):
                 # The HAR file was created by chrome/chromium and we got the _initiator key
-                self.all_initiator_url[n.initiator_url].append(n.name)
+                self.all_initiator_url[n.initiator_url].add(n.name)
 
             if url_entry['startedDateTime'] in self.har.pages_start_times:
                 for page in self.har.pages_start_times[url_entry['startedDateTime']]:
@@ -536,7 +617,7 @@ class Har2Tree:
             if hasattr(n, 'referer') and i > 0:
                 # NOTE 2021-05-14: referer to self are a real thing: url -> POST to self
                 if n.name != n.referer or ('method' in n.request and n.request['method'] == 'POST'):
-                    self.all_referer[n.referer].append(n.name)
+                    self.all_referer[n.referer].add(n.name)
 
             self._nodes_list.append(n)
             self.all_url_requests[n.name].append(n)
@@ -617,6 +698,14 @@ class Har2Tree:
             for child_node_hostname, child_nodes_url in sub_roots.items():
                 self.make_hostname_tree(child_nodes_url, child_node_hostname)
 
+    def _all_urlnodes_in_host_tree(self) -> None:
+        # debug: check if all the nodes in the URL tree are in the hostnode tree (they must have an UUID)
+        self.logger.warning('Validating host tree....')
+        for urlnode in self.url_tree.traverse():
+            if not hasattr(urlnode, 'hostnode_uuid'):
+                self.logger.error(f'URL Node not un host tree: {urlnode}')
+        self.logger.warning('host tree validated.')
+
     def make_tree(self) -> URLNode:
         """Build URL and Host trees"""
         self._make_subtree(self.url_tree)
@@ -647,6 +736,15 @@ class Har2Tree:
         # Initialize the hostname tree root
         self.hostname_tree.add_url(self.url_tree)
         self.make_hostname_tree(self.url_tree, self.hostname_tree)
+        if dev_debug_mode:
+            self._all_urlnodes_in_host_tree()
+        if isinstance(self.har.frames, dict):
+            if self.har.frames.get('children') and self.har.frames['children'] is not None:
+                # we have frames in the main one
+                for f_child in self.har.frames['children']:
+                    self._load_iframes(self.rendered_node, f_child)
+        else:
+            self.logger.warning(f'Wrong format for the frames ({type(self.har.frames)}), very old capture.')
         return self.url_tree
 
     @trace_make_subtree_fallback
@@ -664,7 +762,7 @@ class Har2Tree:
                             # we got an non-empty response, breaking
                             break
                     # attach to the the first response with something, or to whatever we get.
-                    self._make_subtree(node_with_hostname, [node])
+                    self._make_subtree(node_with_hostname, [node], fallback=True)
                     return
 
         # Sometimes, the har has a list of pages, generally when we have HTTP redirects.
@@ -682,7 +780,7 @@ class Har2Tree:
             page_root_node = self.get_url_node_by_uuid(self.pages_root[node.pageref])
             if dev_debug:
                 self.logger.warning(f'Failed to attach URLNode in the normal process, attaching node to page {node.pageref} - Node: {page_root_node.uuid} - {page_root_node.name}.')
-            self._make_subtree(page_root_node, [node])
+            self._make_subtree(page_root_node, [node], fallback=True)
         elif self.rendered_node != self.url_tree:
             # Generally, when we have a bunch of redirects, they (generally) do not branch out
             # before the final landing page *but* it is not always the case: some intermediary
@@ -694,7 +792,7 @@ class Har2Tree:
             # end of this method anyway
             if dev_debug:
                 self.logger.warning(f'Failed to attach URLNode in the normal process, attaching node to final redirect: {self.har.final_redirect}.')
-            self._make_subtree(self.rendered_node, [node])
+            self._make_subtree(self.rendered_node, [node], fallback=True)
         elif 'pages' in self.har.har['log']:
             # No luck, the node is root for this pageref, let's attach it to the prior page in the list, or the very first node (tree root)
             page_before = self.har.har['log']['pages'][0]
@@ -716,13 +814,42 @@ class Har2Tree:
                 # node to the root node
                 page_root_node = self.url_tree
                 self.logger.warning('The pages in the HAR are in in the wrong order, this should not happen but here we are')
-            self._make_subtree(page_root_node, [node])
+            self._make_subtree(page_root_node, [node], fallback=True)
         else:
             # no way to attach it to anything else, attach to the root node
-            self._make_subtree(self.url_tree, [node])
+            self._make_subtree(self.url_tree, [node], fallback=True)
+
+    def all_real_urls_in_children(self, frame: FramesResponse) -> Iterator[str]:
+        # from a frame, search all the real urls in each of the children, stop at the first one
+        if (frame.get('url') and frame['url'] is not None
+                and not (frame['url'] in ['about:blank', 'about:srcdoc']  # not loading anything, same as empty
+                         or frame['url'].startswith('data')  # base64 encoded content
+                         or frame['url'].startswith('chrome-error')  # not in the HAR/tree
+                         or frame['url'].startswith('blob'))):  # blobs aren't URLs
+            yield frame['url']
+        else:
+            # got no real URL, try the children
+            if frame.get('children') and frame['children'] is not None:
+                for c in frame['children']:
+                    yield from self.all_real_urls_in_children(c)
+
+    def search_in_frames(self, urls: set[str], frame: FramesResponse) -> Iterator[str]:
+        # If the frame doesn't have children, there are no potential URLs to attach
+        if not frame.get('children') or frame['children'] is None:
+            return None
+
+        if frame.get('url'):
+            u = unquote_plus(frame['url'])
+            if urls & {u, u.split('#', 1)[0]}:
+                # got a matching URL, get list of potential iframes urls
+                for c in frame['children']:
+                    yield from self.all_real_urls_in_children(c)
+        for c in frame['children']:
+            yield from self.search_in_frames(urls, c)
 
     @trace_make_subtree
-    def _make_subtree(self, root: URLNode, nodes_to_attach: list[URLNode] | None=None, dev_debug: bool=False) -> None:
+    def _make_subtree(self, root: URLNode, nodes_to_attach: list[URLNode] | None=None,
+                      dev_debug: bool=False, fallback: bool=False) -> None:
         """Recursive method building each level of the tree"""
         matching_urls: list[URLNode]
         if nodes_to_attach is None:
@@ -784,6 +911,26 @@ class Har2Tree:
             if unode.empty_response:
                 continue
 
+            # 2025-11-14
+            # the referer of an iframe is the hostname of the parent, even if the parent
+            # is a URL with a full path. Before using the referer, we need to check if we have
+            # the current url in the frame tree. If we do, find nodes (in the remaining list)
+            # with the URLs of the children - any fragment will be missing - and attach that node
+            possible_iframe_urls = {unode.name, unode.name.split('#', 1)[0]}
+            for possible_url in self.search_in_frames(possible_iframe_urls, self.har.frames):
+                cu = unquote_plus(possible_url)
+                for u in {cu, cu.split('#', 1)[0]}:
+                    if u not in self.all_url_requests:
+                        if '#' not in u:
+                            self.logger.info(f'"{u}" in the frames URLs, but not in the HAR.')
+                        continue
+                    matching_urls = [url_node for url_node in self.all_url_requests[u]
+                                     if url_node in self._nodes_list]
+                    self._nodes_list = [node for node in self._nodes_list if node not in matching_urls]
+                    if dev_debug:
+                        self.logger.warning(f'Found via initiator from {unode.name} to {matching_urls}.')
+                    self._make_subtree(unode, matching_urls)
+
             # The node can have a redirect, but also trigger ressources refering to themselves, we need to trigger this code on each node.
             if self.all_initiator_url.get(unode.name):
                 # The URL (unode.name) is in the list of known urls initiating calls
@@ -815,6 +962,12 @@ class Har2Tree:
             if hasattr(unode, 'external_ressources'):
                 # the url loads external things, and some of them have no referer....
                 for external_tag, links in unode.external_ressources.items():
+                    # 2025-11-06: skip full regex until we're calling this method in the fallback
+                    #             the iframes will often (not always) have a referer set and the URL
+                    #             might be found by the regex and it will not be attached at the
+                    #             right place
+                    if external_tag == 'full_regex' and not fallback:
+                        continue
                     for link in links:
                         if link not in self.all_url_requests or link == self.har.final_redirect:
                             # We have a lot of false positives

{har2tree-1.32.0 → har2tree-1.34.2}/har2tree/helper.py

@@ -364,7 +364,8 @@ def find_external_ressources(mimetype: str, data: bytes, base_url: str, all_requ
     # link: https://www.w3schools.com/TAGs/tag_link.asp -> href
     # object: https://www.w3schools.com/TAGs/tag_object.asp -> data
     external_ressources: dict[str, list[str]] = {'img': [], 'script': [], 'video': [], 'audio': [],
-                                                 'iframe': [], 'embed': [], 'source': [],
+                                                 'iframe': [],
+                                                 'embed': [], 'source': [],
                                                  'link': [],
                                                  'object': [],
                                                  'css': [],

{har2tree-1.32.0 → har2tree-1.34.2}/har2tree/nodes.py

@@ -485,6 +485,7 @@ class URLNode(HarTreeNode):
 
             # Common JS redirect we can catch easily
             # NOTE: it is extremely fragile and doesn't work very often but is kinda better than nothing.
+            # NOTE 2025-08-30: Also, finding that doesn't mean it is in a part of the code that is executed without user interaction. It can be triggered after a user fills a form for example.
             # Source: https://stackoverflow.com/questions/13363174/regular-expression-to-catch-as-many-javascript-redirections-as-possible
             regex = re.compile(br"""((location.href)|(window.location)|(location.replace)|(location.assign))(( ?= ?)|( ?\( ?))("|')([^'"]*)("|')( ?\) ?)?;""", re.I)
             matches = re.findall(regex, self.body.getvalue())
@@ -586,7 +587,7 @@ class URLNode(HarTreeNode):
             return href
 
         if not hasattr(self, 'rendered_html') or not self.rendered_html:
-            raise Har2TreeError('Not the node of a page rendered ({self.uuid}), invalid request.')
+            raise Har2TreeError(f'Not the node of a page rendered ({self.uuid}), invalid request.')
         urls: set[str] = set()
 
         # The simple ones: the links.

{har2tree-1.32.0 → har2tree-1.34.2}/pyproject.toml

@@ -1,32 +1,31 @@
 [project]
 name = "har2tree"
-version = "1.32.0"
+version = "1.34.2"
 description = "HTTP Archive (HAR) to ETE Toolkit generator"
 authors = [
     {name="Raphaël Vinot", email="raphael.vinot@circl.lu"}
 ]
 license = "BSD-3-Clause"
 readme = "README.md"
-requires-python = ">=3.9,<4.0"
+requires-python = ">=3.10,<3.15"
 
 dynamic = [ "classifiers" ]
 
 dependencies = [
     "ete3 (>=3.1.3)",
-    "beautifulsoup4[charset-normalizer,lxml] (>=4.13.5)",
-    "publicsuffixlist (>=1.0.2.20250824)",
+    "beautifulsoup4[charset-normalizer,lxml] (>=4.14.2)",
+    "publicsuffixlist (>=1.0.2.20251115)",
     "filetype (>=1.2.0)",
-    # poetry up fails with the version of numpy forced for python < 3.10.
+    # poetry up fails with the version of numpy forced for python < 3.11.
     # The work around is to comment it, run poetry up, uncomment it. and run poetry update.
-    "numpy (<2.1) ; python_version < \"3.10\"",
-    "numpy (<2.3) ; python_version < \"3.11\"",
-    "numpy (>=2.3.2) ; python_version >= \"3.11\"",
+    "numpy (>=2.2,<2.3) ; python_version < '3.11'",
+    "numpy (>=2.3.4) ; python_version >= \"3.11\" and python_version < \"4.0\"",
     "w3lib (>=2.3.1)",
     "tinycss2 (>=1.4.0)",
-    "legacy-cgi (>=2.6.3) ; python_version >= \"3.13,<4.0\"",
+    "legacy-cgi (>=2.6.4) ; python_version >= \"3.13\" and python_version < \"4.0\"",
     "multipart (>=1.3.0,<2.0.0)",
     "json-stream (>=2.3.3,<3.0.0)",
-    "requests-toolbelt (>=1.0.0,<2.0.0)",
+    "requests-toolbelt (>=1.0.0,<2.0.0)"
 ]
 
 [project.urls]
@@ -48,9 +47,9 @@ classifiers = [
 docs = ["Sphinx (>=8.2.3) ; python_version >= \"3.11\"", "six (>=1.17.0)"]
 
 [tool.poetry.group.dev.dependencies]
-mypy = "^1.17.1"
-pytest-cov = "^6.2.1"
-coverage = "^7.10.5"
+mypy = "^1.18.2"
+pytest-cov = "^7.0.0"
+coverage = "^7.11.3"
 types-beautifulsoup4 = "^4.12.0.20250516"
 
 [build-system]