[Screenshot] Replace imgkit with weasyprint

.github/workflows/linux.yml

@@ -26,8 +26,6 @@ jobs:
           sudo apt-get install python3 python3-dev python-is-python3
           sudo apt-get install build-essential clang systemtap-sdt-dev
           sudo apt-get install libfuzzy-dev graphviz libgraphviz-dev tesseract-ocr unrar xfonts-base xfonts-75dpi
-          wget https://github.com/wkhtmltopdf/packaging/releases/download/0.12.6-1/wkhtmltox_0.12.6-1.bionic_amd64.deb
-          sudo dpkg -i wkhtmltox_0.12.6-1.bionic_amd64.deb
 
       - name: STEP 4. Install libemu
         run: |

docs/source/logging.rst

@@ -340,7 +340,7 @@ type is PE).
 screenshots
 ^^^^^^^^^^^
 
-The collection *screenshots* is used to store base-64 encoded screenshots (JPG format)
+The collection *screenshots* is used to store base-64 encoded screenshots (PDF format)
 of the analyzed page.
 
 .. code-block:: sh

pyproject.toml

@@ -81,8 +81,8 @@ exploitgraph = [
 ]
 
 image = [
-    "imgkit==1.1.0",
     "pytesseract",
+    "weasyprint",
 ]
 
 shellcode = [

thug/Analysis/screenshot/Screenshot.py

@@ -1,13 +1,13 @@
-import sys
+import io
 import logging
 import bs4
 
 try:
-    import imgkit
+    import weasyprint
 
-    IMGKIT_MODULE = True
+    WEASYPRINT_MODULE = True
 except ImportError:  # pragma: no cover
-    IMGKIT_MODULE = False
+    WEASYPRINT_MODULE = False
 
 log = logging.getLogger("Thug")
 
@@ -16,7 +16,7 @@ class Screenshot:
     content_types = ("text/html",)
 
     def __init__(self):
-        self.enable = IMGKIT_MODULE
+        self.enable = WEASYPRINT_MODULE
 
     def run(self, window, url, response, ctype):
         if not self.enable or not log.ThugOpts.screenshot:
@@ -37,13 +37,14 @@ def run(self, window, url, response, ctype):
                 img["src"] = norm_src
 
         content = soup.prettify(formatter=None)
-        options = {"quiet": ""}
-
-        if sys.platform in ("linux",):
-            options["xvfb"] = ""
 
         try:
-            screenshot = imgkit.from_string(content, False, options=options)
-            log.ThugLogging.log_screenshot(url, screenshot)
+            html = weasyprint.HTML(string=content)
+            document = html.render()
+
+            with io.BytesIO() as screenshot:
+                document.write_pdf(screenshot)
+                screenshot.seek(0)
+                log.ThugLogging.log_screenshot(url, screenshot.read())
         except Exception as e:  # pragma: no cover,pylint:disable=broad-except
             log.warning("[SCREENSHOT] Error: %s", str(e))

thug/Logging/ThugLogging.py

@@ -453,7 +453,7 @@ def log_screenshot(self, url, screenshot):
         @screenshot Screenshot
         """
         dirname = os.path.join(self.baseDir, "analysis", "screenshots")
-        filename = f"{hashlib.sha256(screenshot).hexdigest()}.jpg"
+        filename = f"{hashlib.sha256(screenshot).hexdigest()}.pdf"
         self.store_content(dirname, filename, screenshot)
 
         for m in self.resolve_method("log_screenshot"):  # pragma: no cover