Merge pull request #3974 from seleniumbase/more-pdf-improvements

mdmintz · web-flow · commit 677b2b036a34 · 2025-09-18T17:21:45.000-04:00
More PDF improvements
diff --git a/README.md b/README.md
@@ -65,43 +65,68 @@
 
 --------
 
-<p align="left">📗 Here's <a href="https://github.com/seleniumbase/SeleniumBase/blob/master/examples/raw_google.py">raw_google.py</a>, which performs a Google search:</p>
+<p align="left">📗 For performing a Google Search without hitting the "unusual traffic" page, you can use SeleniumBase UC Mode. Here's <a href="https://github.com/seleniumbase/SeleniumBase/blob/master/examples/raw_google.py">SeleniumBase/examples/raw_google.py</a>, which exports the search results into different formats (PDF, HTML, PNG):</p>
 
 ```python
 from seleniumbase import SB
 
 with SB(test=True, uc=True) as sb:
     sb.open("https://google.com/ncr")
-    sb.type('[title="Search"]', "SeleniumBase GitHub page\n")
-    sb.click('[href*="github.com/seleniumbase/"]')
-    sb.save_screenshot_to_logs()  # ./latest_logs/
+    sb.type('[title="Search"]', "SeleniumBase GitHub page")
+    sb.click("div:not([jsname]) > * > input")
     print(sb.get_page_title())
+    sb.sleep(2)  # Wait for the "AI Overview" result
+    if sb.is_text_visible("Generating"):
+        sb.wait_for_text("AI Overview")
+    sb.save_as_pdf_to_logs()  # Saved to ./latest_logs/
+    sb.save_page_source_to_logs()
+    sb.save_screenshot_to_logs()
 ```
 
 > `python raw_google.py`
 
-<a href="https://github.com/seleniumbase/SeleniumBase/blob/master/examples/raw_google.py"><img src="https://seleniumbase.github.io/cdn/gif/google_search.gif" alt="SeleniumBase Test" title="SeleniumBase Test" width="480" /></a>
+<a href="https://github.com/seleniumbase/SeleniumBase/blob/master/examples/raw_google.py"><img src="https://seleniumbase.github.io/cdn/img/google_sb_result.png" alt="SeleniumBase on Google" title="SeleniumBase on Google" width="440" /></a>
 
 --------
 
-<p align="left">📗 Here's an example of bypassing Cloudflare's challenge page: <a href="https://github.com/seleniumbase/SeleniumBase/blob/master/examples/cdp_mode/raw_gitlab.py">SeleniumBase/examples/cdp_mode/raw_gitlab.py</a></p>
+<p align="left">📗 Here's an example of bypassing Cloudflare's challenge page with UC Mode + CDP Mode: <a href="https://github.com/seleniumbase/SeleniumBase/blob/master/examples/cdp_mode/raw_gitlab.py">SeleniumBase/examples/cdp_mode/raw_gitlab.py</a></p>
 
 ```python
 from seleniumbase import SB
 
 with SB(uc=True, test=True, locale="en") as sb:
     url = "https://gitlab.com/users/sign_in"
     sb.activate_cdp_mode(url)
-    sb.sleep(1)
+    sb.sleep(2.2)
     sb.uc_gui_click_captcha()
-    sb.sleep(2)
+    sb.assert_text("Username", '[for="user_login"]', timeout=3)
+    sb.assert_element('label[for="user_login"]')
+    sb.highlight('button:contains("Sign in")')
+    sb.highlight('h1:contains("GitLab.com")')
+    sb.post_message("SeleniumBase wasn't detected", duration=4)
 ```
 
 <img src="https://seleniumbase.github.io/other/cf_sec.jpg" title="SeleniumBase" width="332"> <img src="https://seleniumbase.github.io/other/gitlab_bypass.png" title="SeleniumBase" width="288">
 
+<p align="left">📙 You can also use SeleniumBase's pure CDP Mode, which doesn't use chromedriver, Selenium, or a Python context manager at all: <a href="https://github.com/seleniumbase/SeleniumBase/blob/master/examples/cdp_mode/raw_cdp_gitlab.py">SeleniumBase/examples/cdp_mode/raw_cdp_gitlab.py</a></p>
+
+```python
+from seleniumbase import sb_cdp
+
+url = "https://gitlab.com/users/sign_in"
+sb = sb_cdp.Chrome(url)
+sb.sleep(2.5)
+sb.gui_click_captcha()
+sb.highlight('h1:contains("GitLab.com")')
+sb.highlight('button:contains("Sign in")')
+sb.driver.stop()
+```
+
+> (Due to a change in Chrome 137 where the --load-extension switch was removed, one limitation with this format is that you can't load extensions directly. The other formats weren't affected by this change.)
+
 --------
 
-<p align="left">📗 Here's <a href="https://github.com/seleniumbase/SeleniumBase/blob/master/examples/test_get_swag.py">test_get_swag.py</a>, which tests an e-commerce site:</p>
+<p align="left">📗 Here's <a href="https://github.com/seleniumbase/SeleniumBase/blob/master/examples/test_get_swag.py">SeleniumBase/examples/test_get_swag.py</a>, which tests an e-commerce site:</p>
 
 ```python
 from seleniumbase import BaseCase
@@ -133,7 +158,7 @@ class MyTestClass(BaseCase):
 
 --------
 
-<p align="left">📗 Here's <a href="https://github.com/seleniumbase/SeleniumBase/blob/master/examples/test_coffee_cart.py" target="_blank">test_coffee_cart.py</a>, which verifies an e-commerce site:</p>
+<p align="left">📗 Here's <a href="https://github.com/seleniumbase/SeleniumBase/blob/master/examples/test_coffee_cart.py" target="_blank">SeleniumBase/examples/test_coffee_cart.py</a>, which verifies an e-commerce site:</p>
 
 ```zsh
 pytest test_coffee_cart.py --demo
@@ -147,7 +172,7 @@ pytest test_coffee_cart.py --demo
 
 <a id="multiple_examples"></a>
 
-<p align="left">📗 Here's <a href="https://github.com/seleniumbase/SeleniumBase/blob/master/examples/test_demo_site.py" target="_blank">test_demo_site.py</a>, which covers several actions:</p>
+<p align="left">📗 Here's <a href="https://github.com/seleniumbase/SeleniumBase/blob/master/examples/test_demo_site.py" target="_blank">SeleniumBase/examples/test_demo_site.py</a>, which covers several actions:</p>
 
 ```zsh
 pytest test_demo_site.py
diff --git a/examples/cdp_mode/ReadMe.md b/examples/cdp_mode/ReadMe.md
@@ -521,6 +521,7 @@ sb.cdp.scroll_up(amount=25)
 sb.cdp.scroll_down(amount=25)
 sb.cdp.save_screenshot(name, folder=None, selector=None)
 sb.cdp.print_to_pdf(name, folder=None)
+sb.cdp.save_as_pdf(name, folder=None)
 ```
 
 --------
diff --git a/examples/raw_google.py b/examples/raw_google.py
@@ -2,8 +2,12 @@
 
 with SB(test=True, uc=True) as sb:
     sb.open("https://google.com/ncr")
-    sb.type('[title="Search"]', "SeleniumBase GitHub page\n")
-    sb.sleep(1)
-    sb.click('[href*="github.com/seleniumbase/"]')
-    sb.save_screenshot_to_logs()  # ./latest_logs/
+    sb.type('[title="Search"]', "SeleniumBase GitHub page")
+    sb.click("div:not([jsname]) > * > input")
     print(sb.get_page_title())
+    sb.sleep(2)  # Wait for the "AI Overview" result
+    if sb.is_text_visible("Generating"):
+        sb.wait_for_text("AI Overview")
+    sb.save_as_pdf_to_logs()  # Saved to ./latest_logs/
+    sb.save_page_source_to_logs()
+    sb.save_screenshot_to_logs()
diff --git a/help_docs/method_summary.md b/help_docs/method_summary.md
@@ -247,6 +247,7 @@ self.switch_to_driver(driver)
 self.switch_to_default_driver()
 self.save_screenshot(name, folder=None, selector=None, by="css selector")
 self.save_screenshot_to_logs(name=None, selector=None, by="css selector")
+self.save_as_pdf_to_logs(name=None)
 self.save_data_to_logs(data, file_name=None)
 self.append_data_to_logs(data, file_name=None)
 self.save_page_source(name, folder=None)
@@ -342,6 +343,8 @@ self.save_data_as(data, file_name, destination_folder=None)
 self.append_data_to_file(data, file_name, destination_folder=None)
 self.get_file_data(file_name, folder=None)
 self.print_to_pdf(name, folder=None)
+# Duplicates:
+# self.save_as_pdf(name, folder=None)
 self.get_downloads_folder()
 self.get_browser_downloads_folder()
 self.get_downloaded_files(regex=None, browser=False)
diff --git a/mkdocs_build/requirements.txt b/mkdocs_build/requirements.txt
@@ -14,7 +14,7 @@ pathspec==0.12.1
 Babel==2.17.0
 paginate==0.5.7
 mkdocs==1.6.1
-mkdocs-material==9.6.19
+mkdocs-material==9.6.20
 mkdocs-exclude-search==0.6.6
 mkdocs-simple-hooks==0.1.5
 mkdocs-material-extensions==1.3.1
diff --git a/sbase/steps.py b/sbase/steps.py
@@ -1191,6 +1191,15 @@ def set_attributes(context, selector, attribute, value):
     sb.set_attributes(selector, attribute, value)
 
 
+@step("Save as PDF to logs")
+@step("Save as PDF to the logs")
+@step("User saves page as PDF to logs")
+@step("User saves page as PDF to the logs")
+def save_as_pdf_to_logs(context):
+    sb = context.sb
+    sb.save_as_pdf_to_logs()
+
+
 @step("Save page source to logs")
 @step("Save the page source to the logs")
 @step("User saves page source to logs")
diff --git a/seleniumbase/__version__.py b/seleniumbase/__version__.py
@@ -1,2 +1,2 @@
 # seleniumbase package
-__version__ = "4.41.6"
+__version__ = "4.41.7"
diff --git a/seleniumbase/behave/behave_helper.py b/seleniumbase/behave/behave_helper.py
@@ -522,6 +522,8 @@ def generate_gherkin(srt_actions):
                 )
         elif action[0] == "ss_tl":
             sb_actions.append("Save screenshot to logs")
+        elif action[0] == "pdftl":
+            sb_actions.append("Save as PDF to logs")
         elif action[0] == "spstl":
             sb_actions.append("Save page source to logs")
         elif action[0] == "sh_fc":
diff --git a/seleniumbase/core/browser_launcher.py b/seleniumbase/core/browser_launcher.py
@@ -865,6 +865,7 @@ def uc_open_with_cdp_mode(driver, url=None, **kwargs):
     cdp.scroll_down = CDPM.scroll_down
     cdp.save_screenshot = CDPM.save_screenshot
     cdp.print_to_pdf = CDPM.print_to_pdf
+    cdp.save_as_pdf = CDPM.save_as_pdf
     cdp.page = page  # async world
     cdp.driver = driver.cdp_base  # async world
     cdp.tab = cdp.page  # shortcut (original)
diff --git a/seleniumbase/core/recorder_helper.py b/seleniumbase/core/recorder_helper.py
@@ -556,6 +556,9 @@ def generate_sbase_code(srt_actions):
         elif action[0] == "ss_tl":
             method = "save_screenshot_to_logs"
             sb_actions.append("self.%s()" % method)
+        elif action[0] == "pdftl":
+            method = "save_as_pdf_to_logs"
+            sb_actions.append("self.%s()" % method)
         elif action[0] == "spstl":
             method = "save_page_source_to_logs"
             sb_actions.append("self.%s()" % method)
diff --git a/seleniumbase/core/sb_cdp.py b/seleniumbase/core/sb_cdp.py
@@ -2571,6 +2571,9 @@ def print_to_pdf(self, name, folder=None):
             filename = os.path.join(folder, name)
         self.loop.run_until_complete(self.page.print_to_pdf(filename))
 
+    def save_as_pdf(self, *args, **kwargs):
+        self.print_to_pdf(*args, **kwargs)
+
 
 class Chrome(CDPMethods):
     def __init__(self, url=None, **kwargs):
diff --git a/seleniumbase/fixtures/base_case.py b/seleniumbase/fixtures/base_case.py
@@ -134,6 +134,7 @@ def __initialize_variables(self):
         self.__requests_timeout = None
         self.__page_source_count = 0
         self.__screenshot_count = 0
+        self.__saved_pdf_count = 0
         self.__logs_data_count = 0
         self.__last_data_file = None
         self.__level_0_visual_f = False
@@ -4492,7 +4493,8 @@ def save_screenshot_to_logs(
         If a provided selector is not found, then takes a full-page screenshot.
         (The last_page / failure screenshot is always "screenshot.png")
         The screenshot will be in PNG format."""
-        self.wait_for_ready_state_complete()
+        if not self.__is_cdp_swap_needed():
+            self.wait_for_ready_state_complete()
         test_logpath = os.path.join(self.log_path, self.__get_test_id())
         self.__create_log_path_as_needed(test_logpath)
         if name:
@@ -4510,6 +4512,11 @@ def save_screenshot_to_logs(
         if selector and by:
             selector, by = self.__recalculate_selector(selector, by)
             if page_actions.is_element_present(self.driver, selector, by):
+                if self.__is_cdp_swap_needed():
+                    selector = self.convert_to_css_selector(selector, by=by)
+                    return self.cdp.save_screenshot(
+                        name, folder=test_logpath, selector=selector
+                    )
                 return page_actions.save_screenshot(
                     self.driver, name, test_logpath, selector, by
                 )
@@ -4523,8 +4530,49 @@ def save_screenshot_to_logs(
                         action = ["ss_tl", "", origin, time_stamp]
                         self.__extra_actions.append(action)
         sb_config._has_logs = True
+        if self.__is_cdp_swap_needed():
+            return self.cdp.save_screenshot(name, folder=test_logpath)
         return page_actions.save_screenshot(self.driver, name, test_logpath)
 
+    def save_as_pdf(self, name, folder=None):
+        """Same as self.print_to_pdf()"""
+        return self.print_to_pdf(name, folder=folder)
+
+    def save_as_pdf_to_logs(self, name=None):
+        """Saves the page as a PDF to the "latest_logs/" folder.
+        Naming is automatic:
+            If NO NAME provided: "_1_PDF.pdf", "_2_PDF.pdf", etc.
+            If NAME IS provided, then: "_1_name.pdf", "_2_name.pdf", etc."""
+        if not self.__is_cdp_swap_needed():
+            self.wait_for_ready_state_complete()
+        test_logpath = os.path.join(self.log_path, self.__get_test_id())
+        self.__create_log_path_as_needed(test_logpath)
+        if name:
+            name = str(name)
+        self.__saved_pdf_count += 1
+        if not name or len(name) == 0:
+            name = "_%s_PDF.pdf" % self.__saved_pdf_count
+        else:
+            pre_name = "_%s_" % self.__saved_pdf_count
+            if len(name) >= 4 and name[-4:].lower() == ".pdf":
+                name = name[:-4]
+                if len(name) == 0:
+                    name = "PDF"
+            name = "%s%s.pdf" % (pre_name, name)
+        if self.recorder_mode:
+            url = self.get_current_url()
+            if url and len(url) > 0:
+                if ("http:") in url or ("https:") in url or ("file:") in url:
+                    if self.get_session_storage_item("pause_recorder") == "no":
+                        time_stamp = self.execute_script("return Date.now();")
+                        origin = self.get_origin()
+                        action = ["pdftl", "", origin, time_stamp]
+                        self.__extra_actions.append(action)
+        sb_config._has_logs = True
+        if self.__is_cdp_swap_needed():
+            return self.cdp.print_to_pdf(name, folder=test_logpath)
+        return self.print_to_pdf(name, test_logpath)
+
     def save_page_source_to_logs(self, name=None):
         """Saves the page HTML to the "latest_logs/" folder.
         Naming is automatic:
@@ -5517,6 +5565,7 @@ def __process_recorded_actions(self):
         ext_actions.append("s_scr")
         ext_actions.append("ss_tf")
         ext_actions.append("ss_tl")
+        ext_actions.append("pdftl")
         ext_actions.append("spstl")
         ext_actions.append("da_el")
         ext_actions.append("da_ep")
diff --git a/setup.py b/setup.py
@@ -270,7 +270,7 @@
             'pdfminer.six==20250324;python_version<"3.9"',
             'pdfminer.six==20250506;python_version>="3.9"',
             'cryptography==39.0.2;python_version<"3.9"',
-            'cryptography==45.0.7;python_version>="3.9"',
+            'cryptography==46.0.1;python_version>="3.9"',
             'cffi==1.17.1;python_version<"3.9"',
             'cffi==2.0.0;python_version>="3.9"',
             'pycparser==2.22;python_version<"3.9"',
@@ -296,7 +296,7 @@
         ],
         # pip install -e .[psutil]
         "psutil": [
-            "psutil==7.0.0",
+            "psutil==7.1.0",
         ],
         # pip install -e .[pyautogui]
         "pyautogui": [

Original file line number	Diff line number	Diff line change
`@@ -1,2 +1,2 @@`
`1`	`1`	`# seleniumbase package`
`2`		`-__version__ = "4.41.6"`
	`2`	`+__version__ = "4.41.7"`
Original file line number	Diff line number	Diff line change
`@@ -522,6 +522,8 @@ def generate_gherkin(srt_actions):`
`522`	`522`	`)`
`523`	`523`	`elif action[0] == "ss_tl":`
`524`	`524`	`sb_actions.append("Save screenshot to logs")`
	`525`	`+ elif action[0] == "pdftl":`
	`526`	`+ sb_actions.append("Save as PDF to logs")`
`525`	`527`	`elif action[0] == "spstl":`
`526`	`528`	`sb_actions.append("Save page source to logs")`
`527`	`529`	`elif action[0] == "sh_fc":`