test - also include header offset

Co-authored-by: silversub <silversub@gmail.com>

changedetectionio/content_fetchers/puppeteer.py

@@ -147,7 +147,7 @@ class fetcher(Fetcher):
                          is_binary,
                          empty_pages_are_a_change
                          ):
-
+        import re
         self.delete_browser_steps_screenshots()
         extra_wait = int(os.getenv("WEBDRIVER_DELAY_BEFORE_CONTENT_READY", 5)) + self.render_extract_delay
 
@@ -172,6 +172,17 @@ class fetcher(Fetcher):
         # headless - ask a new page
         self.page = (pages := await browser.pages) and len(pages) or await browser.newPage()
 
+        if '--window-size' in self.browser_connection_url:
+            # Be sure the viewport is always the window-size, this is often not the same thing
+            match = re.search(r'--window-size=(\d+),(\d+)', self.browser_connection_url)
+            if match:
+                logger.debug(f"Setting viewport to same as --window-size in browser connection URL {int(match.group(1))},{int(match.group(2))}")
+                await self.page.setViewport({
+                    "width": int(match.group(1)),
+                    "height": int(match.group(2))
+                })
+                logger.debug(f"Puppeteer viewport size {self.page.viewport}")
+
         try:
             from pyppeteerstealth import inject_evasions_into_page
         except ImportError:
@@ -218,7 +229,6 @@ class fetcher(Fetcher):
 
         response = await self.page.goto(url, waitUntil="load")
 
-
         if response is None:
             await self.page.close()
             await browser.close()

changedetectionio/content_fetchers/res/stock-not-in-stock.js

@@ -125,6 +125,20 @@ async () => {
         // so it's good to filter to just the 'above the fold' elements
         // and it should be atleast 100px from the top to ignore items in the toolbar, sometimes menu items like "Coming soon" exist
 
+        function elementIsInEyeBallRange(element) {
+            // outside the 'fold' or some weird text in the heading area
+            // .getBoundingClientRect() was causing a crash in chrome 119, can only be run on contentVisibility != hidden
+            // Note: theres also an automated test that places the 'out of stock' text fairly low down
+            // Skip text that could be in the header area
+            if (element.getBoundingClientRect().bottom + window.scrollY <= 300 ) {
+                return false;
+            }
+            // Skip text that could be much further down (like a list of "you may like" products that have 'sold out' in there
+            if (element.getBoundingClientRect().bottom + window.scrollY >= 1300 ) {
+                return false;
+            }
+            return true;
+        }
 
 // @todo - if it's SVG or IMG, go into image diff mode
 
@@ -161,9 +175,7 @@ async () => {
         for (let i = elementsToScan.length - 1; i >= 0; i--) {
             const element = elementsToScan[i];
 
-            // outside the 'fold' or some weird text in the heading area
-            // .getBoundingClientRect() was causing a crash in chrome 119, can only be run on contentVisibility != hidden
-            if (element.getBoundingClientRect().top + window.scrollY >= vh || element.getBoundingClientRect().top + window.scrollY <= 100) {
+            if (!elementIsInEyeBallRange(element)) {
                 continue
             }
 
@@ -177,11 +189,11 @@ async () => {
             } catch (e) {
                 console.warn('stock-not-in-stock.js scraper - handling element for gettext failed', e);
             }
-
             if (elementText.length) {
                 // try which ones could mean its in stock
                 if (negateOutOfStockRegex.test(elementText) && !elementText.includes('(0 products)')) {
                     console.log(`Negating/overriding 'Out of Stock' back to "Possibly in stock" found "${elementText}"`)
+                    element.style.border = "2px solid green"; // highlight the element that was detected as in stock
                     return 'Possibly in stock';
                 }
             }
@@ -190,10 +202,8 @@ async () => {
         // OTHER STUFF THAT COULD BE THAT IT'S OUT OF STOCK
         for (let i = elementsToScan.length - 1; i >= 0; i--) {
             const element = elementsToScan[i];
-            // outside the 'fold' or some weird text in the heading area
-            // .getBoundingClientRect() was causing a crash in chrome 119, can only be run on contentVisibility != hidden
-            // Note: theres also an automated test that places the 'out of stock' text fairly low down
-            if (element.getBoundingClientRect().top + window.scrollY >= vh + 250 || element.getBoundingClientRect().top + window.scrollY <= 100) {
+
+            if (!elementIsInEyeBallRange(element)) {
                 continue
             }
             elementText = "";
@@ -208,6 +218,7 @@ async () => {
                 for (const outOfStockText of outOfStockTexts) {
                     if (elementText.includes(outOfStockText)) {
                         console.log(`Selected 'Out of Stock' - found text "${outOfStockText}" - "${elementText}" - offset top ${element.getBoundingClientRect().top}, page height is ${vh}`)
+                        element.style.border = "2px solid red"; // highlight the element that was detected as out of stock
                         return outOfStockText; // item is out of stock
                     }
                 }

changedetectionio/content_fetchers/res/xpath_element_scraper.js

@@ -202,7 +202,6 @@ async (options) => {
         // Foreach filter, go and find it on the page and add it to the results so we can visualise it again
         for (const f of include_filters) {
             bbox = false;
-            q = false;
 
             if (!f.length) {
                 console.log("xpath_element_scraper: Empty filter, skipping");
@@ -255,7 +254,7 @@ async (options) => {
                             console.log("xpath_element_scraper: Got filter by ownerElement element, scroll from top was " + scroll_y)
                         } catch (e) {
                             console.log(e)
-                            console.log("xpath_element_scraper: error looking up q.ownerElement")
+                            console.log("xpath_element_scraper: error looking up node.ownerElement")
                         }
                     }
 

changedetectionio/tests/restock/test_restock.py

@@ -14,6 +14,8 @@ from changedetectionio.notification import (
 def set_original_response():
     test_return_data = """<html>
        <body>
+       <section id=header style="padding: 50px; height: 350px">This is the header which should be ignored always - <span>add to cart</span></section>
+       <!-- stock-not-in-stock.js will ignore text in the first 300px, see elementIsInEyeBallRange(), sometimes "add to cart" and other junk is here -->
      Some initial text<br>
      <p>Which is across multiple lines</p>
      <br>
@@ -52,8 +54,6 @@ def test_restock_detection(client, live_server, measure_memory_usage):
 
     set_original_response()
     #assert os.getenv('PLAYWRIGHT_DRIVER_URL'), "Needs PLAYWRIGHT_DRIVER_URL set for this test"
-
-    time.sleep(1)
     live_server_setup(live_server)
     #####################
     notification_url = url_for('test_notification_endpoint', _external=True).replace('http://localhost', 'http://changedet').replace('http', 'json')
@@ -84,7 +84,8 @@ def test_restock_detection(client, live_server, measure_memory_usage):
     # Is it correctly show as NOT in stock?
     wait_for_all_checks(client)
     res = client.get(url_for("watchlist.index"))
-    assert b'not-in-stock' in res.data
+    assert b'processor-restock_diff' in res.data # Should have saved in restock mode
+    assert b'not-in-stock' in res.data # should be out of stock
 
     # Is it correctly shown as in stock
     set_back_in_stock_response()

changedetectionio/tests/test_ignore_text.py

@@ -32,13 +32,14 @@ def test_strip_text_func():
     stripped_content = html_tools.strip_ignore_text(test_content, ignore)
     assert stripped_content == "Some initial text\n\nWhich is across multiple lines\n\n\n\nSo let's see what happens."
 
-def set_original_ignore_response():
-    test_return_data = """<html>
+def set_original_ignore_response(ver_stamp="123"):
+    test_return_data = f"""<html>
        <body>
      Some initial text<br>
      <p>Which is across multiple lines</p>
      <br>
      So let's see what happens.  <br>
+     <link href="https://www.somesite/wp-content/themes/cooltheme/style2.css?v={ver_stamp}" rel="stylesheet"/>
      </body>
      </html>
 
@@ -48,13 +49,14 @@ def set_original_ignore_response():
         f.write(test_return_data)
 
 
-def set_modified_original_ignore_response():
-    test_return_data = """<html>
+def set_modified_original_ignore_response(ver_stamp="123"):
+    test_return_data = f"""<html>
        <body>
      Some NEW nice initial text<br>
      <p>Which is across multiple lines</p>
      <br>
      So let's see what happens.  <br>
+     <link href="https://www.somesite/wp-content/themes/cooltheme/style2.css?v={ver_stamp}" rel="stylesheet"/>
      <p>new ignore stuff</p>
      <p>blah</p>
      </body>
@@ -67,14 +69,15 @@ def set_modified_original_ignore_response():
 
 
 # Is the same but includes ZZZZZ, 'ZZZZZ' is the last line in ignore_text
-def set_modified_ignore_response():
-    test_return_data = """<html>
+def set_modified_ignore_response(ver_stamp="123"):
+    test_return_data = f"""<html>
        <body>
      Some initial text<br>
      <p>Which is across multiple lines</p>
      <P>ZZZZz</P>
      <br>
      So let's see what happens.  <br>
+     <link href="https://www.somesite/wp-content/themes/cooltheme/style2.css?v={ver_stamp}" rel="stylesheet"/>
      </body>
      </html>
 
@@ -165,9 +168,9 @@ def test_check_ignore_text_functionality(client, live_server, measure_memory_usa
     assert b'Deleted' in res.data
 
 # When adding some ignore text, it should not trigger a change, even if something else on that line changes
-def test_check_global_ignore_text_functionality(client, live_server, measure_memory_usage):
-    #live_server_setup(live_server)
-    ignore_text = "XXXXX\r\nYYYYY\r\nZZZZZ"
+def _run_test_global_ignore(client, as_source=False, extra_ignore=""):
+    ignore_text = "XXXXX\r\nYYYYY\r\nZZZZZ\r\n"+extra_ignore
+
     set_original_ignore_response()
 
     # Goto the settings page, add our ignore text
@@ -186,6 +189,10 @@ def test_check_global_ignore_text_functionality(client, live_server, measure_mem
 
     # Add our URL to the import page
     test_url = url_for('test_endpoint', _external=True)
+    if as_source:
+        # Switch to source mode so we can test that too!
+        test_url = "source:"+test_url
+
     res = client.post(
         url_for("imports.import_page"),
         data={"urls": test_url},
@@ -203,12 +210,15 @@ def test_check_global_ignore_text_functionality(client, live_server, measure_mem
         follow_redirects=True
     )
     assert b"Updated watch." in res.data
-
+    wait_for_all_checks(client)
     # Check it saved
     res = client.get(
         url_for("settings.settings_page"),
     )
-    assert bytes(ignore_text.encode('utf-8')) in res.data
+
+    for i in ignore_text.splitlines():
+        assert bytes(i.encode('utf-8')) in res.data
+
 
     # Trigger a check
     client.get(url_for("ui.form_watch_checknow"), follow_redirects=True)
@@ -221,7 +231,8 @@ def test_check_global_ignore_text_functionality(client, live_server, measure_mem
 
     # Make a change which includes the ignore text, it should be ignored and no 'change' triggered
     # It adds text with "ZZZZzzzz" and "ZZZZ" is in the ignore list
-    set_modified_ignore_response()
+    # And tweaks the ver_stamp which should be picked up by global regex ignore
+    set_modified_ignore_response(ver_stamp=time.time())
 
     # Trigger a check
     client.get(url_for("ui.form_watch_checknow"), follow_redirects=True)
@@ -243,3 +254,11 @@ def test_check_global_ignore_text_functionality(client, live_server, measure_mem
 
     res = client.get(url_for("ui.form_delete", uuid="all"), follow_redirects=True)
     assert b'Deleted' in res.data
+
+def test_check_global_ignore_text_functionality(client, live_server):
+    #live_server_setup(live_server)
+    _run_test_global_ignore(client, as_source=False)
+
+def test_check_global_ignore_text_functionality_as_source(client, live_server):
+    #live_server_setup(live_server)
+    _run_test_global_ignore(client, as_source=True, extra_ignore='/\?v=\d/')

docker-compose.yml

@@ -72,7 +72,7 @@ services:
   
       # Comment out ports: when using behind a reverse proxy , enable networks: etc.
       ports:
-        - 5000:5000
+        - 127.0.0.1:5000:5000
       restart: unless-stopped
 
      # Used for fetching pages via WebDriver+Chrome where you need Javascript support.
@@ -82,7 +82,7 @@ services:
      # If WEBDRIVER or PLAYWRIGHT are enabled, changedetection container depends on that
      # and must wait before starting (substitute "browser-chrome" with "playwright-chrome" if last one is used)
 #      depends_on:
-#          sockpuppetbrowser:
+#          browser-sockpuppet-chrome:
 #              condition: service_started