Merge branch 'master' into price-scraper-ML-integration

.github/workflows/test-stack-reusable-workflow.yml

@@ -52,6 +52,10 @@ jobs:
           docker run --network changedet-network -d -e "LOG_LEVEL=TRACE" --cap-add=SYS_ADMIN --name sockpuppetbrowser --hostname sockpuppetbrowser --rm -p 3000:3000 dgtlmoon/sockpuppetbrowser:latest                    
           docker run --network changedet-network -d -e "LOG_LEVEL=TRACE" --cap-add=SYS_ADMIN --name sockpuppetbrowser-custom-url --hostname sockpuppetbrowser-custom-url  -p 3001:3000 --rm dgtlmoon/sockpuppetbrowser:latest
 
+          # CDIO AI Element scraper for prices
+          # Run CDIO with PRICE_SCRAPER_ML_ENDPOINT=http://cdio-ai-price-element:5005/price-element
+          docker run --network changedet-network -d -e "LOG_LEVEL=TRACE" --name cdio-ai-price-element --hostname cdio-ai-price-element -p 5005:5005 --rm dgtlmoon/changedetection.io-ai:latest
+
       - name: Spin up ancillary SMTP+Echo message test server
         run: |
           # Debug SMTP server/echo message back server
@@ -95,6 +99,11 @@ jobs:
           # Settings headers playwright tests - Call back in from Sockpuppetbrowser, check headers
           docker run --name "changedet" --hostname changedet --rm -e "FLASK_SERVER_NAME=changedet" -e "PLAYWRIGHT_DRIVER_URL=ws://sockpuppetbrowser:3000?dumpio=true" --network changedet-network test-changedetectionio  bash -c 'find .; cd changedetectionio; pytest --live-server-host=0.0.0.0  --live-server-port=5004 tests/test_request.py; pwd;find .'
 
+# PLAYWRIGHT/NODE-> CDP
+      - name: ML/AI Price element scraper via Playwright+dgtlmoon/changedetection.io-ai
+        run: |
+          docker run --rm -e "FLASK_SERVER_NAME=cdio" -e "PLAYWRIGHT_DRIVER_URL=ws://sockpuppetbrowser:3000" -e "PRICE_SCRAPER_ML_ENDPOINT=ws://cdio-ai-price-element:5005/price-element" --network changedet-network --hostname=cdio test-changedetectionio  bash -c 'cd changedetectionio;pytest --live-server-host=0.0.0.0 --live-server-port=5004 tests/ml_price_scraper/test_scrape_price_element.py'
+
       - name: Playwright and SocketPuppetBrowser - Restock detection
         run: |                            
           # restock detection via playwright - added name=changedet here so that playwright and sockpuppetbrowser can connect to it

changedetectionio/__init__.py

@@ -2,7 +2,7 @@
 
 # Read more https://github.com/dgtlmoon/changedetection.io/wiki
 
-__version__ = '0.46.01'
+__version__ = '0.46.04'
 
 from changedetectionio.strtobool import strtobool
 from json.decoder import JSONDecodeError

changedetectionio/blueprint/browser_steps/__init__.py

@@ -85,7 +85,8 @@ def construct_blueprint(datastore: ChangeDetectionStore):
         browsersteps_start_session['browserstepper'] = browser_steps.browsersteps_live_ui(
             playwright_browser=browsersteps_start_session['browser'],
             proxy=proxy,
-            start_url=datastore.data['watching'][watch_uuid].get('url')
+            start_url=datastore.data['watching'][watch_uuid].get('url'),
+            headers=datastore.data['watching'][watch_uuid].get('headers')
         )
 
         # For test

changedetectionio/content_fetchers/__init__.py

@@ -4,7 +4,7 @@ from loguru import logger
 from changedetectionio.content_fetchers.exceptions import BrowserStepsStepException
 import os
 
-visualselector_xpath_selectors = 'div,span,form,table,tbody,tr,td,a,p,ul,li,h1,h2,h3,h4,header,footer,section,article,aside,details,main,nav,section,summary'
+visualselector_xpath_selectors = 'div,span,form,table,tbody,tr,td,a,p,ul,li,h1,h2,h3,h4,header,footer,section,article,aside,details,main,nav,section,summary,bdi,strong'
 
 # available_fetchers() will scan this implementation looking for anything starting with html_
 # this information is used in the form selections

changedetectionio/content_fetchers/base.py

@@ -65,8 +65,8 @@ class Fetcher():
 
     def __init__(self):
         import importlib.resources
-        self.xpath_element_js = importlib.resources.files("changedetectionio.content_fetchers.res").joinpath('xpath_element_scraper.js').read_text()
-        self.instock_data_js = importlib.resources.files("changedetectionio.content_fetchers.res").joinpath('stock-not-in-stock.js').read_text()
+        self.xpath_element_js = importlib.resources.files("changedetectionio.content_fetchers.res").joinpath('xpath_element_scraper.js').read_text(encoding='utf-8')
+        self.instock_data_js = importlib.resources.files("changedetectionio.content_fetchers.res").joinpath('stock-not-in-stock.js').read_text(encoding='utf-8')
 
     @abstractmethod
     def get_error(self):
@@ -81,7 +81,8 @@ class Fetcher():
             request_method,
             ignore_status_codes=False,
             current_include_filters=None,
-            is_binary=False):
+            is_binary=False,
+            empty_pages_are_a_change=False):
         # Should set self.error, self.status_code and self.content
         pass
 

changedetectionio/content_fetchers/playwright.py

@@ -83,7 +83,8 @@ class fetcher(Fetcher):
             request_method,
             ignore_status_codes=False,
             current_include_filters=None,
-            is_binary=False):
+            is_binary=False,
+            empty_pages_are_a_change=False):
 
         from playwright.sync_api import sync_playwright
         import playwright._impl._errors
@@ -130,7 +131,7 @@ class fetcher(Fetcher):
             if response is None:
                 context.close()
                 browser.close()
-                logger.debug("Content Fetcher > Response object was none")
+                logger.debug("Content Fetcher > Response object from the browser communication was none")
                 raise EmptyReply(url=url, status_code=None)
 
             try:
@@ -166,10 +167,10 @@ class fetcher(Fetcher):
 
                 raise Non200ErrorCodeReceived(url=url, status_code=self.status_code, screenshot=screenshot)
 
-            if len(self.page.content().strip()) == 0:
+            if not empty_pages_are_a_change and len(self.page.content().strip()) == 0:
+                logger.debug("Content Fetcher > Content was empty, empty_pages_are_a_change = False")
                 context.close()
                 browser.close()
-                logger.debug("Content Fetcher > Content was empty")
                 raise EmptyReply(url=url, status_code=response.status)
 
             # Run Browser Steps here

changedetectionio/content_fetchers/puppeteer.py

@@ -75,7 +75,8 @@ class fetcher(Fetcher):
                          request_method,
                          ignore_status_codes,
                          current_include_filters,
-                         is_binary
+                         is_binary,
+                         empty_pages_are_a_change
                          ):
 
         from changedetectionio.content_fetchers import visualselector_xpath_selectors
@@ -153,7 +154,7 @@ class fetcher(Fetcher):
         if response is None:
             await self.page.close()
             await browser.close()
-            logger.warning("Content Fetcher > Response object was none")
+            logger.warning("Content Fetcher > Response object was none (as in, the response from the browser was empty, not just the content)")
             raise EmptyReply(url=url, status_code=None)
 
         self.headers = response.headers
@@ -186,10 +187,11 @@ class fetcher(Fetcher):
 
             raise Non200ErrorCodeReceived(url=url, status_code=self.status_code, screenshot=screenshot)
         content = await self.page.content
-        if len(content.strip()) == 0:
+
+        if not empty_pages_are_a_change and len(content.strip()) == 0:
+            logger.error("Content Fetcher > Content was empty (empty_pages_are_a_change is False), closing browsers")
             await self.page.close()
             await browser.close()
-            logger.error("Content Fetcher > Content was empty")
             raise EmptyReply(url=url, status_code=response.status)
 
         # Run Browser Steps here
@@ -247,7 +249,7 @@ class fetcher(Fetcher):
         await self.fetch_page(**kwargs)
 
     def run(self, url, timeout, request_headers, request_body, request_method, ignore_status_codes=False,
-            current_include_filters=None, is_binary=False):
+            current_include_filters=None, is_binary=False, empty_pages_are_a_change=False):
 
         #@todo make update_worker async which could run any of these content_fetchers within memory and time constraints
         max_time = os.getenv('PUPPETEER_MAX_PROCESSING_TIMEOUT_SECONDS', 180)
@@ -262,7 +264,8 @@ class fetcher(Fetcher):
                 request_method=request_method,
                 ignore_status_codes=ignore_status_codes,
                 current_include_filters=current_include_filters,
-                is_binary=is_binary
+                is_binary=is_binary,
+                empty_pages_are_a_change=empty_pages_are_a_change
             ), timeout=max_time))
         except asyncio.TimeoutError:
             raise(BrowserFetchTimedOut(msg=f"Browser connected but was unable to process the page in {max_time} seconds."))

changedetectionio/content_fetchers/requests.py

@@ -1,9 +1,8 @@
+from loguru import logger
+import chardet
 import hashlib
 import os
-
-import chardet
 import requests
-
 from changedetectionio import strtobool
 from changedetectionio.content_fetchers.exceptions import BrowserStepsInUnsupportedFetcher, EmptyReply, Non200ErrorCodeReceived
 from changedetectionio.content_fetchers.base import Fetcher
@@ -26,7 +25,8 @@ class fetcher(Fetcher):
             request_method,
             ignore_status_codes=False,
             current_include_filters=None,
-            is_binary=False):
+            is_binary=False,
+            empty_pages_are_a_change=False):
 
         if self.browser_steps_get_valid_steps():
             raise BrowserStepsInUnsupportedFetcher(url=url)
@@ -53,7 +53,7 @@ class fetcher(Fetcher):
             session.mount('file://', FileAdapter())
 
         r = session.request(method=request_method,
-                            data=request_body,
+                            data=request_body.encode('utf-8') if type(request_body) is str else request_body,
                             url=url,
                             headers=request_headers,
                             timeout=timeout,
@@ -74,7 +74,10 @@ class fetcher(Fetcher):
         self.headers = r.headers
 
         if not r.content or not len(r.content):
-            raise EmptyReply(url=url, status_code=r.status_code)
+            if not empty_pages_are_a_change:
+                raise EmptyReply(url=url, status_code=r.status_code)
+            else:
+                logger.debug(f"URL {url} gave zero byte content reply with Status Code {r.status_code}, but empty_pages_are_a_change = True")
 
         # @todo test this
         # @todo maybe you really want to test zero-byte return pages?

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

@@ -75,6 +75,7 @@ function isItemInStock() {
         'vergriffen',
         'vorbestellen',
         'vorbestellung ist bald möglich',
+        'we don\'t currently have any',
         'we couldn\'t find any products that match',
         'we do not currently have an estimate of when this product will be back in stock.',
         'we don\'t know when or if this item will be back in stock.',
@@ -173,7 +174,8 @@ function isItemInStock() {
         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 + 150 || element.getBoundingClientRect().top + window.scrollY <= 100) {
+        // 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) {
             continue
         }
         elementText = "";
@@ -187,7 +189,7 @@ function isItemInStock() {
             // and these mean its out of stock
             for (const outOfStockText of outOfStockTexts) {
                 if (elementText.includes(outOfStockText)) {
-                    console.log(`Selected 'Out of Stock' - found text "${outOfStockText}" - "${elementText}"`)
+                    console.log(`Selected 'Out of Stock' - found text "${outOfStockText}" - "${elementText}" - offset top ${element.getBoundingClientRect().top}, page height is ${vh}`)
                     return outOfStockText; // item is out of stock
                 }
             }

changedetectionio/content_fetchers/res/xpath_element_scraper.js

@@ -15,6 +15,7 @@ try {
     console.log(e);
 }
 
+const percentageNumerical = str => Math.round((str.match(/\d/g) || []).length / str.length * 100);
 
 // Include the getXpath script directly, easier than fetching
 function getxpath(e) {
@@ -77,6 +78,30 @@ const findUpTag = (el) => {
     }
     return null;
 }
+// Text width scraper for ML training/detection
+// Create a single canvas and get its 2D context
+const canvas = document.createElement("canvas");
+const context = canvas.getContext("2d");
+
+// Function to get the width and height of the text inside an element and round them to the nearest integer
+function getTextWidthAndHeightinPx(element) {
+    // Set the font to match the style of the text in the element
+    context.font = window.getComputedStyle(element).font;
+
+    // Get the text inside the element
+    const text = element.textContent || element.innerText;
+
+    // Measure the text width
+    const metrics = context.measureText(text);
+    const width = Math.round(metrics.width);
+
+    // Get the font size from the computed style
+    const fontSize = parseFloat(window.getComputedStyle(element).fontSize);
+    const height = Math.round(fontSize); // Using font size as an approximation of height
+
+    // Return both width and height as an object
+    return { textWidth: width, textHeight: height };
+}
 
 
 // @todo - if it's SVG or IMG, go into image diff mode
@@ -122,8 +147,10 @@ const visibleElementsArray = [];
 // Call collectVisibleElements with the starting parent element
 collectVisibleElements(document.body, visibleElementsArray);
 
+// Append any custom selectors to the visibleElementsArray
 
-visibleElementsArray.forEach(function (element) {
+
+function get_element_metadata(element) {
 
     bbox = element.getBoundingClientRect();
 
@@ -164,18 +191,68 @@ visibleElementsArray.forEach(function (element) {
         }
     }
 
+    let label = "none" // A placeholder, the actual labels for training are done by hand for now
 
-    size_pos.push({
+    // Check if the element was found and get its text , not including any child element
+    let text = Array.from(element.childNodes)
+        .filter(node => node.nodeType === Node.TEXT_NODE)
+        .map(node => node.textContent)
+        .join('');
+
+    // Remove any gaps in sequences of newlines and tabs inside the string
+    text = text.trim().replace(/[\s\t\n\r]{2,}/g, ' ').trim();
+
+    // Try to identify any possible currency amounts "Sale: 4000" or "Sale now 3000 Kc", can help with the training.
+    // @todo could be instead of USD/AUD etc [A-Z]{2,3} ?
+    //const hasDigitCurrency = (/\d/.test(text.slice(0, 6)) || /\d/.test(text.slice(-6)) ) &&  /([€£$¥₩₹]|USD|AUD|EUR|Kč|kr|SEK|RM|,–)/.test(text) ;
+    const hasDigitCurrency = (/\d/.test(text.slice(0, 6)) || /\d/.test(text.slice(-6)) ) &&  /([€£$¥₩₹]|USD|AUD|EUR|Kč|kr|SEK|RM|,–)/.test(text) ;
+    const hasDigit = /[0-9]/.test(text) ;
+
+    // Sizing of the actual text inside the element can be very different from the elements size
+    const { textWidth, textHeight } = getTextWidthAndHeightinPx(element);
+
+    const computedStyle = window.getComputedStyle(element);
+    let red, green, blue;
+
+    if (text.length) {
+        // Extract the RGB values from the color string (format: rgb(r, g, b))
+        [red, green, blue] = computedStyle.color.match(/\d+/g).map(Number);
+    } else {
+        // Assign default values if text is empty
+        [red, green, blue] = [0, 0, 0];
+    }
+    return {
         xpath: xpath_result,
         width: Math.round(bbox['width']),
         height: Math.round(bbox['height']),
         left: Math.floor(bbox['left']),
         top: Math.floor(bbox['top']) + scroll_y,
+        // tagName used by Browser Steps
         tagName: (element.tagName) ? element.tagName.toLowerCase() : '',
+        // tagtype used by Browser Steps
         tagtype: (element.tagName.toLowerCase() === 'input' && element.type) ? element.type.toLowerCase() : '',
-        isClickable: window.getComputedStyle(element).cursor == "pointer"
-    });
+        isClickable: window.getComputedStyle(element).cursor === "pointer",
+        // Used by the keras/pytorch trainer
+        fontSize: window.getComputedStyle(element).getPropertyValue('font-size'),
+        fontWeight: window.getComputedStyle(element).getPropertyValue('font-weight'),
+        pcNumerical: text.length && percentageNumerical(text),
+        hasDigit: hasDigit,
+        hasDigitCurrency: hasDigitCurrency,
+        textWidth: textWidth,
+        textHeight: textHeight,
+        textLength: text.length,
+        t_r: red,
+        t_g: green,
+        t_b: blue,
+        label: label,
+    };
+}
 
+visibleElementsArray.forEach(function (element) {
+    let metadata = get_element_metadata(element);
+    if(metadata) {
+        size_pos.push(metadata);
+    }
 });
 
 
@@ -184,7 +261,19 @@ visibleElementsArray.forEach(function (element) {
 if (include_filters.length) {
     let results;
     // Foreach filter, go and find it on the page and add it to the results so we can visualise it again
+    outerLoop:
     for (const f of include_filters) {
+        // Quick check so we dont end up with duplicates in the training data
+        for (let index = 0; index < size_pos.length; index++) {
+            let item = size_pos[index];
+            if (item.xpath === f) {
+                item.highlight_as_custom_filter = true;
+                item.found_as_duplicate = true;
+                item.label = "price";
+                continue outerLoop;
+            }
+        }
+
         bbox = false;
         q = false;
 
@@ -205,7 +294,6 @@ if (include_filters.length) {
                 }
             } else {
                 console.log("[css] Scanning for included filter " + f)
-                console.log("[css] Scanning for included filter " + f);
                 results = document.querySelectorAll(f);
             }
         } catch (e) {
@@ -242,17 +330,15 @@ if (include_filters.length) {
                         console.log("xpath_element_scraper: error looking up q.ownerElement")
                     }
                 }
-
-                if (bbox && bbox['width'] > 0 && bbox['height'] > 0) {
-                    size_pos.push({
-                        xpath: f,
-                        width: parseInt(bbox['width']),
-                        height: parseInt(bbox['height']),
-                        left: parseInt(bbox['left']),
-                        top: parseInt(bbox['top']) + scroll_y,
-                        highlight_as_custom_filter: true
-                    });
+                element_info = get_element_metadata(node);
+                if(element_info) {
+                    // Be sure we use exactly what was written
+                    element_info['xpath'] = f;
+                    element_info['highlight_as_custom_filter'] = true;
+                    element_info['label'] = "price";
+                    size_pos.push(element_info);
                 }
+
             });
         }
     }

changedetectionio/content_fetchers/webdriver_selenium.py

@@ -56,7 +56,8 @@ class fetcher(Fetcher):
             request_method,
             ignore_status_codes=False,
             current_include_filters=None,
-            is_binary=False):
+            is_binary=False,
+            empty_pages_are_a_change=False):
 
         from selenium import webdriver
         from selenium.webdriver.chrome.options import Options as ChromeOptions

changedetectionio/flask_app.py

@@ -792,9 +792,9 @@ def changedetection_app(config=None, datastore_o=None):
             # Re #286 - We wait for syncing new data to disk in another thread every 60 seconds
             # But in the case something is added we should save straight away
             datastore.needs_write_urgent = True
-
-            # Queue the watch for immediate recheck, with a higher priority
-            update_q.put(queuedWatchMetaData.PrioritizedItem(priority=1, item={'uuid': uuid, 'skip_when_checksum_same': False}))
+            if not datastore.data['watching'][uuid].get('paused'):
+                # Queue the watch for immediate recheck, with a higher priority
+                update_q.put(queuedWatchMetaData.PrioritizedItem(priority=1, item={'uuid': uuid, 'skip_when_checksum_same': False}))
 
             # Diff page [edit] link should go back to diff page
             if request.args.get("next") and request.args.get("next") == 'diff':
@@ -1377,17 +1377,19 @@ def changedetection_app(config=None, datastore_o=None):
         import brotli
 
         watch = datastore.data['watching'].get(uuid)
-        if watch and os.path.isdir(watch.watch_data_dir):
-            latest_filename = list(watch.history.keys())[0]
+        if watch and watch.history.keys() and os.path.isdir(watch.watch_data_dir):
+            latest_filename = list(watch.history.keys())[-1]
             html_fname = os.path.join(watch.watch_data_dir, f"{latest_filename}.html.br")
-            if html_fname.endswith('.br'):
-                # Read and decompress the Brotli file
-                with open(html_fname, 'rb') as f:
+            with open(html_fname, 'rb') as f:
+                if html_fname.endswith('.br'):
+                    # Read and decompress the Brotli file
                     decompressed_data = brotli.decompress(f.read())
+                else:
+                    decompressed_data = f.read()
 
-                buffer = BytesIO(decompressed_data)
+            buffer = BytesIO(decompressed_data)
 
-                return send_file(buffer, as_attachment=True, download_name=f"{latest_filename}.html", mimetype='text/html')
+            return send_file(buffer, as_attachment=True, download_name=f"{latest_filename}.html", mimetype='text/html')
 
 
         # Return a 500 error
@@ -1600,6 +1602,15 @@ def changedetection_app(config=None, datastore_o=None):
 
             flash(f"{len(uuids)} watches were tagged")
 
+        elif op.startswith('mode:'):
+            mode = op.replace('mode:','')
+            for uuid in uuids:
+                uuid = uuid.strip()
+                if datastore.data['watching'].get(uuid):
+                    datastore.data['watching'][uuid]['processor'] = mode
+            flash(f"{len(uuids)} watches changed modes")
+
+
         return redirect(url_for('index'))
 
     @app.route("/api/share-url", methods=['GET'])

changedetectionio/model/Watch.py

@@ -518,7 +518,7 @@ class model(watch_base):
         self.ensure_data_dir_exists()
 
         with open(target_path, 'w') as f:
-            f.write(json.dumps(data))
+            f.write(json.dumps(data, indent=2))
             f.close()
 
     # Save as PNG, PNG is larger but better for doing visual diff in the future

changedetectionio/notification.py

@@ -107,7 +107,7 @@ def apprise_custom_api_call_wrapper(body, title, notify_type, *args, **kwargs):
 
     r(results.get('url'),
       auth=auth,
-      data=body,
+      data=body.encode('utf-8') if type(body) is str else body,
       headers=headers,
       params=params
       )

changedetectionio/processors/__init__.py

@@ -26,6 +26,8 @@ class difference_detection_processor():
 
     def call_browser(self):
         from requests.structures import CaseInsensitiveDict
+        from changedetectionio.content_fetchers.exceptions import EmptyReply
+
         # Protect against file:// access
         if re.search(r'^file://', self.watch.get('url', '').strip(), re.IGNORECASE):
             if not strtobool(os.getenv('ALLOW_FILE_URI', 'false')):
@@ -133,8 +135,18 @@ class difference_detection_processor():
         is_binary = self.watch.is_pdf
 
         # And here we go! call the right browser with browser-specific settings
-        self.fetcher.run(url, timeout, request_headers, request_body, request_method, ignore_status_codes, self.watch.get('include_filters'),
-                    is_binary=is_binary)
+        empty_pages_are_a_change = self.datastore.data['settings']['application'].get('empty_pages_are_a_change', False)
+
+        self.fetcher.run(url=url,
+                         timeout=timeout,
+                         request_headers=request_headers,
+                         request_body=request_body,
+                         request_method=request_method,
+                         ignore_status_codes=ignore_status_codes,
+                         current_include_filters=self.watch.get('include_filters'),
+                         is_binary=is_binary,
+                         empty_pages_are_a_change=empty_pages_are_a_change
+                         )
 
         #@todo .quit here could go on close object, so we can run JS if change-detected
         self.fetcher.quit()

changedetectionio/processors/restock_diff/__init__.py

@@ -1,11 +1,12 @@
 
-from changedetectionio.model.Watch import model as BaseWatch
-import re
 from babel.numbers import parse_decimal
+from changedetectionio.model.Watch import model as BaseWatch
+from typing import Union
+import re
 
 class Restock(dict):
 
-    def parse_currency(self, raw_value: str) -> float:
+    def parse_currency(self, raw_value: str) -> Union[float, None]:
         # Clean and standardize the value (ie 1,400.00 should be 1400.00), even better would be store the whole thing as an integer.
         standardized_value = raw_value
 
@@ -21,8 +22,11 @@ class Restock(dict):
         # Remove any non-numeric characters except for the decimal point
         standardized_value = re.sub(r'[^\d.-]', '', standardized_value)
 
-        # Convert to float
-        return float(parse_decimal(standardized_value, locale='en'))
+        if standardized_value:
+            # Convert to float
+            return float(parse_decimal(standardized_value, locale='en'))
+
+        return None
 
     def __init__(self, *args, **kwargs):
         # Define default values

changedetectionio/processors/restock_diff/processor.py

@@ -3,10 +3,13 @@ from ..exceptions import ProcessorException
 from . import Restock
 from loguru import logger
 import hashlib
+import os
 import re
 import urllib3
 import time
 
+from ...html_tools import html_to_text
+
 urllib3.disable_warnings(urllib3.exceptions.InsecureRequestWarning)
 name = 'Re-stock & Price detection for single product pages'
 description = 'Detects if the product goes back to in-stock'
@@ -34,23 +37,28 @@ def get_itemprop_availability(html_content) -> Restock:
     Kind of funny/cool way to find price/availability in one many different possibilities.
     Use 'extruct' to find any possible RDFa/microdata/json-ld data, make a JSON string from the output then search it.
     """
+
     from jsonpath_ng import parse
 
     now = time.time()
     import extruct
     logger.trace(f"Imported extruct module in {time.time() - now:.3f}s")
 
-    value = {}
     now = time.time()
+
     # Extruct is very slow, I'm wondering if some ML is going to be faster (800ms on my i7), 'rdfa' seems to be the heaviest.
-
     syntaxes = ['dublincore', 'json-ld', 'microdata', 'microformat', 'opengraph']
+    try:
+        data = extruct.extract(html_content, syntaxes=syntaxes)
+    except Exception as e:
+        logger.warning(f"Unable to extract data, document parsing with extruct failed with {type(e).__name__} - {str(e)}")
+        return Restock()
 
-    data = extruct.extract(html_content, syntaxes=syntaxes)
     logger.trace(f"Extruct basic extract of all metadata done in {time.time() - now:.3f}s")
 
     # First phase, dead simple scanning of anything that looks useful
     value = Restock()
+    return value
     if data:
         logger.debug(f"Using jsonpath to find price/availability/etc")
         price_parse = parse('$..(price|Price)')
@@ -118,6 +126,52 @@ class perform_site_check(difference_detection_processor):
     screenshot = None
     xpath_data = None
 
+    def ML_scrape_for_price_data(self, ML_price_scraper_url):
+        import requests
+        from changedetectionio import html_tools
+
+        price_info = None
+
+        # Perform the POST request
+        response = requests.post(ML_price_scraper_url, json=self.fetcher.xpath_data)
+        logger.debug(f"ML Price scraper - {ML_price_scraper_url} Response OK? - '{response.ok}'")
+        # Check if the response contains a dict
+        if response.ok:  # This checks if the request was successful (status code 200-299)
+            response_json = response.json()
+            logger.debug(f"ML Price scraper: response - {response_json}'")
+            if isinstance(response_json, dict) and 'idx' in response_json.keys():
+                suggested_xpath_idx = response_json.get('idx')
+                if response_json.get('score') <0.80 or response_json.get('score') > 1.0:
+                    logger.warning(f"Predict score was outside normal range, aborting ML/AI price check, needs better training data in this case?")
+                    return None
+
+                # Use the path provided to extra the price text
+                from price_parser import Price
+                scrape_element = self.fetcher.xpath_data.get('size_pos', {})[suggested_xpath_idx]
+                logger.debug(f"Predicted selector with price information is {scrape_element['xpath']}")
+
+                result_s = None
+                if scrape_element['xpath'][0] == '/' or scrape_element['xpath'].startswith('xpath'):
+                    result_s = html_tools.xpath_filter(xpath_filter=scrape_element['xpath'],
+                                                       html_content=self.fetcher.content)
+                else:
+                    # CSS Filter, extract the HTML that matches and feed that into the existing inscriptis::get_text
+                    result_s = html_tools.include_filters(include_filters=scrape_element['xpath'],
+                                                          html_content=self.fetcher.content)
+
+                if result_s:
+                    text = html_to_text(result_s)
+                    logger.debug(f"Guessed the text '{text}' as the price information")
+                    if text:
+                        price_info = Price.fromstring(text)
+            else:
+                logger.error(f"ML Price scraper: missing xpath index (IDX) in response?")
+        else:
+            print(f"ML Price scraper: Request failed with status code: {response.status_code}")
+
+#@TODO THROW HELPFUL MESSAGE WITH LINK TO TUTORIAL IF IT CANT CONNECT!
+        return price_info
+
     def run_changedetection(self, watch, skip_when_checksum_same=True):
         if not watch:
             raise Exception("Watch no longer exists.")
@@ -174,6 +228,21 @@ class perform_site_check(difference_detection_processor):
                 else:
                     update_obj['restock']['in_stock'] = False
 
+        # Attempt to pass the elements off to the machine-learning endpoint if its enabled
+        # This might return a confident guess as to which element contains the price data
+        if not itemprop_availability.get('price'):
+            ML_price_scraper_url = os.getenv("PRICE_SCRAPER_ML_ENDPOINT")
+            if self.fetcher.xpath_data and ML_price_scraper_url:
+                price_info = self.ML_scrape_for_price_data(ML_price_scraper_url)
+                if price_info and price_info.amount:
+                    logger.success(f"ML Price scraper: Got price data {price_info}")
+                    itemprop_availability['price'] = f"{price_info.amount}"
+                    update_obj['restock']['price'] = f"{price_info.amount}"
+                if price_info and price_info.currency:
+                    itemprop_availability['currency'] = price_info.currency
+                    update_obj['restock']['currency'] = price_info.currency
+
+
         # Main detection method
         fetched_md5 = None
 

changedetectionio/templates/settings.html

@@ -76,7 +76,7 @@
                     </div>
                     <div class="pure-control-group">
                         {{ render_checkbox_field(form.application.form.empty_pages_are_a_change) }}
-                        <span class="pure-form-message-inline">When a page contains HTML, but no renderable text appears (empty page), is this considered a change?</span>
+                        <span class="pure-form-message-inline">When a request returns no content, or the HTML does not contain any text, is this considered a change?</span>
                     </div>
                 {% if form.requests.proxy %}
                     <div class="pure-control-group inline-radio">

changedetectionio/templates/watch-overview.html

@@ -37,6 +37,8 @@
         <button class="pure-button button-secondary button-xsmall" name="op" value="assign-tag" id="checkbox-assign-tag">Tag</button>
         <button class="pure-button button-secondary button-xsmall" name="op" value="mark-viewed">Mark viewed</button>
         <button class="pure-button button-secondary button-xsmall" name="op" value="notification-default">Use default notification</button>
+        <button class="pure-button button-secondary button-xsmall" name="op" value="mode:text_json_diff">Mode: Page changes</button>
+        <button class="pure-button button-secondary button-xsmall" name="op" value="mode:restock_diff">Mode: Price/Restock</button>
         <button class="pure-button button-secondary button-xsmall" name="op" value="clear-errors">Clear errors</button>
         <button class="pure-button button-secondary button-xsmall" style="background: #dd4242;" name="op" value="clear-history">Clear/reset history</button>
         <button class="pure-button button-secondary button-xsmall" style="background: #dd4242;" name="op" value="delete">Delete</button>

changedetectionio/tests/conftest.py

@@ -8,7 +8,7 @@ from changedetectionio import changedetection_app
 from changedetectionio import store
 import os
 import sys
-from loguru import logger
+
 
 # https://github.com/pallets/flask/blob/1.1.2/examples/tutorial/tests/test_auth.py
 # Much better boilerplate than the docs

changedetectionio/tests/fetchers/test_content.py

@@ -1,6 +1,5 @@
 #!/usr/bin/env python3
 
-import time
 from flask import url_for
 from ..util import live_server_setup, wait_for_all_checks
 import logging

changedetectionio/tests/ml_price_scraper/test_scrape_price_element.py

@@ -0,0 +1,127 @@
+import os
+
+from flask import url_for
+from changedetectionio.tests.util import set_original_response, set_modified_response, set_more_modified_response, live_server_setup, \
+    wait_for_all_checks, \
+    set_longer_modified_response
+
+import time
+
+# No semantic data just some text, we should be able to find the product price.
+def set_response(price="121.95"):
+    html_content = f"""
+    <!DOCTYPE html>
+    <html lang="en">
+    <head>
+        <meta charset="UTF-8">
+        <meta name="viewport" content="width=device-width, initial-scale=1.0">
+        <title>Ajax Widget</title>
+        <style>
+            body {{
+                font-family: Arial, sans-serif;
+                margin: 0;
+                padding: 0;
+                display: flex;
+                justify-content: center;
+                align-items: center;
+                height: 100vh;
+                background-color: #f4f4f4;
+            }}
+            .container {{
+                display: flex;
+                flex-direction: row;
+                background-color: #fff;
+                border: 1px solid #ddd;
+                padding: 20px;
+                border-radius: 5px;
+                box-shadow: 0 0 10px rgba(0, 0, 0, 0.1);
+                width: 80%;
+                max-width: 800px;
+            }}
+            .description {{
+                flex: 2;
+                margin-right: 20px;
+            }}
+            .description h1 {{
+                margin-top: 0;
+            }}
+            .price {{
+                flex: 1;
+                text-align: right;
+                font-size: 24px;
+                color: #333;
+            }}
+            .price span {{
+                font-size: 32px;
+                font-weight: bold;
+            }}
+            .buy-button {{
+                display: inline-block;
+                margin-top: 20px;
+                padding: 10px 20px;
+                background-color: #28a745;
+                color: #fff;
+                text-decoration: none;
+                border-radius: 5px;
+                font-size: 16px;
+            }}
+            .buy-button:hover {{
+                background-color: #218838;
+            }}
+        </style>
+    </head>
+    <body>
+        <div class="container">
+            <div class="description">
+                <h1>Ajax Widget</h1>
+                <p>The Ajax Widget is the ultimate solution for all your widget needs. Crafted with precision and using the latest technology, this widget offers unmatched performance and durability. Whether you're using it for personal or professional purposes, the Ajax Widget will not disappoint. It's easy to use, reliable, and comes with a sleek design that complements any setup. Don't settle for less; get the best with the Ajax Widget today!</p>
+            </div>
+            <div class="price">
+                <span>${price}</span>
+                <br>
+                <a href="#" class="buy-button">Buy Now</a><br>
+                IN STOCK
+            </div>
+        </div>
+    </body>
+    </html>
+    """
+
+    with open("test-datastore/endpoint-content.txt", "w") as f:
+        f.write(html_content)
+    time.sleep(1)
+    return None
+
+
+
+
+def test_restock_itemprop_basic(client, live_server):
+
+    # needs to be set and something like 'ws://127.0.0.1:3000'
+    assert os.getenv('PLAYWRIGHT_DRIVER_URL'), "Needs PLAYWRIGHT_DRIVER_URL set for this test"
+    assert os.getenv('PRICE_SCRAPER_ML_ENDPOINT'), "Needs PRICE_SCRAPER_ML_ENDPOINT set for this test"
+
+
+    live_server_setup(live_server)
+
+    set_response(price="123.99")
+
+    # because it needs to access itself from within the sockpuppetbrowser
+    test_url = url_for('test_endpoint', _external=True)
+    test_url = test_url.replace('localhost.localdomain', 'cdio')
+    test_url = test_url.replace('localhost', 'cdio')
+
+    client.post(
+        url_for("form_quick_watch_add"),
+        data={"url": test_url, "tags": 'restock tests', 'processor': 'restock_diff'},
+        follow_redirects=True
+    )
+    wait_for_all_checks(client)
+    res = client.get(url_for("index"))
+
+    assert b'123.99' in res.data
+    assert b' in-stock' in res.data
+    assert b' not-in-stock' not in res.data
+
+    res = client.get(url_for("form_delete", uuid="all"), follow_redirects=True)
+    assert b'Deleted' in res.data

changedetectionio/tests/test_add_replace_remove_filter.py

@@ -112,7 +112,7 @@ def test_check_add_line_contains_trigger(client, live_server, measure_memory_usa
     res = client.post(
         url_for("settings_page"),
         data={"application-notification_title": "New ChangeDetection.io Notification - {{ watch_url }}",
-              "application-notification_body": 'triggered text was -{{triggered_text}}-',
+              "application-notification_body": 'triggered text was -{{triggered_text}}- 网站监测 内容更新了',
               # https://github.com/caronc/apprise/wiki/Notify_Custom_JSON#get-parameter-manipulation
               "application-notification_urls": test_notification_url,
               "application-minutes_between_check": 180,
@@ -167,9 +167,10 @@ def test_check_add_line_contains_trigger(client, live_server, measure_memory_usa
     # Takes a moment for apprise to fire
     time.sleep(3)
     assert os.path.isfile("test-datastore/notification.txt"), "Notification fired because I can see the output file"
-    with open("test-datastore/notification.txt", 'r') as f:
-        response= f.read()
-        assert '-Oh yes please-' in response
+    with open("test-datastore/notification.txt", 'rb') as f:
+        response = f.read()
+        assert b'-Oh yes please-' in response
+        assert '网站监测 内容更新了'.encode('utf-8') in response
 
 
     res = client.get(url_for("form_delete", uuid="all"), follow_redirects=True)

changedetectionio/tests/test_backend.py

@@ -69,6 +69,12 @@ def test_check_basic_change_detection_functionality(client, live_server, measure
 
     wait_for_all_checks(client)
 
+    uuid = extract_UUID_from_client(client)
+
+    # Check the 'get latest snapshot works'
+    res = client.get(url_for("watch_get_latest_html", uuid=uuid))
+    assert b'which has this one new line' in res.data
+
     # Now something should be ready, indicated by having a 'unviewed' class
     res = client.get(url_for("index"))
     assert b'unviewed' in res.data
@@ -86,7 +92,7 @@ def test_check_basic_change_detection_functionality(client, live_server, measure
     assert expected_url.encode('utf-8') in res.data
 
     # Following the 'diff' link, it should no longer display as 'unviewed' even after we recheck it a few times
-    res = client.get(url_for("diff_history_page", uuid="first"))
+    res = client.get(url_for("diff_history_page", uuid=uuid))
     assert b'selected=""' in res.data, "Confirm diff history page loaded"
 
     # Check the [preview] pulls the right one
@@ -143,18 +149,12 @@ def test_check_basic_change_detection_functionality(client, live_server, measure
     assert b'unviewed' not in res.data
 
     # #2458 "clear history" should make the Watch object update its status correctly when the first snapshot lands again
-    uuid = extract_UUID_from_client(client)
     client.get(url_for("clear_watch_history", uuid=uuid))
     client.get(url_for("form_watch_checknow"), follow_redirects=True)
     wait_for_all_checks(client)
     res = client.get(url_for("index"))
     assert b'preview/' in res.data
 
-
-    # Check the 'get latest snapshot works'
-    res = client.get(url_for("watch_get_latest_html", uuid=uuid))
-    assert b'<head><title>head title</title></head>' in res.data
-
     #
     # Cleanup everything
     res = client.get(url_for("form_delete", uuid="all"), follow_redirects=True)

changedetectionio/tests/test_nonrenderable_pages.py

@@ -1,12 +1,8 @@
 #!/usr/bin/env python3
 
-import time
 from flask import url_for
-from urllib.request import urlopen
-from .util import set_original_response, set_modified_response, live_server_setup
-
-sleep_time_for_fetch_thread = 3
-
+from .util import set_original_response, set_modified_response, live_server_setup, wait_for_all_checks
+import time
 
 def set_nonrenderable_response():
     test_return_data = """<html>
@@ -22,6 +18,13 @@ def set_nonrenderable_response():
 
     return None
 
+def set_zero_byte_response():
+
+    with open("test-datastore/endpoint-content.txt", "w") as f:
+        f.write("")
+
+    return None
+
 def test_check_basic_change_detection_functionality(client, live_server, measure_memory_usage):
     set_original_response()
     live_server_setup(live_server)
@@ -35,18 +38,11 @@ def test_check_basic_change_detection_functionality(client, live_server, measure
 
     assert b"1 Imported" in res.data
 
-    time.sleep(sleep_time_for_fetch_thread)
+    wait_for_all_checks(client)
 
-    # Do this a few times.. ensures we dont accidently set the status
-    for n in range(3):
-        client.get(url_for("form_watch_checknow"), follow_redirects=True)
-
-        # Give the thread time to pick it up
-        time.sleep(sleep_time_for_fetch_thread)
-
-        # It should report nothing found (no new 'unviewed' class)
-        res = client.get(url_for("index"))
-        assert b'unviewed' not in res.data
+    # It should report nothing found (no new 'unviewed' class)
+    res = client.get(url_for("index"))
+    assert b'unviewed' not in res.data
 
 
     #####################
@@ -64,7 +60,7 @@ def test_check_basic_change_detection_functionality(client, live_server, measure
     client.get(url_for("form_watch_checknow"), follow_redirects=True)
 
     # Give the thread time to pick it up
-    time.sleep(sleep_time_for_fetch_thread)
+    wait_for_all_checks(client)
 
     # It should report nothing found (no new 'unviewed' class)
     res = client.get(url_for("index"))
@@ -86,14 +82,21 @@ def test_check_basic_change_detection_functionality(client, live_server, measure
     client.get(url_for("form_watch_checknow"), follow_redirects=True)
 
     # Give the thread time to pick it up
-    time.sleep(sleep_time_for_fetch_thread)
+    wait_for_all_checks(client)
 
     # It should report nothing found (no new 'unviewed' class)
     res = client.get(url_for("index"))
     assert b'unviewed' in res.data
+    client.get(url_for("mark_all_viewed"), follow_redirects=True)
 
-
-
+    # A totally zero byte (#2528) response should also not trigger an error
+    set_zero_byte_response()
+    time.sleep(2)
+    client.get(url_for("form_watch_checknow"), follow_redirects=True)
+    wait_for_all_checks(client)
+    res = client.get(url_for("index"))
+    assert b'unviewed' in res.data # A change should have registered because empty_pages_are_a_change is ON
+    assert b'fetch-error' not in res.data
 
     #
     # Cleanup everything

changedetectionio/tests/test_notification.py

@@ -291,11 +291,11 @@ def test_notification_custom_endpoint_and_jinja2(client, live_server, measure_me
         data={
               "application-fetch_backend": "html_requests",
               "application-minutes_between_check": 180,
-              "application-notification_body": '{ "url" : "{{ watch_url }}", "secret": 444 }',
+              "application-notification_body": '{ "url" : "{{ watch_url }}", "secret": 444, "somebug": "网站监测 内容更新了" }',
               "application-notification_format": default_notification_format,
               "application-notification_urls": test_notification_url,
               # https://github.com/caronc/apprise/wiki/Notify_Custom_JSON#get-parameter-manipulation
-              "application-notification_title": "New ChangeDetection.io Notification - {{ watch_url }}",
+              "application-notification_title": "New ChangeDetection.io Notification - {{ watch_url }} ",
               },
         follow_redirects=True
     )
@@ -324,6 +324,7 @@ def test_notification_custom_endpoint_and_jinja2(client, live_server, measure_me
         j = json.loads(x)
         assert j['url'].startswith('http://localhost')
         assert j['secret'] == 444
+        assert j['somebug'] == '网站监测 内容更新了'
 
     # URL check, this will always be converted to lowercase
     assert os.path.isfile("test-datastore/notification-url.txt")
@@ -354,9 +355,10 @@ def test_notification_custom_endpoint_and_jinja2(client, live_server, measure_me
 #2510
 def test_global_send_test_notification(client, live_server, measure_memory_usage):
 
-
     #live_server_setup(live_server)
     set_original_response()
+    if os.path.isfile("test-datastore/notification.txt"):
+        os.unlink("test-datastore/notification.txt")
 
     # otherwise other settings would have already existed from previous tests in this file
     res = client.post(
@@ -364,7 +366,8 @@ def test_global_send_test_notification(client, live_server, measure_memory_usage
         data={
             "application-fetch_backend": "html_requests",
             "application-minutes_between_check": 180,
-            "application-notification_body": 'change detection is cool',
+            #1995 UTF-8 content should be encoded
+            "application-notification_body": 'change detection is cool 网站监测 内容更新了',
             "application-notification_format": default_notification_format,
             "application-notification_urls": "",
             "application-notification_title": "New ChangeDetection.io Notification - {{ watch_url }}",
@@ -399,8 +402,7 @@ def test_global_send_test_notification(client, live_server, measure_memory_usage
 
     with open("test-datastore/notification.txt", 'r') as f:
         x = f.read()
-        assert 'change detection is coo' in x
-
+        assert 'change detection is cool 网站监测 内容更新了' in x
 
     os.unlink("test-datastore/notification.txt")
 
@@ -420,7 +422,7 @@ def test_global_send_test_notification(client, live_server, measure_memory_usage
     with open("test-datastore/notification.txt", 'r') as f:
         x = f.read()
         # Should come from notification.py default handler when there is no notification body to pull from
-        assert 'change detection is coo' in x
+        assert 'change detection is cool 网站监测 内容更新了' in x
 
     client.get(
         url_for("form_delete", uuid="all"),

changedetectionio/update_worker.py

@@ -1,6 +1,5 @@
 from .processors.exceptions import ProcessorException
-from . import content_fetchers
-
+import changedetectionio.content_fetchers.exceptions as content_fetchers_exceptions
 from changedetectionio.processors.text_json_diff.processor import FilterNotFoundInResponse
 from changedetectionio import html_tools
 
@@ -301,7 +300,7 @@ class update_worker(threading.Thread):
                         self.datastore.update_watch(uuid=uuid, update_obj={'last_error': e.message})
                         process_changedetection_results = False
 
-                    except content_fetchers.exceptions.ReplyWithContentButNoText as e:
+                    except content_fetchers_exceptions.ReplyWithContentButNoText as e:
                         # Totally fine, it's by choice - just continue on, nothing more to care about
                         # Page had elements/content but no renderable text
                         # Backend (not filters) gave zero output
@@ -327,7 +326,7 @@ class update_worker(threading.Thread):
                             
                         process_changedetection_results = False
 
-                    except content_fetchers.exceptions.Non200ErrorCodeReceived as e:
+                    except content_fetchers_exceptions.Non200ErrorCodeReceived as e:
                         if e.status_code == 403:
                             err_text = "Error - 403 (Access denied) received"
                         elif e.status_code == 404:
@@ -380,23 +379,23 @@ class update_worker(threading.Thread):
 
                         process_changedetection_results = False
 
-                    except content_fetchers.exceptions.checksumFromPreviousCheckWasTheSame as e:
+                    except content_fetchers_exceptions.checksumFromPreviousCheckWasTheSame as e:
                         # Yes fine, so nothing todo, don't continue to process.
                         process_changedetection_results = False
                         changed_detected = False
-                    except content_fetchers.exceptions.BrowserConnectError as e:
+                    except content_fetchers_exceptions.BrowserConnectError as e:
                         self.datastore.update_watch(uuid=uuid,
                                                     update_obj={'last_error': e.msg
                                                                 }
                                                     )
                         process_changedetection_results = False
-                    except content_fetchers.exceptions.BrowserFetchTimedOut as e:
+                    except content_fetchers_exceptions.BrowserFetchTimedOut as e:
                         self.datastore.update_watch(uuid=uuid,
                                                     update_obj={'last_error': e.msg
                                                                 }
                                                     )
                         process_changedetection_results = False
-                    except content_fetchers.exceptions.BrowserStepsStepException as e:
+                    except content_fetchers_exceptions.BrowserStepsStepException as e:
 
                         if not self.datastore.data['watching'].get(uuid):
                             continue
@@ -438,25 +437,25 @@ class update_worker(threading.Thread):
 
                         process_changedetection_results = False
 
-                    except content_fetchers.exceptions.EmptyReply as e:
+                    except content_fetchers_exceptions.EmptyReply as e:
                         # Some kind of custom to-str handler in the exception handler that does this?
                         err_text = "EmptyReply - try increasing 'Wait seconds before extracting text', Status Code {}".format(e.status_code)
                         self.datastore.update_watch(uuid=uuid, update_obj={'last_error': err_text,
                                                                            'last_check_status': e.status_code})
                         process_changedetection_results = False
-                    except content_fetchers.exceptions.ScreenshotUnavailable as e:
+                    except content_fetchers_exceptions.ScreenshotUnavailable as e:
                         err_text = "Screenshot unavailable, page did not render fully in the expected time or page was too long - try increasing 'Wait seconds before extracting text'"
                         self.datastore.update_watch(uuid=uuid, update_obj={'last_error': err_text,
                                                                            'last_check_status': e.status_code})
                         process_changedetection_results = False
-                    except content_fetchers.exceptions.JSActionExceptions as e:
+                    except content_fetchers_exceptions.JSActionExceptions as e:
                         err_text = "Error running JS Actions - Page request - "+e.message
                         if e.screenshot:
                             watch.save_screenshot(screenshot=e.screenshot, as_error=True)
                         self.datastore.update_watch(uuid=uuid, update_obj={'last_error': err_text,
                                                                            'last_check_status': e.status_code})
                         process_changedetection_results = False
-                    except content_fetchers.exceptions.PageUnloadable as e:
+                    except content_fetchers_exceptions.PageUnloadable as e:
                         err_text = "Page request from server didnt respond correctly"
                         if e.message:
                             err_text = "{} - {}".format(err_text, e.message)
@@ -468,7 +467,7 @@ class update_worker(threading.Thread):
                                                                            'last_check_status': e.status_code,
                                                                            'has_ldjson_price_data': None})
                         process_changedetection_results = False
-                    except content_fetchers.exceptions.BrowserStepsInUnsupportedFetcher as e:
+                    except content_fetchers_exceptions.BrowserStepsInUnsupportedFetcher as e:
                         err_text = "This watch has Browser Steps configured and so it cannot run with the 'Basic fast Plaintext/HTTP Client', either remove the Browser Steps or select a Chrome fetcher."
                         self.datastore.update_watch(uuid=uuid, update_obj={'last_error': err_text})
                         process_changedetection_results = False

docker-compose.yml

@@ -18,7 +18,7 @@ services:
   #
   #        Log levels are in descending order. (TRACE is the most detailed one)
   #        Log output levels: TRACE, DEBUG(default), INFO, SUCCESS, WARNING, ERROR, CRITICAL
-  #      - LOGGER_LEVEL=DEBUG
+  #      - LOGGER_LEVEL=TRACE
   #
   #       Alternative WebDriver/selenium URL, do not use "'s or 's!
   #      - WEBDRIVER_URL=http://browser-chrome:4444/wd/hub
@@ -29,8 +29,9 @@ services:
   #
   #             https://selenium-python.readthedocs.io/api.html#module-selenium.webdriver.common.proxy
   #
-  #       Alternative Playwright URL, do not use "'s or 's!
-  #      - PLAYWRIGHT_DRIVER_URL=ws://playwright-chrome:3000
+  #       Alternative target "Chrome" Playwright URL, do not use "'s or 's!
+  #       "Playwright" is a driver/librarythat allows changedetection to talk to a Chrome or similar browser.
+  #      - PLAYWRIGHT_DRIVER_URL=ws://sockpuppetbrowser:3000
   #
   #       Playwright proxy settings playwright_proxy_server, playwright_proxy_bypass, playwright_proxy_username, playwright_proxy_password
   #
@@ -57,6 +58,10 @@ services:
   #
   #        Absolute minimum seconds to recheck, overrides any watch minimum, change to 0 to disable
   #      - MINIMUM_SECONDS_RECHECK_TIME=3
+  #
+  #        Scrape prices from web pages automatically where the page has no embedded price information (see below also)
+  #      - PRICE_SCRAPER_ML_ENDPOINT=http://cdio-price-scraper:5005
+
       # Comment out ports: when using behind a reverse proxy , enable networks: etc.
       ports:
         - 5000:5000
@@ -73,10 +78,10 @@ services:
 #              condition: service_started
 
 
-     # Used for fetching pages via Playwright+Chrome where you need Javascript support.
+     # Sockpuppetbrowser is basically chrome wrapped in an API for allowing fast fetching of web-pages.
      # RECOMMENDED FOR FETCHING PAGES WITH CHROME
-#    playwright-chrome:
-#        hostname: playwright-chrome
+#    sockpuppetbrowser:
+#        hostname: sockpuppetbrowser
 #        image: dgtlmoon/sockpuppetbrowser:latest
 #        cap_add:
 #            - SYS_ADMIN
@@ -103,6 +108,13 @@ services:
 #            # Workaround to avoid the browser crashing inside a docker container
 #            # See https://github.com/SeleniumHQ/docker-selenium#quick-start
 #            - /dev/shm:/dev/shm
+#        restart: unless-stopped
+
+     # Machine Learning/AI - Use "Visual Selector" elements data to scrape price data
+
+#    cdio-keras-price-scraper:
+#        hostname: cdio-price-scraper
+#        image: dgtlmoon/changedetection-AI-pricescraper
 #        restart: unless-stopped
 
 volumes:

requirements.txt

@@ -35,7 +35,7 @@ dnspython==2.6.1 # related to eventlet fixes
 # jq not available on Windows so must be installed manually
 
 # Notification library
-apprise~=1.8.0
+apprise~=1.8.1
 
 # apprise mqtt https://github.com/dgtlmoon/changedetection.io/issues/315
 # and 2.0.0 https://github.com/dgtlmoon/changedetection.io/issues/2241 not yet compatible
@@ -79,9 +79,10 @@ pyppeteerstealth>=0.0.4
 pytest ~=7.2
 pytest-flask ~=1.2
 
-# Pin jsonschema version to prevent build errors on armv6 while rpds-py wheels aren't available (1708)
-jsonschema==4.17.3
+# Anything 4.0 and up but not 5.0
+jsonschema ~= 4.0
 
+price_parser
 loguru
 
 # For scraping all possible metadata relating to products so we can do better restock detection