diff --git a/changedetectionio/processors/restock_diff.py b/changedetectionio/processors/restock_diff.py
index a40ad262a..09a09558a 100644
--- a/changedetectionio/processors/restock_diff.py
+++ b/changedetectionio/processors/restock_diff.py
@@ -1,5 +1,8 @@
from . import difference_detection_processor
-from ..html_tools import xpath1_filter as xpath_filter # xpath1 is a lot faster and is sufficient here
+from ..html_tools import xpath1_filter as xpath_filter
+# xpath1 is a lot faster and is sufficient here
+from ..html_tools import extract_json_as_string, has_ldjson_product_info
+
from copy import deepcopy
from loguru import logger
import hashlib
@@ -36,9 +39,14 @@ def get_itemprop_availability(html_content):
value = None
try:
- value = xpath_filter("//*[@itemtype='https://schema.org/Offer']//*[@itemprop='availability']/@href", html_content)
+ if has_ldjson_product_info(html_content):
+ value = extract_json_as_string(html_content.lower(), "json:$..offers.availability", ensure_is_ldjson_info_type=True)
+
+# value = xpath_filter("//*[@itemtype='https://schema.org/Offer']//*[@itemprop='availability']/@href", ldjson)
+
+
if value:
- value = re.sub(r'(?i)^http(s)+://schema.org/', '', value.strip())
+ value = re.sub(r'(?i)^(https|http)://schema.org/', '', value.strip(' "\''))
except Exception as e:
print("Exception getting get_itemprop_availability (itemprop='availability')", str(e))
@@ -82,13 +90,14 @@ class perform_site_check(difference_detection_processor):
if availability:
self.fetcher.instock_data = availability # Stored as the text snapshot
+ # @todo: Configurable?
if any(availability in s for s in
[
- 'InStock',
- 'InStoreOnly',
- 'LimitedAvailability',
- 'OnlineOnly',
- 'PreSale' # Debatable?
+ 'instock',
+ 'Instoreonly',
+ 'limitedavailability',
+ 'onlineonly',
+ 'presale' # Debatable?
]):
update_obj['in_stock'] = True
else:
@@ -99,13 +108,14 @@ class perform_site_check(difference_detection_processor):
# 'Possibly in stock' comes from stock-not-in-stock.js when no string found above the fold.
update_obj['in_stock'] = True if self.fetcher.instock_data == 'Possibly in stock' else False
logger.debug(f"Watch UUID {uuid} restock check returned '{self.fetcher.instock_data}' from JS scraper.")
- else:
+
+ if not self.fetcher.instock_data:
raise UnableToExtractRestockData(status_code=self.fetcher.status_code)
# Main detection method
fetched_md5 = None
- if self.fetcher.instock_data:
- fetched_md5 = hashlib.md5(self.fetcher.instock_data.encode('utf-8')).hexdigest()
+
+ fetched_md5 = hashlib.md5(self.fetcher.instock_data.encode('utf-8')).hexdigest()
# The main thing that all this at the moment comes down to :)
changed_detected = False
diff --git a/changedetectionio/tests/test_restock_itemprop.py b/changedetectionio/tests/test_restock_itemprop.py
new file mode 100644
index 000000000..6dd88ca7c
--- /dev/null
+++ b/changedetectionio/tests/test_restock_itemprop.py
@@ -0,0 +1,69 @@
+#!/usr/bin/python3
+from flask import url_for
+from .util import live_server_setup, wait_for_all_checks, extract_UUID_from_client
+
+
+
+instock_props = [
+ '',
+ '',
+]
+
+out_of_stock_props = [
+ # out of stock AND contains multiples
+ ''
+]
+
+def set_original_response(props_markup):
+ test_return_data = f"""
+
+ Some initial text
+ Which is across multiple lines
+
+ So let's see what happens.
+ price: $10.99
+ {props_markup}
+
+
+ """
+
+ with open("test-datastore/endpoint-content.txt", "w") as f:
+ f.write(test_return_data)
+ return None
+
+
+
+def test_restock_itemprop_basic(client, live_server):
+
+ live_server_setup(live_server)
+
+ test_url = url_for('test_endpoint', _external=True)
+
+ for p in instock_props:
+ set_original_response(props_markup=p)
+ client.post(
+ url_for("form_quick_watch_add"),
+ data={"url": test_url, "tags": '', 'processor': 'restock_diff'},
+ follow_redirects=True
+ )
+ wait_for_all_checks(client)
+ res = client.get(url_for("index"))
+ assert b' in-stock' in res.data
+
+ res = client.get(url_for("form_delete", uuid="all"), follow_redirects=True)
+ assert b'Deleted' in res.data
+
+
+ for p in out_of_stock_props:
+ set_original_response(props_markup=p)
+ client.post(
+ url_for("form_quick_watch_add"),
+ data={"url": test_url, "tags": '', 'processor': 'restock_diff'},
+ follow_redirects=True
+ )
+ wait_for_all_checks(client)
+ res = client.get(url_for("index"))
+ assert b'not-in-stock' in res.data
+
+ res = client.get(url_for("form_delete", uuid="all"), follow_redirects=True)
+ assert b'Deleted' in res.data
\ No newline at end of file