Merge branch 'master' into add-system-info-api

.github/workflows/test-container-build.yml

@@ -1,12 +1,21 @@
 name: ChangeDetection.io Container Build Test
 
 # Triggers the workflow on push or pull request events
+
+# This line doesnt work, even tho it is the documented one
+#on: [push, pull_request]
+
 on:
   push:
     paths:
       - requirements.txt
       - Dockerfile
 
+  pull_request:
+    paths:
+      - requirements.txt
+      - Dockerfile
+
   # Changes to requirements.txt packages and Dockerfile may or may not always be compatible with arm etc, so worth testing
   # @todo: some kind of path filter for requirements.txt and Dockerfile
 jobs:

CONTRIBUTING.md

@@ -6,7 +6,7 @@ Otherwise, it's always best to PR into the `dev` branch.
 
 Please be sure that all new functionality has a matching test!
 
-Use `pytest` to validate/test, you can run the existing tests as `pytest tests/test_notifications.py` for example
+Use `pytest` to validate/test, you can run the existing tests as `pytest tests/test_notification.py` for example
 
 ```
 pip3 install -r requirements-dev

Dockerfile

@@ -64,6 +64,7 @@ EXPOSE 5000
 
 # The actual flask app
 COPY changedetectionio /app/changedetectionio
+
 # The eventlet server wrapper
 COPY changedetection.py /app/changedetection.py
 

MANIFEST.in

@@ -2,6 +2,7 @@ recursive-include changedetectionio/api *
 recursive-include changedetectionio/templates *
 recursive-include changedetectionio/static *
 recursive-include changedetectionio/model *
+recursive-include changedetectionio/tests *
 include changedetection.py
 global-exclude *.pyc
 global-exclude node_modules

README.md

@@ -161,50 +161,14 @@ This will re-parse the JSON and apply formatting to the text, making it super ea
 
 ### JSONPath or jq?
 
-For more complex parsing, filtering, and modifying of JSON data, jq is recommended due to the built-in operators and functions. Refer to the [documentation](https://stedolan.github.io/jq/manual/) for more information on jq.
+For more complex parsing, filtering, and modifying of JSON data, jq is recommended due to the built-in operators and functions. Refer to the [documentation](https://stedolan.github.io/jq/manual/) for more specifc information on jq.
 
-Notes:
-- `jq` must be added manually separately from the installation of changedetection.io (simply run `pip3 install jq`)
-- `jq` is not available on Windows or must be manually compiled (No "wheel" package available on pypi)
+One big advantage of `jq` is that you can use logic in your JSON filter, such as filters to only show items that have a value greater than/less than etc.
 
-- The example below adds the price in dollars to each item in the JSON data, and then filters to only show items that are greater than 10.
+See the wiki https://github.com/dgtlmoon/changedetection.io/wiki/JSON-Selector-Filter-help for more information and examples
 
-#### Sample input data from API
-```
-{
-    "items": [
-        {
-           "name": "Product A",
-           "priceInCents": 2500
-        },
-        {
-           "name": "Product B",
-           "priceInCents": 500
-        },
-        {
-           "name": "Product C",
-           "priceInCents": 2000
-        }
-    ]
-}
-```
+Note: `jq` library must be added separately (`pip3 install jq`)
 
-#### Sample jq
-`jq:.items[] | . + { "priceInDollars": (.priceInCents / 100) } | select(.priceInDollars > 10)`
-
-#### Sample output data
-```
-{
-  "name": "Product A",
-  "priceInCents": 2500,
-  "priceInDollars": 25
-}
-{
-  "name": "Product C",
-  "priceInCents": 2000,
-  "priceInDollars": 20
-}
-```
 
 ### Parse JSON embedded in HTML!
 

changedetectionio/__init__.py

@@ -33,7 +33,7 @@ from flask_wtf import CSRFProtect
 from changedetectionio import html_tools
 from changedetectionio.api import api_v1
 
-__version__ = '0.39.20.1'
+__version__ = '0.39.20.4'
 
 datastore = None
 
@@ -194,6 +194,9 @@ def changedetection_app(config=None, datastore_o=None):
     watch_api.add_resource(api_v1.Watch, '/api/v1/watch/<string:uuid>',
                            resource_class_kwargs={'datastore': datastore, 'update_q': update_q})
 
+    watch_api.add_resource(api_v1.SystemInfo, '/api/v1/systeminfo',
+                           resource_class_kwargs={'datastore': datastore, 'update_q': update_q})
+
 
 
 
@@ -816,8 +819,10 @@ def changedetection_app(config=None, datastore_o=None):
 
         newest_file = history[dates[-1]]
 
+        # Read as binary and force decode as UTF-8
+        # Windows may fail decode in python if we just use 'r' mode (chardet decode exception)
         try:
-            with open(newest_file, 'r') as f:
+            with open(newest_file, 'r', encoding='utf-8', errors='ignore') as f:
                 newest_version_file_contents = f.read()
         except Exception as e:
             newest_version_file_contents = "Unable to read {}.\n".format(newest_file)
@@ -830,7 +835,7 @@ def changedetection_app(config=None, datastore_o=None):
             previous_file = history[dates[-2]]
 
         try:
-            with open(previous_file, 'r') as f:
+            with open(previous_file, 'r', encoding='utf-8', errors='ignore') as f:
                 previous_version_file_contents = f.read()
         except Exception as e:
             previous_version_file_contents = "Unable to read {}.\n".format(previous_file)
@@ -907,7 +912,7 @@ def changedetection_app(config=None, datastore_o=None):
         timestamp = list(watch.history.keys())[-1]
         filename = watch.history[timestamp]
         try:
-            with open(filename, 'r') as f:
+            with open(filename, 'r', encoding='utf-8', errors='ignore') as f:
                 tmp = f.readlines()
 
                 # Get what needs to be highlighted

changedetectionio/api/api_v1.py

@@ -122,3 +122,33 @@ class CreateWatch(Resource):
             return {'status': "OK"}, 200
 
         return list, 200
+
+class SystemInfo(Resource):
+    def __init__(self, **kwargs):
+        # datastore is a black box dependency
+        self.datastore = kwargs['datastore']
+        self.update_q = kwargs['update_q']
+
+    @auth.check_token
+    def get(self):
+        import time
+        overdue_watches = []
+
+        # Check all watches and report which have not been checked but should have been
+
+        for uuid, watch in self.datastore.data.get('watching', {}).items():
+            # see if now - last_checked is greater than the time that should have been
+            # this is not super accurate (maybe they just edited it) but better than nothing
+            t = watch.threshold_seconds()
+            if not t:
+                t = self.datastore.threshold_seconds
+            time_since_check = time.time() - watch.get('last_checked')
+            if time_since_check > t:
+                overdue_watches.append(uuid)
+
+        return {
+                   'queue_size': self.update_q.qsize(),
+                   'overdue_watches': overdue_watches,
+                   'uptime': round(time.time() - self.datastore.start_time, 2),
+                   'watch_count': len(self.datastore.data.get('watching', {}))
+               }, 200

changedetectionio/changedetection.py

@@ -102,6 +102,14 @@ def main():
                     has_password=datastore.data['settings']['application']['password'] != False
                     )
 
+    # Monitored websites will not receive a Referer header
+    # when a user clicks on an outgoing link.
+    @app.after_request
+    def hide_referrer(response):
+        if os.getenv("HIDE_REFERER", False):
+            response.headers["Referrer-Policy"] = "no-referrer"
+        return response
+
     # Proxy sub-directory support
     # Set environment var USE_X_SETTINGS=1 on this script
     # And then in your proxy_pass settings

changedetectionio/model/Watch.py

@@ -118,7 +118,10 @@ class model(dict):
         if os.path.isfile(fname):
             logging.debug("Reading history index " + str(time.time()))
             with open(fname, "r") as f:
-                tmp_history = dict(i.strip().split(',', 2) for i in f.readlines())
+                for i in f.readlines():
+                    if ',' in i:
+                        k, v = i.strip().split(',', 2)
+                        tmp_history[k] = v
 
         if len(tmp_history):
             self.__newest_history_key = list(tmp_history.keys())[-1]
@@ -151,28 +154,30 @@ class model(dict):
         import uuid
         import logging
 
-        output_path = "{}/{}".format(self.__datastore_path, self['uuid'])
+        output_path = os.path.join(self.__datastore_path, self['uuid'])
 
         self.ensure_data_dir_exists()
+        snapshot_fname = os.path.join(output_path, str(uuid.uuid4()))
 
-        snapshot_fname = "{}/{}.stripped.txt".format(output_path, uuid.uuid4())
         logging.debug("Saving history text {}".format(snapshot_fname))
 
+        # in /diff/ and /preview/ we are going to assume for now that it's UTF-8 when reading
+        # most sites are utf-8 and some are even broken utf-8
         with open(snapshot_fname, 'wb') as f:
             f.write(contents)
             f.close()
 
         # Append to index
         # @todo check last char was \n
-        index_fname = "{}/history.txt".format(output_path)
+        index_fname = os.path.join(output_path, "history.txt")
         with open(index_fname, 'a') as f:
             f.write("{},{}\n".format(timestamp, snapshot_fname))
             f.close()
 
         self.__newest_history_key = timestamp
-        self.__history_n+=1
+        self.__history_n += 1
 
-        #@todo bump static cache of the last timestamp so we dont need to examine the file to set a proper ''viewed'' status
+        # @todo bump static cache of the last timestamp so we dont need to examine the file to set a proper ''viewed'' status
         return snapshot_fname
 
     @property

changedetectionio/run_all_tests.sh

@@ -9,6 +9,8 @@
 # exit when any command fails
 set -e
 
+SCRIPT_DIR=$( cd -- "$( dirname -- "${BASH_SOURCE[0]}" )" &> /dev/null && pwd )
+
 find tests/test_*py -type f|while read test_name
 do
   echo "TEST RUNNING $test_name"
@@ -45,7 +47,9 @@ docker kill $$-test_selenium
 
 echo "TESTING WEBDRIVER FETCH > PLAYWRIGHT/BROWSERLESS..."
 # Not all platforms support playwright (not ARM/rPI), so it's not packaged in requirements.txt
-pip3 install playwright~=1.24
+PLAYWRIGHT_VERSION=$(grep -i -E "RUN pip install.+" "$SCRIPT_DIR/../Dockerfile" | grep --only-matching -i -E "playwright[=><~+]+[0-9\.]+")
+echo "using $PLAYWRIGHT_VERSION"
+pip3 install "$PLAYWRIGHT_VERSION"
 docker run -d --name $$-test_browserless -e "DEFAULT_LAUNCH_ARGS=[\"--window-size=1920,1080\"]" --rm  -p 3000:3000  --shm-size="2g"  browserless/chrome:1.53-chrome-stable
 # takes a while to spin up
 sleep 5

changedetectionio/store.py

@@ -30,14 +30,14 @@ class ChangeDetectionStore:
     def __init__(self, datastore_path="/datastore", include_default_watches=True, version_tag="0.0.0"):
         # Should only be active for docker
         # logging.basicConfig(filename='/dev/stdout', level=logging.INFO)
-        self.needs_write = False
+        self.__data = App.model()
         self.datastore_path = datastore_path
         self.json_store_path = "{}/url-watches.json".format(self.datastore_path)
+        self.needs_write = False
         self.proxy_list = None
+        self.start_time = time.time()
         self.stop_thread = False
 
-        self.__data = App.model()
-
         # Base definition for all watchers
         # deepcopy part of #569 - not sure why its needed exactly
         self.generic_definition = deepcopy(Watch.model(datastore_path = datastore_path, default={}))

changedetectionio/tests/test_api.py

@@ -147,6 +147,16 @@ def test_api_simple(client, live_server):
     # @todo how to handle None/default global values?
     assert watch['history_n'] == 2, "Found replacement history section, which is in its own API"
 
+    # basic systeminfo check
+    res = client.get(
+        url_for("systeminfo"),
+        headers={'x-api-key': api_key},
+    )
+    info = json.loads(res.data)
+    assert info.get('watch_count') == 1
+    assert info.get('uptime') > 0.5
+
+
     # Finally delete the watch
     res = client.delete(
         url_for("watch", uuid=watch_uuid),

changedetectionio/tests/visualselector/test_fetch_data.py

@@ -13,9 +13,9 @@ def test_visual_selector_content_ready(client, live_server):
     live_server_setup(live_server)
     time.sleep(1)
 
-    # Add our URL to the import page, maybe better to use something we control?
-    # We use an external URL because the docker container is too difficult to setup to connect back to the pytest socket
-    test_url = 'https://news.ycombinator.com'
+    # Add our URL to the import page, because the docker container (playwright/selenium) wont be able to connect to our usual test url
+    test_url = "https://changedetection.io/ci-test/test-runjs.html"
+
     res = client.post(
         url_for("form_quick_watch_add"),
         data={"url": test_url, "tag": '', 'edit_and_watch_submit_button': 'Edit > Watch'},
@@ -25,13 +25,27 @@ def test_visual_selector_content_ready(client, live_server):
 
     res = client.post(
         url_for("edit_page", uuid="first", unpause_on_save=1),
-        data={"css_filter": ".does-not-exist", "url": test_url, "tag": "", "headers": "", 'fetch_backend': "html_webdriver"},
+        data={
+              "url": test_url,
+              "tag": "",
+              "headers": "",
+              'fetch_backend': "html_webdriver",
+              'webdriver_js_execute_code': 'document.querySelector("button[name=test-button]").click();'
+        },
         follow_redirects=True
     )
     assert b"unpaused" in res.data
     time.sleep(1)
     wait_for_all_checks(client)
     uuid = extract_UUID_from_client(client)
+
+    # Check the JS execute code before extract worked
+    res = client.get(
+        url_for("preview_page", uuid="first"),
+        follow_redirects=True
+    )
+    assert b'I smell JavaScript' in res.data
+
     assert os.path.isfile(os.path.join('test-datastore', uuid, 'last-screenshot.png')), "last-screenshot.png should exist"
     assert os.path.isfile(os.path.join('test-datastore', uuid, 'elements.json')), "xpath elements.json data should exist"
 

docker-compose.yml

@@ -45,6 +45,9 @@ services:
   #        Respect proxy_pass type settings, `proxy_set_header Host "localhost";` and `proxy_set_header X-Forwarded-Prefix /app;`
   #        More here https://github.com/dgtlmoon/changedetection.io/wiki/Running-changedetection.io-behind-a-reverse-proxy-sub-directory
   #      - USE_X_SETTINGS=1
+  #
+  #        Hides the `Referer` header so that monitored websites can't see the changedetection.io hostname.
+  #      - HIDE_REFERER=true
 
       # Comment out ports: when using behind a reverse proxy , enable networks: etc.
       ports:

requirements.txt

@@ -1,8 +1,8 @@
-flask~= 2.0
+flask ~= 2.0
 flask_wtf
-eventlet>=0.31.0
+eventlet >= 0.31.0
 validators
-timeago ~=1.0
+timeago ~= 1.0
 inscriptis ~= 2.2
 feedgen ~= 0.9
 flask-login ~= 0.5
@@ -47,3 +47,4 @@ selenium ~= 4.1.0
 werkzeug ~= 2.0.0
 
 # playwright is installed at Dockerfile build time because it's not available on all platforms
+