Plugins for conditions (and include Similarity / Levenshtein, wordcount conditions) Re #3108
Some checks failed
Build and push containers / metadata (push) Has been cancelled
Build and push containers / build-push-containers (push) Has been cancelled
Publish Python 🐍distribution 📦 to PyPI and TestPyPI / Build distribution 📦 (push) Has been cancelled
ChangeDetection.io App Test / lint-code (push) Has been cancelled
Publish Python 🐍distribution 📦 to PyPI and TestPyPI / Test the built 📦 package works basically. (push) Has been cancelled
Publish Python 🐍distribution 📦 to PyPI and TestPyPI / Publish Python 🐍 distribution 📦 to PyPI (push) Has been cancelled
ChangeDetection.io App Test / test-application-3-10 (push) Has been cancelled
ChangeDetection.io App Test / test-application-3-11 (push) Has been cancelled
ChangeDetection.io App Test / test-application-3-12 (push) Has been cancelled
ChangeDetection.io App Test / test-application-3-13 (push) Has been cancelled
CodeQL / Analyze (javascript) (push) Has been cancelled
CodeQL / Analyze (python) (push) Has been cancelled

This commit is contained in:
dgtlmoon
2025-04-22 18:19:56 +02:00
committed by GitHub
parent a75f57de43
commit af568d064c
10 changed files with 489 additions and 8 deletions

View File

@@ -0,0 +1,98 @@
# Creating Plugins for changedetection.io
This document describes how to create plugins for changedetection.io. Plugins can be used to extend the functionality of the application in various ways.
## Plugin Types
### UI Stats Tab Plugins
These plugins can add content to the Stats tab in the Edit page. This is useful for adding custom statistics or visualizations about a watch.
#### Creating a UI Stats Tab Plugin
1. Create a Python file in a directory that will be loaded by the plugin system.
2. Use the `global_hookimpl` decorator to implement the `ui_edit_stats_extras` hook:
```python
import pluggy
from loguru import logger
global_hookimpl = pluggy.HookimplMarker("changedetectionio")
@global_hookimpl
def ui_edit_stats_extras(watch):
"""Add custom content to the stats tab"""
# Calculate or retrieve your stats
my_stat = calculate_something(watch)
# Return HTML content as a string
html = f"""
<div class="my-plugin-stats">
<h4>My Plugin Statistics</h4>
<p>My statistic: {my_stat}</p>
</div>
"""
return html
```
3. The HTML you return will be included in the Stats tab.
## Plugin Loading
Plugins can be loaded from:
1. Built-in plugin directories in the codebase
2. External packages using setuptools entry points
To add a new plugin directory, modify the `plugin_dirs` dictionary in `pluggy_interface.py`.
## Example Plugin
Here's a simple example of a plugin that adds a word count statistic to the Stats tab:
```python
import pluggy
from loguru import logger
global_hookimpl = pluggy.HookimplMarker("changedetectionio")
def count_words_in_history(watch):
"""Count words in the latest snapshot"""
try:
if not watch.history.keys():
return 0
latest_key = list(watch.history.keys())[-1]
latest_content = watch.get_history_snapshot(latest_key)
return len(latest_content.split())
except Exception as e:
logger.error(f"Error counting words: {str(e)}")
return 0
@global_hookimpl
def ui_edit_stats_extras(watch):
"""Add word count to the Stats tab"""
word_count = count_words_in_history(watch)
html = f"""
<div class="word-count-stats">
<h4>Content Analysis</h4>
<table class="pure-table">
<tbody>
<tr>
<td>Word count (latest snapshot)</td>
<td>{word_count}</td>
</tr>
</tbody>
</table>
</div>
"""
return html
```
## Testing Your Plugin
1. Place your plugin in one of the directories scanned by the plugin system
2. Restart changedetection.io
3. Go to the Edit page of a watch and check the Stats tab to see your content

View File

@@ -233,6 +233,9 @@ def construct_blueprint(datastore: ChangeDetectionStore, update_q, queuedWatchMe
# Only works reliably with Playwright
# Import the global plugin system
from changedetectionio.pluggy_interface import collect_ui_edit_stats_extras
template_args = {
'available_processors': processors.available_processors(),
'available_timezones': sorted(available_timezones()),
@@ -250,6 +253,7 @@ def construct_blueprint(datastore: ChangeDetectionStore, update_q, queuedWatchMe
'settings_application': datastore.data['settings']['application'],
'system_has_playwright_configured': os.getenv('PLAYWRIGHT_DRIVER_URL'),
'system_has_webdriver_configured': os.getenv('WEBDRIVER_URL'),
'ui_edit_stats_extras': collect_ui_edit_stats_extras(watch),
'visual_selector_data_ready': datastore.visualselector_data_is_ready(watch_uuid=uuid),
'timezone_default_config': datastore.data['settings']['application'].get('timezone'),
'using_global_webdriver_wait': not default['webdriver_delay'],

View File

@@ -102,12 +102,31 @@ def execute_ruleset_against_all_plugins(current_watch_uuid: str, application_dat
if complete_rules:
# Give all plugins a chance to update the data dict again (that we will test the conditions against)
for plugin in plugin_manager.get_plugins():
new_execute_data = plugin.add_data(current_watch_uuid=current_watch_uuid,
application_datastruct=application_datastruct,
ephemeral_data=ephemeral_data)
if new_execute_data and isinstance(new_execute_data, dict):
EXECUTE_DATA.update(new_execute_data)
try:
import concurrent.futures
import time
with concurrent.futures.ThreadPoolExecutor() as executor:
future = executor.submit(
plugin.add_data,
current_watch_uuid=current_watch_uuid,
application_datastruct=application_datastruct,
ephemeral_data=ephemeral_data
)
# Set a timeout of 10 seconds
try:
new_execute_data = future.result(timeout=10)
if new_execute_data and isinstance(new_execute_data, dict):
EXECUTE_DATA.update(new_execute_data)
except concurrent.futures.TimeoutError:
# The plugin took too long, abort processing for this watch
raise Exception(f"Plugin {plugin.__class__.__name__} took more than 10 seconds to run.")
except Exception as e:
# Log the error but continue with the next plugin
import logging
logging.error(f"Error executing plugin {plugin.__class__.__name__}: {str(e)}")
continue
# Create the ruleset
ruleset = convert_to_jsonlogic(logic_operator=logic_operator, rule_dict=complete_rules)
@@ -132,3 +151,18 @@ for plugin in plugin_manager.get_plugins():
if isinstance(new_field_choices, list):
field_choices.extend(new_field_choices)
def collect_ui_edit_stats_extras(watch):
"""Collect and combine HTML content from all plugins that implement ui_edit_stats_extras"""
extras_content = []
for plugin in plugin_manager.get_plugins():
try:
content = plugin.ui_edit_stats_extras(watch=watch)
if content:
extras_content.append(content)
except Exception as e:
# Skip plugins that don't implement the hook or have errors
pass
return "\n".join(extras_content) if extras_content else ""

View File

@@ -1,5 +1,8 @@
import pluggy
from . import default_plugin # Import the default plugin
import os
import importlib
import sys
from . import default_plugin
# ✅ Ensure that the namespace in HookspecMarker matches PluginManager
PLUGIN_NAMESPACE = "changedetectionio_conditions"
@@ -30,6 +33,11 @@ class ConditionsSpec:
def add_data(current_watch_uuid, application_datastruct, ephemeral_data):
"""Add to the datadict"""
pass
@hookspec
def ui_edit_stats_extras(watch):
"""Return HTML content to add to the stats tab in the edit view"""
pass
# ✅ Set up Pluggy Plugin Manager
plugin_manager = pluggy.PluginManager(PLUGIN_NAMESPACE)
@@ -40,5 +48,27 @@ plugin_manager.add_hookspecs(ConditionsSpec)
# ✅ Register built-in plugins manually
plugin_manager.register(default_plugin, "default_plugin")
# ✅ Load plugins from the plugins directory
def load_plugins_from_directory():
plugins_dir = os.path.join(os.path.dirname(__file__), 'plugins')
if not os.path.exists(plugins_dir):
return
# Get all Python files (excluding __init__.py)
for filename in os.listdir(plugins_dir):
if filename.endswith(".py") and filename != "__init__.py":
module_name = filename[:-3] # Remove .py extension
module_path = f"changedetectionio.conditions.plugins.{module_name}"
try:
module = importlib.import_module(module_path)
# Register the plugin with pluggy
plugin_manager.register(module, module_name)
except (ImportError, AttributeError) as e:
print(f"Error loading plugin {module_name}: {e}")
# Load plugins from the plugins directory
load_plugins_from_directory()
# ✅ Discover installed plugins from external packages (if any)
plugin_manager.load_setuptools_entrypoints(PLUGIN_NAMESPACE)

View File

@@ -0,0 +1 @@
# Import plugins package to make them discoverable

View File

@@ -0,0 +1,102 @@
import pluggy
from loguru import logger
# Support both plugin systems
conditions_hookimpl = pluggy.HookimplMarker("changedetectionio_conditions")
global_hookimpl = pluggy.HookimplMarker("changedetectionio")
def levenshtein_ratio_recent_history(watch, incoming_text=None):
try:
from Levenshtein import ratio, distance
k = list(watch.history.keys())
if len(k) >= 2:
# When called from ui_edit_stats_extras, we don't have incoming_text
if incoming_text is None:
a = watch.get_history_snapshot(timestamp=k[-1]) # Latest snapshot
b = watch.get_history_snapshot(timestamp=k[-2]) # Previous snapshot
else:
a = watch.get_history_snapshot(timestamp=k[-2]) # Second newest, incoming_text will be "newest"
b = incoming_text
distance_value = distance(a, b)
ratio_value = ratio(a, b)
return {
'distance': distance_value,
'ratio': ratio_value,
'percent_similar': round(ratio_value * 100, 2)
}
except Exception as e:
logger.warning(f"Unable to calc similarity: {str(e)}")
return ''
@conditions_hookimpl
def register_operators():
pass
@conditions_hookimpl
def register_operator_choices():
pass
@conditions_hookimpl
def register_field_choices():
return [
("levenshtein_ratio", "Levenshtein - Text similarity ratio"),
("levenshtein_distance", "Levenshtein - Text change distance"),
]
@conditions_hookimpl
def add_data(current_watch_uuid, application_datastruct, ephemeral_data):
res = {}
watch = application_datastruct['watching'].get(current_watch_uuid)
# ephemeral_data['text'] will be the current text after filters, they may have edited filters but not saved them yet etc
if watch and 'text' in ephemeral_data:
lev_data = levenshtein_ratio_recent_history(watch, ephemeral_data['text'])
if isinstance(lev_data, dict):
res['levenshtein_ratio'] = lev_data.get('ratio', 0)
res['levenshtein_similarity'] = lev_data.get('percent_similar', 0)
res['levenshtein_distance'] = lev_data.get('distance', 0)
return res
@global_hookimpl
def ui_edit_stats_extras(watch):
"""Add Levenshtein stats to the UI using the global plugin system"""
"""Generate the HTML for Levenshtein stats - shared by both plugin systems"""
if len(watch.history.keys()) < 2:
return "<p>Not enough history to calculate Levenshtein metrics</p>"
try:
lev_data = levenshtein_ratio_recent_history(watch)
if not lev_data or not isinstance(lev_data, dict):
return "<p>Unable to calculate Levenshtein metrics</p>"
html = f"""
<div class="levenshtein-stats">
<h4>Levenshtein Text Similarity Details</h4>
<table class="pure-table">
<tbody>
<tr>
<td>Raw distance (edits needed)</td>
<td>{lev_data['distance']}</td>
</tr>
<tr>
<td>Similarity ratio</td>
<td>{lev_data['ratio']:.4f}</td>
</tr>
<tr>
<td>Percent similar</td>
<td>{lev_data['percent_similar']}%</td>
</tr>
</tbody>
</table>
<p style="font-size: 80%;">Levenshtein metrics compare the last two snapshots, measuring how many character edits are needed to transform one into the other.</p>
</div>
"""
return html
except Exception as e:
logger.error(f"Error generating Levenshtein UI extras: {str(e)}")
return "<p>Error calculating Levenshtein metrics</p>"

View File

@@ -0,0 +1,82 @@
import pluggy
from loguru import logger
# Support both plugin systems
conditions_hookimpl = pluggy.HookimplMarker("changedetectionio_conditions")
global_hookimpl = pluggy.HookimplMarker("changedetectionio")
def count_words_in_history(watch, incoming_text=None):
"""Count words in snapshot text"""
try:
if incoming_text is not None:
# When called from add_data with incoming text
return len(incoming_text.split())
elif watch.history.keys():
# When called from UI extras to count latest snapshot
latest_key = list(watch.history.keys())[-1]
latest_content = watch.get_history_snapshot(latest_key)
return len(latest_content.split())
return 0
except Exception as e:
logger.error(f"Error counting words: {str(e)}")
return 0
# Implement condition plugin hooks
@conditions_hookimpl
def register_operators():
# No custom operators needed
return {}
@conditions_hookimpl
def register_operator_choices():
# No custom operator choices needed
return []
@conditions_hookimpl
def register_field_choices():
# Add a field that will be available in conditions
return [
("word_count", "Word count of content"),
]
@conditions_hookimpl
def add_data(current_watch_uuid, application_datastruct, ephemeral_data):
"""Add word count data for conditions"""
result = {}
watch = application_datastruct['watching'].get(current_watch_uuid)
if watch and 'text' in ephemeral_data:
word_count = count_words_in_history(watch, ephemeral_data['text'])
result['word_count'] = word_count
return result
def _generate_stats_html(watch):
"""Generate the HTML content for the stats tab"""
word_count = count_words_in_history(watch)
html = f"""
<div class="word-count-stats">
<h4>Content Analysis</h4>
<table class="pure-table">
<tbody>
<tr>
<td>Word count (latest snapshot)</td>
<td>{word_count}</td>
</tr>
</tbody>
</table>
<p style="font-size: 80%;">Word count is a simple measure of content length, calculated by splitting text on whitespace.</p>
</div>
"""
return html
@conditions_hookimpl
def ui_edit_stats_extras(watch):
"""Add word count stats to the UI through conditions plugin system"""
return _generate_stats_html(watch)
@global_hookimpl
def ui_edit_stats_extras(watch):
"""Add word count stats to the UI using the global plugin system"""
return _generate_stats_html(watch)

View File

@@ -0,0 +1,82 @@
import pluggy
import os
import importlib
import sys
# Global plugin namespace for changedetection.io
PLUGIN_NAMESPACE = "changedetectionio"
hookspec = pluggy.HookspecMarker(PLUGIN_NAMESPACE)
hookimpl = pluggy.HookimplMarker(PLUGIN_NAMESPACE)
class ChangeDetectionSpec:
"""Hook specifications for extending changedetection.io functionality."""
@hookspec
def ui_edit_stats_extras(watch):
"""Return HTML content to add to the stats tab in the edit view.
Args:
watch: The watch object being edited
Returns:
str: HTML content to be inserted in the stats tab
"""
pass
# Set up Plugin Manager
plugin_manager = pluggy.PluginManager(PLUGIN_NAMESPACE)
# Register hookspecs
plugin_manager.add_hookspecs(ChangeDetectionSpec)
# Load plugins from subdirectories
def load_plugins_from_directories():
# Dictionary of directories to scan for plugins
plugin_dirs = {
'conditions': os.path.join(os.path.dirname(__file__), 'conditions', 'plugins'),
# Add more plugin directories here as needed
}
# Note: Removed the direct import of example_word_count_plugin as it's now in the conditions/plugins directory
for dir_name, dir_path in plugin_dirs.items():
if not os.path.exists(dir_path):
continue
# Get all Python files (excluding __init__.py)
for filename in os.listdir(dir_path):
if filename.endswith(".py") and filename != "__init__.py":
module_name = filename[:-3] # Remove .py extension
module_path = f"changedetectionio.{dir_name}.plugins.{module_name}"
try:
module = importlib.import_module(module_path)
# Register the plugin with pluggy
plugin_manager.register(module, module_name)
except (ImportError, AttributeError) as e:
print(f"Error loading plugin {module_name}: {e}")
# Load plugins
load_plugins_from_directories()
# Discover installed plugins from external packages (if any)
plugin_manager.load_setuptools_entrypoints(PLUGIN_NAMESPACE)
# Helper function to collect UI stats extras from all plugins
def collect_ui_edit_stats_extras(watch):
"""Collect and combine HTML content from all plugins that implement ui_edit_stats_extras"""
extras_content = []
# Get all plugins that implement the ui_edit_stats_extras hook
results = plugin_manager.hook.ui_edit_stats_extras(watch=watch)
# If we have results, add them to our content
if results:
for result in results:
if result: # Skip empty results
extras_content.append(result)
return "\n".join(extras_content) if extras_content else ""

View File

@@ -450,6 +450,13 @@ Math: {{ 1 + 1 }}") }}
</tr>
</tbody>
</table>
{% if ui_edit_stats_extras %}
<div class="plugin-stats-extras"> <!-- from pluggy plugin -->
{{ ui_edit_stats_extras|safe }}
</div>
{% endif %}
{% if watch.history_n %}
<p>
<a href="{{url_for('ui.ui_edit.watch_get_latest_html', uuid=uuid)}}" class="pure-button button-small">Download latest HTML snapshot</a>

View File

@@ -45,11 +45,15 @@ def set_number_out_of_range_response(number="150"):
f.write(test_return_data)
def test_setup(client, live_server):
"""Test that both text and number conditions work together with AND logic."""
live_server_setup(live_server)
def test_conditions_with_text_and_number(client, live_server):
"""Test that both text and number conditions work together with AND logic."""
set_original_response("50")
live_server_setup(live_server)
#live_server_setup(live_server)
test_url = url_for('test_endpoint', _external=True)
@@ -195,3 +199,40 @@ def test_condition_validate_rule_row(client, live_server):
# If there was only a change in the whitespacing, then we shouldnt have a change detected
def test_wordcount_conditions_plugin(client, live_server, measure_memory_usage):
#live_server_setup(live_server)
test_return_data = """<html>
<body>
Some initial text<br>
<p>Which is across multiple lines</p>
<br>
So let's see what happens. <br>
</body>
</html>
"""
with open("test-datastore/endpoint-content.txt", "w") as f:
f.write(test_return_data)
# Add our URL to the import page
test_url = url_for('test_endpoint', _external=True)
res = client.post(
url_for("imports.import_page"),
data={"urls": test_url},
follow_redirects=True
)
assert b"1 Imported" in res.data
# Give the thread time to pick it up
wait_for_all_checks(client)
# Check it saved
res = client.get(
url_for("ui.ui_edit.edit_page", uuid="first"),
)
# Assert the word count is counted correctly
assert b'<td>13</td>' in res.data