Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
38 changes: 38 additions & 0 deletions content/terms/explanation/filters.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,38 @@
---
title: "Filters"
weight: 3
---

# Filters

Filters enable solving [noise]({{< relref "/terms/guideline/declaring/#usual-noise" >}}) issues in versions that cannot be addressed with direct selection or removal of content using selectors.

## When filters are needed

Use filters when:

- **Content selectors are insufficient**, for example when noise appears within content that can't be targeted with CSS selectors or [range selectors]({{< relref "terms/explanation/range-selectors" >}}) with the [`select`]({{< relref "terms/reference/declaration/#ref-select" >}}) and [`remove`]({{< relref "terms/reference/declaration/#ref-remove" >}}) properties.
- **Content is dynamically generated**, for example when elements change on each page load with changing classes or IDs that cannot be targeted with [attribute selectors](https://developer.mozilla.org/en-US/docs/Web/CSS/Attribute_selectors).
- **Complex tasks are needed**, for example when content transformation is required such as converting images to base64 to store them in the terms version or converting date-based content to a stable format (like “Updated X days ago” to “Last updated on YYYY-MM-DD”).

## How filters work

Filters are JavaScript functions that can manipulate the DOM structure directly. They modify the document structure and content in-place.

## Filter design principles

Filters should follow these core principles:

- **Specific**: target only the noise to remove. Avoid broad selectors that might accidentally remove important content.

> For example, if a filter converts relative dates to absolute dates, make sure to scope the targeted dates. This might translate to selecting with `.metadata time`, not `time`, which might also affect important effective dates within the terms content.

- **Idempotent**: filters should produce the same result even if run multiple times on their own output. This ensures consistency.

> For example, if a filter adds section numbers like "1." to headings, it should check if the numbers already exist, to prevent "1. Privacy Policy" from becoming "1. 1. Privacy Policy" on repeated runs.

- **Efficient**: DOM queries should be optimised and filters should avoid unnecessary operations, processing only the elements needed.

> For example, if a filter updates timestamp elements with a specific class, using `document.querySelectorAll('.timestamp')` is more efficient than `document.querySelectorAll('*')` followed by filtering for timestamp elements.

- **Safe**: filters must not accidentally remove important content. The generated version should always be checked after adding a filter to ensure it still contains the whole terms content.
161 changes: 161 additions & 0 deletions content/terms/how-to/apply-filters.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,161 @@
---
title: Apply filters
weight: 7
---

# How to apply filters

This guide explains how to add filters to existing declarations to remove meaningless content that cannot be removed with CSS selectors, to prevent noise in the versions.

## Prerequisites

- An existing terms declaration file.
- Having already identified the noise to remove and having double-checked it cannot be removed with CSS selectors with the [`remove`]({{< relref "terms/reference/declaration/#ref-remove" >}}) property.

## Step 1: Check for built-in filters
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think I'll use a structure based on the principle that we most often use a built-in filter and optionally a custom filter, so I would put everything related to creating a custom filter on a dedicated how-to page.

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think we currently do not have enough builtin filters to justify splitting into two pages


Built-in filters are pre-defined functions that handle common noise patterns. They are the easiest way to clean up content.

Review the available [built-in filters]({{< relref "/terms/reference/built-in-filters" >}}) to find if one matches your needs.

If you find a suitable built-in filter, proceed to [Step 3](#step-3-declare-the-filter), otherwise you will need to create a custom filter.

## Step 2: Create a custom filter _(optional)_

If no built-in filter matches your needs, you will need to create a custom filter. This requires JavaScript knowledge and familiarity with DOM manipulation.

### Create the filter file

Create a JavaScript file in the same folder and with the same name as your service declaration, but with `.filters.js` extension.

> For example, if your declaration is `declarations/MyService.json`, create `declarations/MyService.filters.js`.

### Write the filter function

Define your filter function with the following signature:

```js
export function myCustomFilter(document, [parameters]) {
// Your filter logic here
}
```

#### Parameters

- `document`: JSDOM document instance representing the web page
- `parameters`: values passed from the declaration _(optional)_

#### Example: Remove session IDs from text content

For example, let's say you want to remove session IDs from text content:

```html
<p>We collect your data for the following purposes:</p>
<ul>
<li>To provide our services</li>
<li>To improve user experience</li>
</ul>
<p class="session-id">Last updated on 2023-12-07 (Session: abc123def456)</p>
```

You can implement this filter as follows:

```js
export function removeSessionIds(document) {
// Find all paragraphs that might contain session IDs
const paragraphs = document.querySelectorAll('.session-id');

paragraphs.forEach(paragraph => {
let text = paragraph.textContent;
// Remove session ID patterns like "Session: abc123" or "(Session: def456)"
text = text.replace(/\s*\(?Session:\s*[a-zA-Z0-9]+\)?/g, '');
paragraph.textContent = text.trim();
});
}
```

Result after applying the filter:

```diff
<p>We collect your data for the following purposes:</p>
<ul>
<li>To provide our services</li>
<li>To improve user experience</li>
</ul>
- <p class="session-id">Last updated on 2023-12-07 (Session: abc123def456)</p>
+ <p class="session-id">Last updated on 2023-12-07</p>
```

## Step 3: Declare the filter

Open your service declaration file (e.g. `declarations/MyService.json`) and locate the `filter` property of the specific terms you want to apply the filter to. If it doesn't exist, add it as an array.

### Filter without parameters

For filters that don’t require parameters, add the filter name as a string:

```json
{
"name": "MyService",
"terms": {
"Privacy Policy": {
"fetch": "https://my.service.example/en/privacy-policy",
"select": ".textcontent",
"filter": [
"removeSessionIds"
]
}
}
}
```

### Filter with parameters

For filters that take parameters, use an object format, for example with the built-in filter `removeQueryParams` to remove query parameters from URLs:

```json
{
"name": "MyService",
"terms": {
"Privacy Policy": {
"fetch": "https://my.service.example/en/privacy-policy",
"select": ".textcontent",
"filter": [
{
"removeQueryParams": ["utm_source", "utm_medium", "utm_campaign"]
}
]
}
}
}
```

### Multiple filters

You can combine multiple filters in the same declaration:

```json
{
"name": "MyService",
"terms": {
"Privacy Policy": {
"fetch": "https://my.service.example/en/privacy-policy",
"select": ".textcontent",
"filter": [
{
"removeQueryParams": ["utm_source", "utm_medium"]
},
"removeSessionIds"
]
}
}
}
```

## Step 4: Test the filter

After adding the filter, test your declaration to ensure it works correctly:

1. Start the terms tracking process
2. Check that the noise has been removed
3. Verify that important content is preserved
27 changes: 27 additions & 0 deletions content/terms/reference/built-in-filters.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,27 @@
---
title: "Built-in filters"
---

# Built-in filters

This reference details all available built-in [filters]({{< relref "terms/explanation/filters" >}}) that can be applied to avoid noise in versions.

{{< refItem
name="removeQueryParams"
description="Removes specified query parameters from URLs in links and images."
>}}

```json
"filter": [
{
"removeQueryParams": ["utm_source", "utm_medium"]
}
]
```

```diff
- <p>Read the <a href="https://example.com/example-page?utm_source=OGB&utm_medium=website&lang=en">list of our affiliates</a>.</p>
+ <p>Read the <a href="https://example.com/example-page?lang=en">list of our affiliates</a>.</p>
```

{{< /refItem >}}
16 changes: 12 additions & 4 deletions content/terms/reference/declaration.md
Original file line number Diff line number Diff line change
Expand Up @@ -139,10 +139,18 @@ As an array of those:

{{< refItem
name="filter"
type="array of strings"
description="Array of filter function names to apply. Function will be executed in the order of the array. See the [Filters]({{< relref \"terms/reference/filters\" >}}) section for more information."
example="[\"filterName1\", \"filterName2\"]"
/>}}
type="array of strings or objects"
description="Array of filter functions to apply. Each item can be either a string (function name) or an object (function name as key, parameters as value). Functions will be executed in the order of the array. See the [Filters]({{< relref \"terms/reference/filters\" >}}) section for more information."
>}}
```json
"filter": [
"filterName1",
{
"filterName2": "param"
}
]
```
{{< /refItem >}}

{{< refItem
name="combine"
Expand Down
Loading