Extraction Helpers

Extract data from PR content using regular expressions.

Overview

Extraction helpers use regex patterns with capture groups to pull specific data from PR descriptions, titles, or comments. The extracted values can be used in field updates, task creation, or PR comments.

All extraction helpers return the first capture group from the regex match, or an empty string if no match is found.

Helpers

extract_from_body

Extract data from the PR description (body).

Syntax:

{{extract_from_body "pattern"}}

Parameters:

• pattern - Regular expression with a capture group

Returns: First capture group value, or empty string

Example:

update_fields:
  '1234567890': '{{extract_from_body "Version: ([\\d.]+)"}}'

extract_from_title

Extract data from the PR title.

Syntax:

{{extract_from_title "pattern"}}

Parameters:

• pattern - Regular expression with a capture group

Returns: First capture group value, or empty string

Example:

update_fields:
  '1234567890': '{{extract_from_title "\\[(\\w+)\\]"}}'

extract_from_comments

Extract data from PR comments.

Syntax:

{{extract_from_comments "pattern"}}

Parameters:

• pattern - Regular expression with a capture group

Returns: First capture group value, or empty string

Example:

update_fields:
  '1234567890': '{{extract_from_comments "Build #(\\d+)"}}'

Automatic Comment Fetching

The action automatically detects when your rules use extract_from_comments and only fetches PR comments when needed. You don't need to configure anything—it's optimized for you.

How it works:

1. Action analyzes your rules configuration at startup
2. If any rule contains extract_from_comments, comments are fetched
3. If no rule uses it, comments are NOT fetched (saves 1 API call)
4. Comments are fetched once and reused for all rules

Performance Impact

• Without extract_from_comments: 0 comment API calls
• With extract_from_comments: 1 comment API call (regardless of how many rules use it)

The optimization happens automatically—you don't need to worry about API efficiency.

Regex Patterns

Capture Groups

Use parentheses () to define what to extract:

Version: ([\\d.]+)
         ^^^^^^^
         This is captured

Common Patterns

Pattern	Description	Example Match	Captures
([\\d.]+)	Version numbers	1.2.3	1.2.3
(\\d+)	Integers	42	42
\\[(\\w+)\\]	Bracketed text	[urgent]	urgent
@(\\w+)	Username	@octocat	octocat
#(\\d+)	Issue/PR number	#123	123

Escaping

In YAML strings, backslashes must be doubled:

# ❌ Wrong - single backslash
'\\d+'

# ✅ Correct - double backslash
'\\\\d+'

Examples

Extract Version Number

From PR body like:

Version: 2.1.0

rules:
  - when:
      event: pull_request
      action: opened
    then:
      update_fields:
        '1234567890': '{{extract_from_body "Version: ([\\d.]+)"}}'

Extract Category from Title

From PR title like: [backend] Add user authentication

rules:
  - when:
      event: pull_request
      action: opened
    then:
      update_fields:
        '1234567890': '{{extract_from_title "\\[(\\w+)\\]"}}'

Extract Build Number from Comments

From comment like: Build #12345 created

rules:
  - when:
      event: pull_request
      action: labeled
      label: 'build_created'
    then:
      update_fields:
        '1234567890': '{{extract_from_comments "Build #(\\d+)"}}'

Extract Multiple Values

Use multiple extractions:

rules:
  - when:
      event: pull_request
      action: opened
    then:
      update_fields:
        '1111111111': '{{extract_from_body "Version: ([\\d.]+)"}}'
        '2222222222': '{{extract_from_body "Environment: (\\w+)"}}'
        '3333333333': '{{extract_from_title "\\[(\\w+)\\]"}}'

Extract with Fallback

Use with or helper for defaults:

update_fields:
  '1234567890': '{{or (extract_from_body "Version: ([\\d.]+)") "Unknown"}}'

Complex Patterns

Extract from structured text:

# From: "Deployed to: production (v2.1.0)"
update_fields:
  '1234567890': '{{extract_from_comments "Deployed to: (\\w+)"}}'  # "production"
  '1111111111': '{{extract_from_comments "\\(v([\\d.]+)\\)"}}'     # "2.1.0"

Empty Results

If no match is found, extraction returns empty string:

# If pattern doesn't match, field update is skipped
update_fields:
  '1234567890': '{{extract_from_body "NotFound: (\\w+)"}}'  # Empty, skipped

TIP

Empty template values skip field updates. This is intentional - you can safely extract without worrying about setting fields to empty strings.

Common Use Cases

Dependency Versions

Track dependency update versions:

rules:
  - when:
      event: pull_request
      action: opened
      author: 'dependabot[bot]'
    then:
      create_task:
        project: '1234567890'
        workspace: '0987654321'
        title: '{{pr.title}}'
        initial_fields:
          '1111111111': '{{extract_from_title "from [\\d.]+ to ([\\d.]+)"}}'

Build Numbers

Track CI build numbers:

rules:
  - when:
      event: pull_request
      action: labeled
      label: 'build_created'
    then:
      update_fields:
        '1234567890': '{{extract_from_comments "Build: #(\\d+)"}}'
        '1111111111': '{{extract_from_comments "Version: ([\\d.]+)"}}'

Environment Tags

Extract deployment environment:

rules:
  - when:
      event: pull_request
      action: labeled
      label: 'deployed'
    then:
      update_fields:
        '1234567890': '{{extract_from_comments "Deployed to (\\w+)"}}'

Issue References

Extract linked issue numbers:

rules:
  - when:
      event: pull_request
      action: opened
    then:
      update_fields:
        '1234567890': '{{extract_from_body "Fixes #(\\d+)"}}'

Regex Tips

Test Your Patterns

Use regex101.com to test patterns:

1. Select "JavaScript" flavor
2. Paste your pattern (single backslashes)
3. Test against sample text
4. Double backslashes for YAML

Common Mistakes

# ❌ Wrong - forgot capture group
'Version: \\d+'

# ✅ Correct - added capture group
'Version: (\\d+)'

# ❌ Wrong - single backslash
'\\d+'

# ✅ Correct - double backslash in YAML
'\\\\d+'

# ❌ Wrong - greedy match gets too much
'Version: (.+)'

# ✅ Correct - specific pattern
'Version: ([\\d.]+)'

Error Handling

Invalid regex patterns log errors but don't fail the action:

# Invalid pattern - logs error, returns empty string
'{{extract_from_body "(?invalid)"}}'

Check workflow logs for regex errors.

Performance

Extractions are cached per template evaluation. Multiple uses of the same extraction don't re-run the regex.

See Also

• or Helper - Provide fallback values
• clean_title - Remove commit prefixes
• Context Variables - Available template variables
• Template System - Handlebars overview