Add text replacement functionality to bulk import API

Summary

Enhance the bulk import API to support text pattern replacement during migration, allowing users to automatically update references, URLs, and other text content when migrating between GitLab instances.

Problem to Solve

When migrating projects/groups from on-premises GitLab to GitLab.com (or between instances), many text references become outdated:

Issue/MR descriptions contain old instance URLs
Comments reference old project paths
Documentation links point to the previous instance
Team and Person mentions are inaccurate
Manual post-migration cleanup is time-consuming and error-prone, also it messes up timestamps

Proposal

Add a text_replacements parameter to the bulk import API that accepts an array of find-and-replace patterns:

{
  "configuration": {
    "url": "https://source-gitlab.example.com",
    "access_token": "glpat-xxxxxxxxxxxxxxxxxxxx"
  },
  "entities": [
    {
      "source_type": "group_entity",
      "source_full_path": "my-group",
      "destination_name": "my-group",
      "destination_namespace": "my-org"
    }
  ],
  "text_replacements": [
    {
      "source_text": "https://on-prem.gitlab.example/group1",
      "target_text": "https://gitlab.com/organization1/group1"
    },
    {
      "source_text": "https://on-prem.gitlab.example/group2", 
      "target_text": "https://gitlab.com/organization2/group1"
    },
    {
      "source_text": "@internal-team",
      "target_text": "@company-org/internal-team"
    },
    {
      "source_text": "JIRA-\\d+",
      "target_text": "https://company.atlassian.net/browse/$0",
      "regex": true
    }
  ]
}

Scope

Apply text replacements to:

Issue descriptions and comments
Merge request descriptions and comments
Epic descriptions and comments
Milestone descriptions
Wiki content (?)
Project/group descriptions
Release descriptions

Implementation Considerations

Add validation for text_replacements parameter
Implement replacement logic in relevant pipeline classes
Support regex patterns for advanced matching
Provide dry-run option to preview changes
Add logging for replacement operations
Consider performance impact on large migrations

Benefits

Reduces manual post-migration work
Ensures consistent reference updates
Improves migration experience for enterprise users
Maintains content integrity across instances

Current API Context

The bulk import API currently accepts the following structure according to the official documentation:

curl --request POST --header "PRIVATE-TOKEN: <your_access_token>" \
     --header "Content-Type: application/json" \
     --data '{
       "configuration": {
         "url": "https://source.gitlab.example.com",
         "access_token": "glpat-xxxxxxxxxxxxxxxxxxxx"
       },
       "entities": [
         {
           "source_type": "group_entity",
           "source_full_path": "my-group",
           "destination_name": "my-group", 
           "destination_namespace": "my-org"
         }
       ]
     }' \
     "https://gitlab.example.com/api/v4/bulk_imports"

Technical Implementation Details

The text replacement feature should integrate with the existing BulkImports pipeline architecture:

Affected Pipeline Classes:

BulkImports::Groups::Pipelines::* - Group-level content
BulkImports::Projects::Pipelines::* - Project-level content
Specifically: IssuesPipeline, MergeRequestsPipeline, MilestonesPipeline, etc.

Processing Location: Text replacements should occur during the transformation phase of each pipeline, before data is written to the target instance.

Use Cases

Enterprise Migration Scenario:

{
  "text_replacements": [
    {
      "source_text": "https://gitlab.internal.company.com",
      "target_text": "https://gitlab.com/company-org"
    },
    {
      "source_text": "@internal-team",
      "target_text": "@company-org/internal-team"
    },
    {
      "source_text": "JIRA-\\d+",
      "target_text": "https://company.atlassian.net/browse/$0",
      "regex": true
    }
  ]
}

API Response Enhancement

The bulk import status response should include replacement statistics:

{
  "id": 1,
  "status": "finished",
  "text_replacements_applied": {
    "total_replacements": 247,
    "entities_affected": 89,
    "breakdown": {
      "issues": 45,
      "merge_requests": 23,
      "comments": 179
    }
  }
}

Backward Compatibility

The text_replacements parameter should be optional to maintain compatibility with existing integrations.