Monorepo Extraction

Monorepo Extraction

Extract and publish packages from a monorepo as standalone repositories.

The Challenge

Monorepos contain multiple packages, but users may want:

• Individual packages as separate repos
• Clean history for just that package
• Independent versioning
• Simplified dependency management

Basic Extraction

core.workflow(
    name = "extract-my-package",
    origin = git.origin(
        url = "https://github.com/company/monorepo",
        ref = "main",
    ),
    destination = git.github_destination(
        url = "https://github.com/company/my-package",
        push = "main",
    ),
    origin_files = glob(["packages/my-package/**"]),
    transformations = [
        # Flatten to root
        core.move("packages/my-package/", ""),
    ],
    authoring = authoring.pass_thru("Sync <sync@company.com>"),
    mode = "SQUASH",
)

Multi-Package Extraction

Extract multiple packages:

def extract_package(name):
    return core.workflow(
        name = "extract-" + name,
        origin = git.origin(
            url = "https://github.com/company/monorepo",
            ref = "main",
        ),
        destination = git.github_destination(
            url = "https://github.com/company/" + name,
            push = "main",
        ),
        origin_files = glob(["packages/" + name + "/**"]),
        transformations = [
            core.move("packages/" + name + "/", ""),
        ],
        authoring = authoring.pass_thru("Sync <sync@company.com>"),
        mode = "SQUASH",
    )

# Create workflows for each package
extract_package("core")
extract_package("cli")
extract_package("utils")

Run specific extraction:

java -jar copybara.jar migrate copy.bara.sky extract-core
java -jar copybara.jar migrate copy.bara.sky extract-cli

Fixing Import Paths

Update internal imports:

transformations = [
    core.move("packages/my-package/", ""),

    # Update monorepo-style imports
    core.replace(
        before = "@company/my-package",
        after = "my-package",
        paths = glob(["**/*.ts", "**/*.js"]),
    ),

    # Update workspace references
    core.replace(
        before = '"workspace:*"',
        after = '"*"',
        paths = glob(["package.json"]),
    ),
]

With Shared Dependencies

Include shared code that the package depends on:

origin_files = glob([
    "packages/my-package/**",
    "packages/shared/**",  # Include shared deps
])

transformations = [
    core.move("packages/my-package/", ""),
    core.move("packages/shared/", "vendor/shared/"),

    # Update import paths
    core.replace(
        before = "@company/shared",
        after = "./vendor/shared",
    ),
]

Preserving History

Use ITERATIVE mode to preserve commit history:

core.workflow(
    name = "extract-with-history",
    mode = "ITERATIVE",  # Preserve individual commits
    origin_files = glob(["packages/my-package/**"]),
    transformations = [
        core.move("packages/my-package/", ""),
    ],
    ...
)

Package.json Updates

Clean up monorepo-specific fields:

transformations = [
    core.move("packages/my-package/", ""),

    # Remove workspace field
    core.replace(
        before = '"workspaces": [${content}],',
        after = "",
        regex_groups = {"content": "[^]]*"},
        paths = glob(["package.json"]),
    ),

    # Update repository field
    core.replace(
        before = '"directory": "packages/my-package"',
        after = "",
        paths = glob(["package.json"]),
    ),
]

Via Pull Request

Review extractions before publishing:

destination = git.github_pr_destination(
    url = "https://github.com/company/my-package",
    destination_ref = "main",
    pr_branch = "sync/${CONTEXT_REFERENCE}",
    title = "Sync from monorepo",
)

Complete Example

MONOREPO = "https://github.com/company/monorepo"

def extract_package(name, npm_name = None):
    if not npm_name:
        npm_name = name

    return core.workflow(
        name = "extract-" + name,
        origin = git.github_origin(url = MONOREPO, ref = "main"),
        destination = git.github_pr_destination(
            url = "https://github.com/company/" + name,
            destination_ref = "main",
            pr_branch = "sync/${CONTEXT_REFERENCE}",
        ),
        origin_files = glob([
            "packages/" + name + "/**",
            "LICENSE",  # Include license
        ]),
        destination_files = glob(
            include = ["**"],
            exclude = ["README.md"],  # Keep standalone readme
        ),
        authoring = authoring.pass_thru("Extract Bot <bot@company.com>"),
        transformations = [
            core.move("packages/" + name + "/", ""),
            core.replace(
                before = "@company/" + npm_name,
                after = npm_name,
            ),
            metadata.add_header("Extracted from: " + MONOREPO),
        ],
        mode = "SQUASH",
    )

extract_package("core", "company-core")
extract_package("cli", "company-cli")
extract_package("utils")

Next Steps

• Documentation sync
• GitHub Actions