feat: add opt-in default npm package excludes from yarn

MODULE.bazel

@@ -153,9 +153,6 @@ npm.npm_translate_lock(
         "//npm/private/test/vendored/is-odd:package.json",
         "//npm/private/test/vendored/semver-max:package.json",
     ],
-    exclude_package_contents = {
-        "chalk": ["**/README*"],
-    },
     generate_bzl_library_targets = True,
     lifecycle_hooks = {
         # We fetch @kubernetes/client-node from source and it has a `prepare` lifecycle hook that needs to be run
@@ -215,6 +212,12 @@ npm.npm_translate_lock(
     verify_node_modules_ignored = "//:.bazelignore",
     verify_patches = "//examples/npm_deps/patches:patches",
 )
+
+# Configure package exclusions using tag class
+npm.npm_exclude_package_contents(
+    package = "chalk",
+    patterns = ["**/README*"],
+)
 use_repo(
     npm,
     "npm",

docs/troubleshooting.md

@@ -175,19 +175,27 @@ If you are experiencing slower than expected builds, you can try disabling or re
 Npm packages sometimes include unnecessary files such as tests, test data etc. Large files or a large number of files
 can effect performance and are sometimes worth explicitly excluding content.
 
-In these cases you can add such packages and the respective files/folders you want to exclude to your npm_translate_lock
-rule in the `exclude_package_contents` attribute like so:
+In these cases you can add such packages and the respective files/folders you want to exclude.
 
+**For WORKSPACE builds:**
 ```starlark
-npm.npm_translate_lock(
+npm_translate_lock(
     ...
     exclude_package_contents = {
         "resolve": ["**/test/*"],
     },
 )
 ```
 
-This example will remove the test folder.
+**For Bzlmod builds (MODULE.bazel):**
+```starlark
+npm.npm_exclude_package_contents(
+    package = "resolve",
+    patterns = ["**/test/*"],
+)
+```
+
+These examples will remove the test folder from the "resolve" package.
 
 You can use this to remove whatever you find to be not needed for your project.
 

e2e/npm_translate_lock_exclude_package_contents/MODULE.bazel

@@ -16,15 +16,20 @@ npm = use_extension(
 npm.npm_translate_lock(
     name = "npm",
     data = ["//:package.json"],
-    exclude_package_contents = {
-        "is-odd@3.0.1": [
-            "**/README*",
-            "",
-            "**/LICENSE*",
-        ],
-    },
     npmrc = "//:.npmrc",
     pnpm_lock = "//:pnpm-lock.yaml",
     verify_node_modules_ignored = "//:.bazelignore",
 )
+npm.npm_exclude_package_contents(
+    package = "*",
+    use_defaults = True,
+)
+npm.npm_exclude_package_contents(
+    package = "is-odd@3.0.1",
+    patterns = [
+        "**/README*",
+        "**/LICENSE*",
+    ],
+    use_defaults = True,
+)
 use_repo(npm, "npm")

e2e/npm_translate_lock_exclude_package_contents/WORKSPACE

@@ -16,11 +16,11 @@ load("@aspect_rules_js//npm:repositories.bzl", "npm_translate_lock")
 npm_translate_lock(
     name = "npm",
     exclude_package_contents = {
+        "*": True,  # Use default exclusions for all packages
         "is-odd@3.0.1": [
             "**/README*",
-            "",
             "**/LICENSE*",
-        ],
+        ],  # Custom exclusions for specific package
     },
     npmrc = "//:.npmrc",
     pnpm_lock = "//:pnpm-lock.yaml",

npm/extensions.bzl

@@ -6,6 +6,7 @@ load("@aspect_bazel_lib//lib:repo_utils.bzl", "repo_utils")
 load("@aspect_bazel_lib//lib:utils.bzl", bazel_lib_utils = "utils")
 load("@bazel_features//:features.bzl", "bazel_features")
 load("//npm:repositories.bzl", "npm_import", "pnpm_repository", _DEFAULT_PNPM_VERSION = "DEFAULT_PNPM_VERSION", _LATEST_PNPM_VERSION = "LATEST_PNPM_VERSION")
+load("//npm/private:exclude_package_contents_default.bzl", "exclude_package_contents_default")
 load("//npm/private:npm_import.bzl", "npm_import_lib", "npm_import_links_lib")
 load("//npm/private:npm_translate_lock.bzl", "npm_translate_lock_lib", "npm_translate_lock_rule")
 load("//npm/private:npm_translate_lock_helpers.bzl", npm_translate_lock_helpers = "helpers")
@@ -24,16 +25,19 @@ def _npm_extension_impl(module_ctx):
         # ctx.actions.declare_symlink was added in Bazel 6
         fail("A minimum version of Bazel 6 required to use rules_js")
 
+    # Collect all exclude_package_contents tags and build exclusion dictionary
+    exclude_package_contents_config = _build_exclude_package_contents_config(module_ctx)
+
     for mod in module_ctx.modules:
         for attr in mod.tags.npm_translate_lock:
-            _npm_translate_lock_bzlmod(attr)
+            _npm_translate_lock_bzlmod(attr, exclude_package_contents_config)
 
             # We cannot read the pnpm_lock file before it has been bootstrapped.
             # See comment in e2e/update_pnpm_lock_with_import/test.sh.
             if attr.pnpm_lock:
                 if hasattr(module_ctx, "watch"):
                     module_ctx.watch(attr.pnpm_lock)
-                _npm_lock_imports_bzlmod(module_ctx, attr)
+                _npm_lock_imports_bzlmod(module_ctx, attr, exclude_package_contents_config)
 
         for i in mod.tags.npm_import:
             _npm_import_bzlmod(i)
@@ -44,7 +48,29 @@ def _npm_extension_impl(module_ctx):
         )
     return module_ctx.extension_metadata()
 
-def _npm_translate_lock_bzlmod(attr):
+def _build_exclude_package_contents_config(module_ctx):
+    """Build exclude_package_contents configuration from tags across all modules."""
+    exclusions = {}
+
+    for mod in module_ctx.modules:
+        for exclude_tag in mod.tags.npm_exclude_package_contents:
+            # Process the package in the tag
+            package = exclude_tag.package
+            if package in exclusions:
+                fail("Duplicate exclude_package_contents tag for package: {}".format(package))
+
+            exclusions[package] = []
+
+            # Add default exclusions if requested
+            if exclude_tag.use_defaults:
+                exclusions[package].extend(exclude_package_contents_default)
+
+            # Add custom patterns
+            exclusions[package].extend(exclude_tag.patterns)
+
+    return exclusions
+
+def _npm_translate_lock_bzlmod(attr, exclude_package_contents_config):
     npm_translate_lock_rule(
         name = attr.name,
         bins = attr.bins,
@@ -74,10 +100,11 @@ def _npm_translate_lock_bzlmod(attr):
         verify_node_modules_ignored = attr.verify_node_modules_ignored,
         verify_patches = attr.verify_patches,
         yarn_lock = attr.yarn_lock,
+        exclude_package_contents = exclude_package_contents_config,
         bzlmod = True,
     )
 
-def _npm_lock_imports_bzlmod(module_ctx, attr):
+def _npm_lock_imports_bzlmod(module_ctx, attr, exclude_package_contents_config):
     state = npm_translate_lock_state.new(attr.name, module_ctx, attr, True)
 
     importers, packages = translate_to_transitive_closure(
@@ -131,6 +158,7 @@ WARNING: Cannot determine home directory in order to load home `.npmrc` file in
         registries = registries,
         default_registry = state.default_registry(),
         npm_auth = npm_auth,
+        exclude_package_contents_config = exclude_package_contents_config,
     )
 
     system_tar = detect_system_tar(module_ctx)
@@ -218,6 +246,7 @@ def _npm_translate_lock_attrs():
 
     # Args not supported or unnecessary in bzlmod
     attrs.pop("repositories_bzl_filename")
+    attrs.pop("exclude_package_contents")  # Use tag classes only for MODULE.bazel
 
     return attrs
 
@@ -237,11 +266,28 @@ def _npm_import_attrs():
 
     return attrs
 
+def _npm_exclude_package_contents_attrs():
+    return {
+        "package": attr.string(
+            doc = "Package name to apply exclusions to. Supports wildcards like '*' for all packages.",
+            mandatory = True,
+        ),
+        "patterns": attr.string_list(
+            doc = "List of glob patterns to exclude from the specified package.",
+            default = [],
+        ),
+        "use_defaults": attr.bool(
+            doc = "Whether to use default exclusion patterns for the specified package. Defaults are as to Yarn autoclean: https://github.com/yarnpkg/yarn/blob/7cafa512a777048ce0b666080a24e80aae3d66a9/src/cli/commands/autoclean.js#L16",
+            default = False,
+        ),
+    }
+
 npm = module_extension(
     implementation = _npm_extension_impl,
     tag_classes = {
         "npm_translate_lock": tag_class(attrs = _npm_translate_lock_attrs()),
         "npm_import": tag_class(attrs = _npm_import_attrs()),
+        "npm_exclude_package_contents": tag_class(attrs = _npm_exclude_package_contents_attrs()),
     },
 )
 

npm/private/BUILD.bazel

@@ -168,6 +168,7 @@ bzl_library(
     name = "npm_translate_lock",
     srcs = ["npm_translate_lock.bzl"],
     deps = [
+        ":exclude_package_contents_default.bzl",
         ":list_sources",
         ":npm_translate_lock_generate",
         ":npm_translate_lock_helpers",

npm/private/exclude_package_contents_default.bzl

@@ -0,0 +1,53 @@
+"""A default files exclude list for common packages.
+
+Based on Yarn autoclean; see
+https://github.com/yarnpkg/yarn/blob/7cafa512a777048ce0b666080a24e80aae3d66a9/src/cli/commands/autoclean.js#L16
+"""
+
+exclude_package_contents_default = [
+    # test directories
+    "**/__tests__/**",
+    "**/test/**",
+    "**/tests/**",
+    "**/powered-test/**",
+
+    # asset directories
+    "**/docs/**",
+    "**/doc/**",
+    "**/website/**",
+    "**/images/**",
+    "**/assets/**",
+
+    # examples
+    "**/example/**",
+    "**/examples/**",
+
+    # code coverage directories
+    "**/coverage/**",
+    "**/.nyc_output/**",
+
+    # build scripts (files)
+    "Makefile",
+    "Gulpfile.js",
+    "Gruntfile.js",
+
+    # configs (files)
+    "appveyor.yml",
+    "circle.yml",
+    "codeship-services.yml",
+    "codeship-steps.yml",
+    "wercker.yml",
+    ".tern-project",
+    ".gitattributes",
+    ".editorconfig",
+    ".*ignore",
+    ".eslintrc",
+    ".jshintrc",
+    ".flowconfig",
+    ".documentup.json",
+    ".yarn-metadata.json",
+    ".travis.yml",
+
+    # misc (files)
+    "*.md",
+]

npm/private/npm_translate_lock.bzl

@@ -28,6 +28,7 @@ Advanced users may want to directly fetch a package from npm rather than start f
 load("@aspect_bazel_lib//lib:utils.bzl", bazel_lib_utils = "utils")
 load("@aspect_bazel_lib//lib:write_source_files.bzl", "write_source_file")
 load("@bazel_skylib//lib:paths.bzl", "paths")
+load(":exclude_package_contents_default.bzl", "exclude_package_contents_default")
 load(":list_sources.bzl", "list_sources")
 load(":npm_translate_lock_generate.bzl", "generate_repository_files")
 load(":npm_translate_lock_helpers.bzl", "helpers")
@@ -43,6 +44,31 @@ RULES_JS_FROZEN_PNPM_LOCK_ENV = "ASPECT_RULES_JS_FROZEN_PNPM_LOCK"
 DEFAULT_REPOSITORIES_BZL_FILENAME = "repositories.bzl"
 DEFAULT_DEFS_BZL_FILENAME = "defs.bzl"
 
+def _normalize_exclude_package_contents(exclude_package_contents):
+    """Normalize exclude_package_contents dictionary for string_list_dict format."""
+    if not exclude_package_contents:
+        return {}
+
+    result = {}
+    for package, value in exclude_package_contents.items():
+        if type(value) == "bool":
+            if value == True:
+                # True means use default exclusions
+                result[package] = exclude_package_contents_default
+            else:
+                # False means no exclusions
+                result[package] = []
+        elif type(value) == "list":
+            # Lists must contain only strings
+            for item in value:
+                if type(item) != "string":
+                    fail("exclude_package_contents list values must be strings. Got: {} in package '{}'".format(type(item), package))
+            result[package] = value
+        else:
+            fail("exclude_package_contents values must be boolean or string list. Got: {} for package '{}'".format(type(value), package))
+
+    return result
+
 _ATTRS = {
     "additional_file_contents": attr.string_list_dict(),
     "bins": attr.string_list_dict(),
@@ -284,18 +310,26 @@ def npm_translate_lock(
 
             Read more: [patching](/docs/pnpm.md#patching)
 
-        exclude_package_contents: A map of package names or package names with their version (e.g., "my-package" or "my-package@v1.2.3")
-            to a list of patterns to exclude from the package's generated node_modules link targets. Multiple matches are additive.
+        exclude_package_contents: Configuration for excluding package contents (WORKSPACE only).
+
+            For MODULE.bazel, use the `exclude_package_contents` tag class instead.
+
+            The configuration is a dictionary that maps package names (or package names with their version, e.g., "my-package" or "my-package@v1.2.3") to exclusion rules.
+
+            Values can be:
+            - `True`: Use default exclusions
+            - List of strings: Multiple exclusion patterns
 
             Versions must match if used.
 
-            For example,
+            Example:
 
             ```
             exclude_package_contents = {
+                "*": True,  # Use defaults for all packages
                 "@foo/bar": ["**/test/**"],
                 "@foo/car@2.0.0": ["**/README*"],
-            },
+            }
             ```
         patch_tool: The patch tool to use. If not specified, the `patch` from `PATH` is used.
 
@@ -591,7 +625,7 @@ def npm_translate_lock(
         npmrc = npmrc,
         use_home_npmrc = use_home_npmrc,
         patches = patches,
-        exclude_package_contents = exclude_package_contents,
+        exclude_package_contents = _normalize_exclude_package_contents(exclude_package_contents),
         patch_tool = patch_tool,
         patch_args = patch_args,
         custom_postinstalls = custom_postinstalls,

npm/private/npm_translate_lock_generate.bzl

@@ -541,7 +541,7 @@ def _gen_npm_import(rctx, system_tar, _import, link_workspace):
     maybe_replace_package = ("""
         replace_package = "%s",""" % _import.replace_package) if _import.replace_package else ""
     maybe_exclude_package_contents = ("""
-        exclude_package_contents = %s,""" % _import.exclude_package_contents) if _import.exclude_package_contents != None else ""
+        exclude_package_contents = %s,""" % starlark_codegen_utils.to_list_attr(_import.exclude_package_contents)) if _import.exclude_package_contents != None else ""
 
     return _NPM_IMPORT_TMPL.format(
         link_packages = starlark_codegen_utils.to_dict_attr(_import.link_packages, 2, quote_value = False),