Merge pull request #3135 from nexB/packagecode-updates

Separate Package parsing functions

Signed-off-by: Philippe Ombredanne <pombredanne@nexb.com>

Author: pombredanne

CHANGELOG.rst

@@ -62,6 +62,18 @@ License detection:
   removed. This new command supports simpler reindexing using custom
   license texts and license rules contributed by plugins or stored in an
   additional directory. 
+
+Package detection:
+~~~~~~~~~~~~~~~~~~~~~
+
+- Code for parsing a Maven POM, npm package.json, freebsd manifest and haxelib
+  JSON have been separated into two functions: one that creates a PackageData
+  object from the parsed Resource, and another that calls the previous function
+  and yields the PackageData. This was done such that we can use the package
+  manifest data parsing code outside of the scancode-toolkit context in other
+  libraries.
+
+
 v31.2.1 - 2022-10-05
 ----------------------------------
 

setup.cfg

@@ -1,6 +1,6 @@
 [metadata]
 name = scancode-toolkit
-version = 31.2.1
+version = 31.2.2
 license = Apache-2.0 AND CC-BY-4.0 AND LicenseRef-scancode-other-permissive AND LicenseRef-scancode-other-copyleft
 
 # description must be on ONE line https://github.com/pypa/setuptools/issues/1390

src/packagedcode/chef.py

@@ -55,7 +55,7 @@ def chef_api_url(name, version):
     return name and version and f'https://supermarket.chef.io/api/v1/cookbooks/{name}/versions/{version}'
 
 
-def get_urls(name, version):
+def get_urls(name, version, **kwargs):
     """
     Return a mapping of URLs given a name and version.
     """

src/packagedcode/cocoapods.py

@@ -434,7 +434,7 @@ def parse(cls, location):
         )
 
 
-def get_urls(name=None, version=None, homepage_url=None, vcs_url=None):
+def get_urls(name=None, version=None, homepage_url=None, vcs_url=None, **kwargs):
     """
     Return a mapping of podspec URLS.
     """

src/packagedcode/freebsd.py

@@ -42,20 +42,13 @@ class CompactManifestHandler(models.DatafileHandler):
     documentation_url = 'https://www.freebsd.org/cgi/man.cgi?pkg-create(8)#MANIFEST_FILE_DETAILS'
 
     @classmethod
-    def parse(cls, location):
-        """
-        Yield one or more Package manifest objects given a file ``location`` pointing to a
-        package archive, manifest or similar.
-        """
-        with io.open(location, encoding='utf-8') as loc:
-            freebsd_manifest = saneyaml.load(loc)
-
+    def _parse(cls, yaml_data):
         package_data = models.PackageData(
             datasource_id=cls.datasource_id,
             type=cls.default_package_type,
             qualifiers=dict(
-                arch=freebsd_manifest.get('arch'),
-                origin=freebsd_manifest.get('origin'),
+                arch=yaml_data.get('arch'),
+                origin=yaml_data.get('origin'),
             )
         )
 
@@ -69,7 +62,7 @@ def parse(cls, location):
         ]
 
         for source, target in plain_fields:
-            value = freebsd_manifest.get(source)
+            value = yaml_data.get(source)
             if value:
                 if isinstance(value, str):
                     value = value.strip()
@@ -87,17 +80,28 @@ def parse(cls, location):
 
         for source, func in field_mappers:
             logger.debug('parse: %(source)r, %(func)r' % locals())
-            value = freebsd_manifest.get(source) or None
+            value = yaml_data.get(source) or None
             if value:
                 func(value, package_data)
 
         # license_mapper needs multiple fields
-        license_mapper(freebsd_manifest, package_data)
+        license_mapper(yaml_data, package_data)
 
         if package_data.declared_license:
             package_data.license_expression = cls.compute_normalized_license(package_data)
 
-        yield package_data
+        return package_data
+
+    @classmethod
+    def parse(cls, location):
+        """
+        Yield one or more Package manifest objects given a file ``location`` pointing to a
+        package archive, manifest or similar.
+        """
+        with io.open(location, encoding='utf-8') as loc:
+            yaml_data = saneyaml.load(loc)
+
+        yield cls._parse(yaml_data)
 
     @classmethod
     def compute_normalized_license(cls, package):

src/packagedcode/haxe.py

@@ -45,26 +45,7 @@ class HaxelibJsonHandler(models.DatafileHandler):
     documentation_url = 'https://lib.haxe.org/documentation/creating-a-haxelib-package/'
 
     @classmethod
-    def parse(cls, location):
-        """
-        Yield one or more Package manifest objects given a file ``location`` pointing to a
-        package_data archive, manifest or similar.
-
-        {
-            "name": "haxelib",
-            "url" : "https://lib.haxe.org/documentation/",
-            "license": "GPL",
-            "tags": ["haxelib", "core"],
-            "description": "The haxelib client",
-            "classPath": "src",
-            "version": "3.4.0",
-            "releasenote": " * Fix password input issue in Windows (#421).\n * ....",
-            "contributors": ["back2dos", "ncannasse", "jason", "Simn", "nadako", "andyli"]
-        }
-        """
-        with io.open(location, encoding='utf-8') as loc:
-            json_data = json.load(loc)
-
+    def _parse(cls, json_data):
         name = json_data.get('name')
         version = json_data.get('version')
 
@@ -110,4 +91,27 @@ def parse(cls, location):
             dep = models.DependentPackage(purl=dep_purl, is_resolved=is_resolved,)
             package_data.dependencies.append(dep)
 
-        yield package_data
+        return package_data
+
+    @classmethod
+    def parse(cls, location):
+        """
+        Yield one or more Package manifest objects given a file ``location`` pointing to a
+        package_data archive, manifest or similar.
+
+        {
+            "name": "haxelib",
+            "url" : "https://lib.haxe.org/documentation/",
+            "license": "GPL",
+            "tags": ["haxelib", "core"],
+            "description": "The haxelib client",
+            "classPath": "src",
+            "version": "3.4.0",
+            "releasenote": " * Fix password input issue in Windows (#421).\n * ....",
+            "contributors": ["back2dos", "ncannasse", "jason", "Simn", "nadako", "andyli"]
+        }
+        """
+        with io.open(location, encoding='utf-8') as loc:
+            json_data = json.load(loc)
+
+        yield cls._parse(json_data)

src/packagedcode/maven.py

@@ -891,12 +891,12 @@ def has_basic_pom_attributes(pom):
     return basics
 
 
-def get_maven_pom(location=None):
+def get_maven_pom(location=None, text=None):
     """
     Return a MavenPom object from a POM file at `location` or provided as a
     `text` string.
     """
-    pom = MavenPom(location=location)
+    pom = MavenPom(location=location, text=text)
 
     extra_properties = {}
 
@@ -1043,7 +1043,7 @@ def get_parties(pom):
     return parties
 
 
-def get_urls(namespace, name, version, qualifiers, base_url='https://repo1.maven.org/maven2'):
+def get_urls(namespace, name, version, qualifiers, base_url='https://repo1.maven.org/maven2', **kwargs):
     """
     Return a mapping of URLs.
     """
@@ -1105,7 +1105,30 @@ def parse(
     Yield Packagedata objects from parsing a Maven pom file at `location` or
     using the provided `text` (one or the other but not both).
     """
-    pom = get_maven_pom(location=location)
+    package = _parse(
+        datasource_id=datasource_id,
+        package_type=package_type,
+        primary_language=primary_language,
+        location=location,
+        base_url=base_url
+    )
+    if package:
+        yield package
+
+
+def _parse(
+    datasource_id,
+    package_type,
+    primary_language,
+    location=None,
+    text=None,
+    base_url='https://repo1.maven.org/maven2',
+):
+    """
+    Yield Packagedata objects from parsing a Maven pom file at `location` or
+    using the provided `text` (one or the other but not both).
+    """
+    pom = get_maven_pom(location=location, text=text)
 
     if not pom:
         return
@@ -1192,7 +1215,8 @@ def parse(
     if not package_data.license_expression and package_data.declared_license:
         package_data.license_expression = models.compute_normalized_license(package_data.declared_license)
 
-    yield package_data
+    return package_data
+
 
 def build_vcs_and_code_view_urls(scm):
     """

src/packagedcode/models.py

@@ -307,7 +307,7 @@ def set_purl(self, package_url):
         for key, value in package_url.to_dict().items():
             self_val = getattr(self, key)
             if not self_val and value:
-                setattr(self, attr, value)
+                setattr(self, key, value)
 
     def to_dict(self, **kwargs):
         mapping = super().to_dict(**kwargs)

src/packagedcode/npm.py

@@ -146,7 +146,7 @@ def walk_npm(cls, resource, codebase, depth=0):
                     yield subchild
 
 
-def get_urls(namespace, name, version):
+def get_urls(namespace, name, version, **kwargs):
     return dict(
         repository_homepage_url=npm_homepage_url(namespace, name, registry='https://www.npmjs.com/package'),
         repository_download_url=npm_download_url(namespace, name, version, registry='https://registry.npmjs.org'),
@@ -163,13 +163,10 @@ class NpmPackageJsonHandler(BaseNpmHandler):
     documentation_url = 'https://docs.npmjs.com/cli/v8/configuring-npm/package-json'
 
     @classmethod
-    def parse(cls, location):
-        with io.open(location, encoding='utf-8') as loc:
-            package_data = json.load(loc)
-
-        name = package_data.get('name')
-        version = package_data.get('version')
-        homepage_url = package_data.get('homepage', '')
+    def _parse(cls, json_data):
+        name = json_data.get('name')
+        version = json_data.get('version')
+        homepage_url = json_data.get('homepage', '')
 
         # a package.json without name and version can be a private package
 
@@ -188,11 +185,11 @@ def parse(cls, location):
             namespace=namespace or None,
             name=name,
             version=version or None,
-            description=package_data.get('description', '').strip() or None,
+            description=json_data.get('description', '').strip() or None,
             homepage_url=homepage_url,
             **urls,
         )
-        vcs_revision = package_data.get('gitHead') or None
+        vcs_revision = json_data.get('gitHead') or None
 
         # mapping of top level package.json items to a function accepting as
         # arguments the package.json element value and returning an iterable of (key,
@@ -214,7 +211,7 @@ def parse(cls, location):
         ]
 
         for source, func in field_mappers:
-            value = package_data.get(source) or None
+            value = json_data.get(source) or None
             if value:
                 if isinstance(value, str):
                     value = value.strip()
@@ -226,14 +223,21 @@ def parse(cls, location):
             package.download_url = npm_download_url(package.namespace, package.name, package.version)
 
         # licenses are a tad special with many different data structures
-        lic = package_data.get('license')
-        lics = package_data.get('licenses')
+        lic = json_data.get('license')
+        lics = json_data.get('licenses')
         package = licenses_mapper(lic, lics, package)
 
         if not package.license_expression and package.declared_license:
             package.license_expression = compute_normalized_license(package.declared_license)
 
-        yield package
+        return package
+
+    @classmethod
+    def parse(cls, location):
+        with io.open(location, encoding='utf-8') as loc:
+            json_data = json.load(loc)
+
+        yield cls._parse(json_data)
 
     @classmethod
     def compute_normalized_license(cls, package):

src/packagedcode/nuget.py

@@ -18,7 +18,7 @@
 # TODO: add dependencies
 
 
-def get_urls(name, version):
+def get_urls(name, version, **kwargs):
     return dict(
         repository_homepage_url=f'https://www.nuget.org/packages/{name}/{version}',
         repository_download_url=f'https://www.nuget.org/api/v2/package/{name}/{version}',