Fix docstests and misc

Signed-off-by: Ayan Sinha Mahapatra <ayansmahapatra@gmail.com>

Author: AyanSinhaMahapatra

src/packagedcode/go_mod.py

@@ -111,8 +111,7 @@ def parse_gomod(location):
     For example::
 
         >>> p = parse_dep_link('github.com/davecgh/go-spew v1.1.1')
-        >>> assert p.group('namespace') == ('github.com/davecgh')
-        >>> assert p.group('name') == ('go-spew')
+        >>> assert p.group('ns_name') == ('github.com/davecgh/go-spew')
         >>> assert p.group('version') == ('v1.1.1')
     """
     with io.open(location, encoding='utf-8', closefd=True) as data:

src/packagedcode/maven.py

@@ -1193,7 +1193,9 @@ def build_vcs_and_code_view_urls(scm):
     ...     connection='scm:git:git@github.com:histogrammar/histogrammar-scala.git',
     ...     tag='HEAD',
     ...     url='https://github.com/histogrammar/histogrammar-scala')
-    >>> expected = {'vcs_url': 'sdfasdasd', 'code_view_url': 'asdasda'}
+    >>> expected = {
+    ...     'vcs_url': 'git+https://github.com/histogrammar/histogrammar-scala.git',
+    ...     'code_view_url': 'https://github.com/histogrammar/histogrammar-scala'}
     >>> assert build_vcs_and_code_view_urls(scm) == expected
     """
 

src/packagedcode/models.py

@@ -158,7 +158,7 @@ def to_tuple(collection):
     Return a tuple of basic Python values by recursively converting a mapping
     and all its sub-mappings.
     For example::
-    >>> as_tuple({7: [1,2,3], 9: {1: [2,6,8]}})
+    >>> to_tuple({7: [1,2,3], 9: {1: [2,6,8]}})
     ((7, (1, 2, 3)), (9, ((1, (2, 6, 8)),)))
     """
     if isinstance(collection, dict):
@@ -820,7 +820,7 @@ class DatafileHandler:
     documentation_url = None
 
     @classmethod
-    def is_datafile(cls, location, filetypes=tuple()):
+    def is_datafile(cls, location, filetypes=tuple(), allow_filename=False):
         """
         Return True if the file at ``location`` is likely a package data file
         that this parser can handle. This implementation is based on:
@@ -834,7 +834,7 @@ def is_datafile(cls, location, filetypes=tuple()):
 
         Subclasses can override to implement more complex data file recognition.
         """
-        if filetype.is_file(location):
+        if filetype.is_file(location) or allow_filename:
             loc = as_posixpath(location)
             if any(fnmatchcase(loc, pat) for pat in cls.path_patterns):
                 filetypes = filetypes or cls.filetypes

src/packagedcode/npm.py

@@ -719,7 +719,7 @@ def npm_homepage_url(namespace, name, registry='https://www.npmjs.com/package'):
     >>> assert npm_homepage_url('', 'angular') == expected
 
     >>> expected = 'https://yarnpkg.com/en/package/angular'
-    >>> assert npm_homepage_url('', 'angular', 'https://yarnpkg.com/en/package/') == expected
+    >>> assert npm_homepage_url('', 'angular', 'https://yarnpkg.com/en/package') == expected
 
     >>> expected = 'https://yarnpkg.com/en/package/@ang/angular'
     >>> assert npm_homepage_url('@ang', 'angular', 'https://yarnpkg.com/en/package') == expected

src/packagedcode/opam.py

@@ -181,51 +181,6 @@ def parse_opam_from_text(text):
     """
     Return a mapping of package data collected from the opam OCaml package
     manifest ``text``.
-
-    Example:
-
-    >>> opam_text = '''
-    ... opam-version: "2.0"
-    ... version: "4.11.0+trunk"
-    ... synopsis: "OCaml development version"
-    ... depends: [
-    ...   "ocaml" {= "4.11.0" & post}
-    ...   "base-unix" {post}
-    ... ]
-    ... conflict-class: "ocaml-core-compiler"
-    ... flags: compiler
-    ... setenv: CAML_LD_LIBRARY_PATH = "%{lib}%/stublibs"
-    ... build: [
-    ...   ["./configure" "--prefix=%{prefix}%"]
-    ...   [make "-j%{jobs}%"]
-    ... ]
-    ... install: [make "install"]
-    ... maintainer: "caml-list@inria.fr"
-    ... homepage: "https://github.com/ocaml/ocaml/"
-    ... bug-reports: "https://github.com/ocaml/ocaml/issues"
-    ... authors: [
-    ...   "Xavier Leroy"
-    ...   "Damien Doligez"
-    ...   "Alain Frisch"
-    ...   "Jacques Garrigue"
-    ... ]'''
-
-    >>> p = parse_opam_from_text(opam_text)
-    >>> for k, v in p.items():
-    >>>     print(k, v)
-    opam-version 2.0
-    version 4.11.0+trunk
-    synopsis OCaml development version
-    depends [Opam(name='ocaml', version='= 4.11.0 & post'), Opam(name='base-unix', version='post')]
-    conflict-class ocaml-core-compiler
-    flags compiler
-    setenv CAML_LD_LIBRARY_PATH = %{lib}%/stublibs
-    build
-    install make install
-    maintainer ['caml-list@inria.fr']
-    homepage https://github.com/ocaml/ocaml/
-    bug-reports https://github.com/ocaml/ocaml/issues
-    authors ['Xavier Leroy', 'Damien Doligez', 'Alain Frisch', 'Jacques Garrigue']
     """
 
     opam_data = {}

src/packagedcode/pypi.py

@@ -597,26 +597,26 @@ class PipRequirementsFileHandler(BaseDependencyFileHandler):
     A pip requirements (or constraints) file.
 
     Some example::
-    >>> PipRequirementsFileHandler.is_datafile('dev-requirements.txt')
+    >>> PipRequirementsFileHandler.is_datafile('dev-requirements.txt', allow_filename=True)
     True
-    >>> PipRequirementsFileHandler.is_datafile('requirements.txt')
+    >>> PipRequirementsFileHandler.is_datafile('requirements.txt', allow_filename=True)
     True
-    >>> PipRequirementsFileHandler.is_datafile('requirement.txt')
+    >>> PipRequirementsFileHandler.is_datafile('requirement.txt', allow_filename=True)
     True
-    >>> PipRequirementsFileHandler.is_datafile('requirements.in')
+    >>> PipRequirementsFileHandler.is_datafile('requirements.in', allow_filename=True)
     True
-    >>> PipRequirementsFileHandler.is_datafile('requirements.pip')
+    >>> PipRequirementsFileHandler.is_datafile('requirements.pip', allow_filename=True)
     True
-    >>> PipRequirementsFileHandler.is_datafile('requirements-dev.txt')
+    >>> PipRequirementsFileHandler.is_datafile('requirements-dev.txt', allow_filename=True)
     True
-    >>> PipRequirementsFileHandler.is_datafile('some-requirements-dev.txt')
+    >>> PipRequirementsFileHandler.is_datafile('some-requirements-dev.txt', allow_filename=True)
     True
-    >>> PipRequirementsFileHandler.is_datafile('requires.txt')
+    >>> PipRequirementsFileHandler.is_datafile('requires.txt', allow_filename=True)
     True
-    >>> PipRequirementsFileHandler.is_datafile('requirements/base.txt')
+    >>> PipRequirementsFileHandler.is_datafile('requirements/base.txt', allow_filename=True)
+    True
+    >>> PipRequirementsFileHandler.is_datafile('reqs.txt', allow_filename=True)
     True
-    >>> PipRequirementsFileHandler.is_datafile('reqs.txt')
-    False
     """
     datasource_id = 'pip_requirements'
 

src/packagedcode/rubygems.py

@@ -398,8 +398,8 @@ def rubygems_api_url(name, version=None):
     >>> assert url == 'https://rubygems.org/api/v2/rubygems/turbolinks/versions/1.0.2.json'
 
     If no version, we return:
-    >>> url = rubygems_api_url(name='turbolinks'')
-    >>> assert url == 'https://rubygems.org/api/v2/rubygems/turbolinks.json'
+    >>> url = rubygems_api_url(name='turbolinks')
+    >>> assert url == 'https://rubygems.org/api/v1/versions/turbolinks.json'
 
     Things we could return: a summary for the latest version, with deps
     https://rubygems.org/api/v1/gems/mqlight.json

tests/packagedcode/data/debian/end-to-end.tgz.expected.json

@@ -14,7 +14,6 @@
       "warnings": [],
       "extra_data": {
         "spdx_license_list_version": "3.16",
-        "OUTDATED": "WARNING: Outdated ScanCode Toolkit version! You are using an outdated version of ScanCode Toolkit: 31.0.0b1 released on: 2021-09-24. A new version 42.5.1 is available with important improvements including bug and security fixes, updated license, copyright and package detection, and improved scanning accuracy. Please download and install the latest version of ScanCode. Visit https://github.com/nexB/scancode-toolkit/releases for details.",
         "files_count": 6
       }
     }