Playwright (#69)

TGWolf · web-flow · commit b42a9ed1dc4c · 2024-12-12T15:21:28.000Z
* Playwight

* cleanup

* fix bandit issues
diff --git a/.github/workflows/cicd.yml b/.github/workflows/cicd.yml
@@ -202,6 +202,9 @@ jobs:
       - name: Install Pytest
         run: pip install pytest pytest-mock
 
+      - name: Setup PlayWright
+        run: playwright install && playwright install-deps
+
       - name: Run Pytest
         run: pytest --no-header -vv
 
diff --git a/.github/workflows/generate-release.yml b/.github/workflows/generate-release.yml
@@ -76,6 +76,9 @@ jobs:
       - name: Install Pytest
         run: pip install pytest pytest-mock
 
+      - name: Setup PlayWright
+        run: playwright install && playwright install-deps
+
       - name: Run Pytest
         run: pytest --no-header -vv
 
diff --git a/.github/workflows/generate-test-release.yml b/.github/workflows/generate-test-release.yml
@@ -75,6 +75,9 @@ jobs:
       - name: Install Pytest
         run: pip install pytest pytest-mock
 
+      - name: Setup PlayWright
+        run: playwright install && playwright install-deps
+
       - name: Run Pytest
         run: pytest --no-header -vv
 
diff --git a/CITATION.cff b/CITATION.cff
@@ -3,8 +3,8 @@ message: If you use this software, please cite it using these metadata.
 title: PyPi Extractor
 abstract: Extract package information for a given user in PyPi.
 type: software
-version: 0.1.2
-date-released: 2024-06-26
+version: 0.1.3
+date-released: 2024-12-12
 repository-code: https://github.com/DevelopersToolbox/pypi-extractor-package
 keywords:
   - "Wolf Software"
diff --git a/README.md b/README.md
@@ -44,6 +44,22 @@ PyPI Extractor is a Python package designed to fetch and process detailed inform
 Python Package Index (PyPI). This package is particularly useful for users who want to retrieve and analyze metadata for packages
 maintained by a specific PyPI user.
 
+## Significant Update From 0.1.3
+
+pypi.org no longer allow you to scrap details using the requests package, or any package that does not support JavaScript. To resolve this we have
+updated this package to utilise [PlayWright](https://pypi.org/project/playwright/) when retrieving a list of packages for a given user. While we have 
+attempted to automate as much as possible you might want to do some of the work manually.
+
+Playwright needs two commands to be run in order for it to function correctly:
+
+```
+playwright install
+playwright install-deps
+```
+
+We have added an `auto_install` option to the main class so that you can instruct the package to do the install for you, this helps when installing the
+package in a fully automated way, e.g. Puppet or similar.
+
 ## Features
 
 - Retrieve a list of packages maintained by a specific PyPI user.
@@ -116,11 +132,13 @@ print(package_details)
 
 A class to fetch and process package details for a given PyPI user.
 
-##### `__init__(self, username: str)`
+##### `__init__(self, username: str, verbose: bool, auto_install: bool)`
 
 - Initializes the `PyPiExtractor` with a username.
 - Parameters:
   - `username` (str): The PyPI username.
+  - `verbose` (bool): Verbose output (Default: False)
+  - `auto_install` (bool): Auto install PlayWright dependencies (Default: False)
 - Raises:
   - `PyPiExtractorError`: If the username is not provided.
 
@@ -132,6 +150,14 @@ A class to fetch and process package details for a given PyPI user.
 - Raises:
   - `PyPiExtractorError`: If the username is not provided.
 
+##### `enable_verbose(self)`
+
+- Enable verbose mode.
+
+##### `enable_auto_install(self)`
+
+- Enable auto install.
+
 ##### `get_user_packages(self) -> list`
 
 - Fetches the list of packages for the given PyPI user.
diff --git a/requirements.txt b/requirements.txt
@@ -1,2 +1,3 @@
 requests==2.32.3
 beautifulsoup4==4.12.3
+playwright==1.49.1
diff --git a/setup.py b/setup.py
@@ -12,7 +12,7 @@
 
 setup(
     name='wolfsoftware.pypi-extractor',
-    version='0.1.2',
+    version='0.1.3',
     author='Wolf Software',
     author_email='pypi@wolfsoftware.com',
     description='Extract package information for a given user in PyPi.',
diff --git a/tests/conftest.py b/tests/conftest.py
@@ -16,32 +16,56 @@
 import requests
 
 
+def raise_error(*args, **kwargs):
+    """Raise an error if the real playwright gets used."""
+    raise RuntimeError("Real Playwright should not be invoked!")
+
+
 @pytest.fixture
-def mock_get_user_packages_success() -> Generator[Union[MagicMock, AsyncMock], Any, None]:
-    """Fixture to mock requests.get for get_user_packages success case."""
-    with patch('requests.get') as mock_get:
-        mock_response = Mock()
-        mock_response.raise_for_status.return_value = None
-        mock_response.text = '''
-        <a class="package-snippet">
-            <h3 class="package-snippet__title">Package1</h3>
-            <p class="package-snippet__description">Description1</p>
-        </a>
-        <a class="package-snippet">
-            <h3 class="package-snippet__title">Package2</h3>
-            <p class="package-snippet__description">Description2</p>
-        </a>
-        '''
-        mock_get.return_value = mock_response
-        yield mock_get
+def mock_playwright() -> Generator[MagicMock, None, None]:
+    """Mock the Playwright sync API."""
+    with patch('wolfsoftware.pypi_extractor.pypi.sync_playwright') as mock_sync_playwright:
+        mock_playwright_instance = MagicMock()
+        mock_browser = MagicMock()
+        mock_context = MagicMock()
+        mock_page = MagicMock()
+
+        # Mock page.goto() and page.wait_for_selector()
+        mock_page.goto.return_value = None
+        mock_page.wait_for_selector.return_value = None
+
+        # Mock page.query_selector_all() to return simulated package elements
+        def mock_query_selector_all(selector):
+            """Handle mocking the right data."""
+            if selector == 'a.package-snippet':
+                return [
+                    MagicMock(query_selector=MagicMock(side_effect=[
+                        MagicMock(inner_text=MagicMock(return_value="Package1")),
+                        MagicMock(inner_text=MagicMock(return_value="Description1")),
+                    ])),
+                    MagicMock(query_selector=MagicMock(side_effect=[
+                        MagicMock(inner_text=MagicMock(return_value="Package2")),
+                        MagicMock(inner_text=MagicMock(return_value="Description2")),
+                    ])),
+                ]
+            return []
+        mock_page.query_selector_all.side_effect = mock_query_selector_all
+
+        mock_context.new_page.return_value = mock_page
+        mock_browser.new_context.return_value = mock_context
+        mock_playwright_instance.chromium.launch.return_value = mock_browser
+        mock_sync_playwright.return_value.__enter__.return_value = mock_playwright_instance
+        yield mock_sync_playwright
 
 
 @pytest.fixture
-def mock_get_user_packages_error() -> Generator[Union[MagicMock, AsyncMock], Any, None]:
-    """Fixture to mock requests.get for get_user_packages error case."""
-    with patch('requests.get') as mock_get:
-        mock_get.side_effect = requests.RequestException("Request error")
-        yield mock_get
+def mock_playwright_error() -> Generator[MagicMock, None, None]:
+    """Fixture to mock Playwright with an error scenario."""
+    with patch('wolfsoftware.pypi_extractor.pypi.sync_playwright') as mock_sync_playwright:
+        mock_playwright_instance = MagicMock()
+        mock_playwright_instance.chromium.launch.side_effect = Exception("Playwright error")
+        mock_sync_playwright.return_value.__enter__.return_value = mock_playwright_instance
+        yield mock_sync_playwright
 
 
 @pytest.fixture
@@ -155,24 +179,13 @@ def mock_get_package_details_error() -> Generator[Union[MagicMock, AsyncMock], A
 
 
 @pytest.fixture
-def mock_get_all_packages_details_success() -> Generator[Union[MagicMock, AsyncMock], Any, None]:
-    """Fixture to mock requests.get for get_all_packages_details success case."""
+def mock_get_all_packages_details_success() -> Generator[MagicMock, None, None]:
+    """Mock requests.get for get_all_packages_details success case."""
     with patch('requests.get') as mock_get:
-        mock_response_user = Mock()
+        # Mock response for the user packages API
+        mock_response_user = MagicMock()
         mock_response_user.raise_for_status.return_value = None
-        mock_response_user.text = '''
-        <a class="package-snippet">
-            <h3 class="package-snippet__title">Package1</h3>
-            <p class="package-snippet__description">Description1</p>
-        </a>
-        <a class="package-snippet">
-            <h3 class="package-snippet__title">Package2</h3>
-            <p class="package-snippet__description">Description2</p>
-        </a>
-        '''
-        mock_response_package1 = Mock()
-        mock_response_package1.raise_for_status.return_value = None
-        mock_response_package1.json.return_value = {
+        mock_response_user.json.return_value = {
             'info': {
                 'name': 'Package1',
                 'version': '1.0.0',
@@ -186,37 +199,30 @@ def mock_get_all_packages_details_success() -> Generator[Union[MagicMock, AsyncM
                 'requires_python': '>=3.6',
             },
             'releases': {
-                '0.9.0': [
-                    {
-                        'upload_time': '2021-01-01T00:00:00',
-                        'upload_time_iso_8601': '2021-01-01T00:00:00Z',
-                        'python_version': 'py3',
-                        'url': 'https://example.com',
-                        'filename': 'package-0.9.0.tar.gz',
-                        'packagetype': 'sdist',
-                        'md5_digest': 'abc123',
-                        'digests': {'sha256': 'def456'},
-                        'size': 12345
-                    }
-                ],
                 '1.0.0': [
                     {
                         'upload_time': '2021-06-01T00:00:00',
                         'upload_time_iso_8601': '2021-06-01T00:00:00Z',
                         'python_version': 'py3',
-                        'url': 'https://example.com',
+                        'url': 'https://example.com/package-1.0.0.tar.gz',
                         'filename': 'package-1.0.0.tar.gz',
                         'packagetype': 'sdist',
-                        'md5_digest': 'ghi789',
-                        'digests': {'sha256': 'jkl012'},
-                        'size': 23456
+                        'md5_digest': 'abc123',
+                        'digests': {'sha256': 'def456'},
+                        'size': 12345
                     }
-                ],
+                ]
             },
             'requires_dist': ['requests', 'beautifulsoup4'],
             'urls': [{'url': 'https://example.com/package-1.0.0.tar.gz'}],
         }
-        mock_response_package2 = Mock()
+
+        # Simulate two different package details responses
+        mock_response_package1 = MagicMock()
+        mock_response_package1.raise_for_status.return_value = None
+        mock_response_package1.json.return_value = mock_response_user.json.return_value
+
+        mock_response_package2 = MagicMock()
         mock_response_package2.raise_for_status.return_value = None
         mock_response_package2.json.return_value = {
             'info': {
@@ -226,41 +232,30 @@ def mock_get_all_packages_details_success() -> Generator[Union[MagicMock, AsyncM
                 'author': 'Author2',
                 'author_email': 'author2@example.com',
                 'license': 'MIT',
-                'home_page': 'https://example.com',
-                'keywords': 'example, package',
+                'home_page': 'https://example.com/package2',
+                'keywords': 'example, package2',
                 'classifiers': ['Development Status :: 5 - Production/Stable'],
                 'requires_python': '>=3.6',
             },
             'releases': {
-                '1.0.0': [
-                    {
-                        'upload_time': '2021-01-01T00:00:00',
-                        'upload_time_iso_8601': '2021-01-01T00:00:00Z',
-                        'python_version': 'py3',
-                        'url': 'https://example.com',
-                        'filename': 'package-1.0.0.tar.gz',
-                        'packagetype': 'sdist',
-                        'md5_digest': 'abc123',
-                        'digests': {'sha256': 'def456'},
-                        'size': 12345
-                    }
-                ],
                 '2.0.0': [
                     {
-                        'upload_time': '2021-06-01T00:00:00',
-                        'upload_time_iso_8601': '2021-06-01T00:00:00Z',
+                        'upload_time': '2022-06-01T00:00:00',
+                        'upload_time_iso_8601': '2022-06-01T00:00:00Z',
                         'python_version': 'py3',
-                        'url': 'https://example.com',
+                        'url': 'https://example.com/package-2.0.0.tar.gz',
                         'filename': 'package-2.0.0.tar.gz',
                         'packagetype': 'sdist',
                         'md5_digest': 'ghi789',
                         'digests': {'sha256': 'jkl012'},
                         'size': 23456
                     }
-                ],
+                ]
             },
             'requires_dist': ['requests', 'beautifulsoup4'],
             'urls': [{'url': 'https://example.com/package-2.0.0.tar.gz'}],
         }
-        mock_get.side_effect = [mock_response_user, mock_response_package1, mock_response_package2]
+
+        # Simulate the sequence of requests
+        mock_get.side_effect = [mock_response_package1, mock_response_package2]
         yield mock_get
diff --git a/tests/test_pypi_extractor.py b/tests/test_pypi_extractor.py
diff --git a/wolfsoftware/pypi_extractor/pypi.py b/wolfsoftware/pypi_extractor/pypi.py

Original file line number	Diff line number	Diff line change
`@@ -1,2 +1,3 @@`
`1`	`1`	`requests==2.32.3`
`2`	`2`	`beautifulsoup4==4.12.3`
	`3`	`+playwright==1.49.1`