tasks: set up harness for installing packages, itest for caching ChromeDriver (#7)

Author: stanislaw

.gitmodules

@@ -1,3 +1,3 @@
 [submodule "submodules/html2pdf"]
 	path = submodules/html2pdf
-	url = git@github.com:mettta/html2pdf.git
+	url = https://github.com/mettta/html2pdf.git

developer/pip_install_hpdf_deps.py

@@ -0,0 +1,101 @@
+import importlib.metadata as importlib_metadata
+import os
+import subprocess
+import sys
+import tempfile
+
+import toml
+from packaging.requirements import Requirement
+
+
+class PackageNotFound(Exception):
+    pass
+
+
+class PackageVersionConflict(Exception):
+    pass
+
+
+# A simplified version inspired by:
+# https://github.com/HansBug/hbutils/blob/37879186c489bced2791309c43d131f1703b7bd4/hbutils/system/python/package.py#L171
+def check_if_package_installed(package_name: str):
+    requirement: Requirement = Requirement(package_name)
+    try:
+        version = importlib_metadata.distribution(requirement.name).version
+    except importlib_metadata.PackageNotFoundError:
+        raise PackageNotFound(requirement) from None
+    if not requirement.specifier.contains(version):
+        raise PackageVersionConflict(version)
+
+
+print(  # noqa: T201
+    "pip_install_hpdf_deps.py: "
+    "checking if the current Python environment has all packages installed"
+    ".",
+    flush=True,
+)
+
+pyproject_content = toml.load("pyproject.toml")
+
+
+# The development dependencies are ignored, because they are managed in tox.ini.
+dependencies = pyproject_content["project"]["dependencies"]
+
+needs_installation = False
+
+for dependency in dependencies:
+    try:
+        check_if_package_installed(dependency)
+    except PackageNotFound:
+        print(  # noqa: T201
+            f"pip_install_hpdf_deps.py: "
+            f"Package is not installed: '{dependency}'.",
+            flush=True,
+        )
+        needs_installation = True
+        break
+    except PackageVersionConflict as exception_:
+        print(  # noqa: T201
+            (
+                f"pip_install_hpdf_deps.py: version conflict between "
+                f"hpdf's requirement '{dependency}' "
+                f"and the already installed package: "
+                f"{exception_.args[0]}."
+            ),
+            flush=True,
+        )
+        needs_installation = True
+        break
+
+if not needs_installation:
+    print(  # noqa: T201
+        "pip_install_hpdf_deps.py: all packages seem to be installed.",
+        flush=True,
+    )
+    sys.exit(0)
+
+print(  # noqa: T201
+    "pip_install_hpdf_deps.py: will install packages.", flush=True
+)
+
+all_packages = "\n".join(dependencies) + "\n"
+
+with tempfile.TemporaryDirectory() as tmp_dir:
+    with open(
+        os.path.join(tmp_dir, "requirements.txt"), "w", encoding="utf8"
+    ) as tmp_requirements_txt_file:
+        tmp_requirements_txt_file.write(all_packages)
+
+    command = [
+        sys.executable,
+        "-m",
+        "pip",
+        "install",
+        "-r",
+        tmp_requirements_txt_file.name,
+    ]
+
+    result = subprocess.run(command, check=True, encoding="utf8")
+    print(  # noqa: T201
+        f"'pip install' command exited with: {result.returncode}", flush=True
+    )

hpdf/hpdf.py

@@ -75,6 +75,7 @@ def find_driver(self, driver: Driver):
         browser_version = self._os_system_manager.get_browser_version_from_os(
             browser_type
         )
+        assert browser_version is not None, browser_version
 
         path_to_cached_chrome_driver_dir = os.path.join(
             path_to_cached_chrome_driver_dir, browser_version, os_type
@@ -84,18 +85,18 @@ def find_driver(self, driver: Driver):
         )
         if os.path.isfile(path_to_cached_chrome_driver):
             print(  # noqa: T201
-                f"html2pdf: chromedriver exists in StrictDoc's local cache: "
+                f"html2pdf: ChromeDriver exists in the local cache: "
                 f"{path_to_cached_chrome_driver}"
             )
             return path_to_cached_chrome_driver
         print(  # noqa: T201
-            f"html2pdf: chromedriver does not exist in StrictDoc's local cache: "
+            f"html2pdf: ChromeDriver does not exist in the local cache: "
             f"{path_to_cached_chrome_driver}"
         )
         path_to_downloaded_chrome_driver = super().find_driver(driver)
         if path_to_downloaded_chrome_driver is None:
             print(  # noqa: T201
-                f"html2pdf: could not get a downloaded Chrome driver: "
+                f"html2pdf: could not get a downloaded ChromeDriver: "
                 f"{path_to_cached_chrome_driver}"
             )
             return None
@@ -241,7 +242,7 @@ def main():
     parser.add_argument(
         "--cache-dir",
         type=str,
-        help="Optional path to a cache directory whereto the Chrome driver is downloaded.",
+        help="Optional path to a cache directory whereto the ChromeDriver is downloaded.",
     )
     parser.add_argument("paths", nargs='+', help="Paths to input HTML file.")
     args = parser.parse_args()
@@ -253,7 +254,7 @@ def main():
         if args.cache_dir is not None
         else (
             os.path.join(
-                tempfile.gettempdir(), "strictdoc_cache", "chromedriver"
+                Path.home(), ".hpdf", "chromedriver"
             )
         )
     )

requirements.txt renamed to requirements.development.txt

@@ -1,4 +1,7 @@
 invoke
+toml
+packaging
+setuptools
 
 #
 # Integration tests

tasks.py

@@ -54,7 +54,8 @@ def list_tasks(context):
 
 @task
 def bootstrap(context):
-    run_invoke(context, "pip install -r requirements.txt")
+    run_invoke(context, "git submodule update --init --recursive")
+    run_invoke(context, "pip install -r requirements.development.txt")
 
 
 @task
@@ -71,6 +72,11 @@ def format_readme(context):
     """)
 
 
+@task
+def lint(context):
+    pass
+
+
 @task
 def test_integration(
     context,
@@ -86,8 +92,12 @@ def test_integration(
     focus_or_none = f"--filter {focus}" if focus else ""
     debug_opts = "-vv --show-all" if debug else ""
 
+    # For now, the --threads are set to 1 because running tests parallelized
+    # will result in race conditions between Web Driver Manager downloading
+    # ChromeDriver.
     itest_command = f"""
         lit
+        --threads 1
         --param HTML2PDF_EXEC="{html2pdf_exec}"
         -v
         {debug_opts}

tests/integration/03_cache_dir_argument/index.html

@@ -0,0 +1,15 @@
+<!doctype html>
+<html lang="">
+
+<head>
+  <meta charset="utf-8">
+  <title>Test page</title>
+  <script src="../../../submodules/html2pdf/dist/bundle.js"></script>
+  <link rel="stylesheet" href="../../../submodules/html2pdf/test/shared/css/main.css">
+</head>
+
+<body>
+  <p>Hello world!</p>
+</body>
+
+</html>

tests/integration/03_cache_dir_argument/test.itest

@@ -0,0 +1,11 @@
+RUN: %html2pdf --cache-dir %S/Output/cache %S/index.html:%S/Output/index.pdf | filecheck %s --dump-input=fail --check-prefix CHECK-RUN1
+RUN: %check_exists --file "%S/Output/index.pdf"
+RUN: python %S/test.py
+
+CHECK-RUN1: html2pdf: ChromeDriver does not exist in the local cache:
+
+RUN: %html2pdf --cache-dir %S/Output/cache %S/index.html:%S/Output/index.pdf | filecheck %s --dump-input=fail --check-prefix CHECK-RUN2
+RUN: %check_exists --file "%S/Output/index.pdf"
+RUN: python %S/test.py
+
+CHECK-RUN2: html2pdf: ChromeDriver exists in the local cache:

tests/integration/03_cache_dir_argument/test.py

@@ -0,0 +1,7 @@
+from pypdf import PdfReader
+
+reader = PdfReader("Output/index.pdf")
+
+assert len(reader.pages) == 1
+
+assert reader.pages[0].extract_text() == "Hello world!"