choose_ci_set.py

import os
import re
import sys
from subprocess import check_output

import pytest

SCRIPT_DIR = os.path.dirname(os.path.abspath(__file__))
CI_MARK = "@pytest.mark.ci"
HEADED_MARK = "@pytest.mark.headed"
OUTPUT_FILE = "selected_tests"


class CollectionPlugin:
    """Mini plugin to get test names"""

    def __init__(self):
        self.tests = []

    def pytest_report_collectionfinish(self, items):
        self.tests = [item.nodeid for item in items]


def snakify(pascal: str) -> str:
    """Convert PascalCase to snake_case"""
    chars = pascal[0].lower()
    for c in pascal[1:]:
        if c == c.upper():
            chars = chars + f"_{c.lower()}"
        else:
            chars = chars + c
    return chars


def pascalify(snake: str) -> str:
    """Convert snake_case to PascalCase"""
    chars = snake[0].upper()
    up_flag = False
    for c in snake[1:]:
        if up_flag:
            chars = chars + c.upper()
            up_flag = False
        elif c == "_":
            up_flag = True
        else:
            chars = chars + c
    return chars


def localify(path: str) -> str:
    """Remove the script dir from an item"""
    return path.replace(SCRIPT_DIR, ".")


def get_tests_by_model(
    model_name: str, test_paths_and_contents: dict, run_list: list
) -> list:
    """
    Given a model name, a dict of paths and their file contents, and the
    list of existing tests/dirs to check, return matches by local paths
    """
    matching_tests = []
    for path, test in test_paths_and_contents.items():
        localpath = localify(path)
        in_run_list = False
        for entry in run_list:
            if entry in localpath:
                in_run_list = True
        if not in_run_list and model_name in test:
            matching_tests.append(localpath)

    return matching_tests


def dedupe(run_list: list, slash: str) -> list:
    """For a run list, remove entries that are covered by more general entries."""
    run_list = list(set(run_list))
    dotslashes = []
    removes = []

    for i, entry in enumerate(run_list):
        if (
            not entry.startswith(".")
            and not entry.startswith("\\")
            and not entry.startswith("/")
        ):
            dotslashes.append(i)

    for dotslash in dotslashes:
        run_list[dotslash] = f".{slash}{run_list[dotslash]}"

    for i, entry_a in enumerate(run_list):
        for j, entry_b in enumerate(run_list):
            if i == j:
                continue
            candidate = max((i, j))
            if entry_a in entry_b and candidate not in removes:
                removes.append(candidate)

    removes.sort(reverse=True)
    for remove in removes:
        del run_list[remove]

    return run_list


if __name__ == "__main__":
    if os.path.exists(".env"):
        with open(".env") as fh:
            contents = fh.read()
            if "TESTRAIL_REPORT='true'" in contents:
                os.environ["TESTRAIL_REPORT"] = "true"
            if "RUN_ALL='true'" in contents:
                os.environ["MANUAL"] = "true"

    if os.environ.get("TESTRAIL_REPORT") or os.environ.get("MANUAL"):
        # Run all tests if this is a scheduled beta or a manual run
        with open(OUTPUT_FILE, "w") as fh:
            fh.write("tests")
            sys.exit(0)

    slash = "/" if "/" in SCRIPT_DIR else "\\"

    re_obj = {
        "test_re_string": r".*/.*/test_.*\.py",
        "suite_conftest_re_string": r".*/.*/conftest\.py",
        "selectors_json_re_string": r"modules/data/.*\.components\.json",
        "object_model_re_string": r"modules/.*.object.*\.py",
        "class_re_string": r"\s*class (\w+)[(A-Za-z0-9_)]*:",
    }
    for k in list(re_obj.keys()):
        if slash == "\\":
            re_obj[k] = re_obj.get(k).replace("/", r"\\")
        short_name = "_".join(k.split("_")[:-1])
        re_obj[short_name] = re.compile(re_obj.get(k))

    run_list = []
    check_output(["git", "fetch", "--quiet", "--depth=1", "origin", "main"])
    committed_files = (
        check_output(["git", "--no-pager", "diff", "--name-only", "origin/main"])
        .decode()
        .replace("/", slash)
        .splitlines()
    )

    main_conftest = "conftest.py"
    base_page = os.path.join("modules", "page_base.py")

    if main_conftest in committed_files or base_page in committed_files:
        # Run all the tests (no files as arguments) if main conftest or basepage changed
        with open(OUTPUT_FILE, "w") as fh:
            fh.write("tests")
        sys.exit(0)

    all_tests = []
    test_paths_and_contents = {}
    for root, _, files in os.walk(os.path.join(SCRIPT_DIR, "tests")):
        for f in files:
            this_file = os.path.join(root, f)
            if re_obj.get("test_re").search(this_file) and "__pycache" not in this_file:
                all_tests.append(os.path.join(this_file))
                with open(this_file, encoding="utf-8") as fh:
                    lines = fh.readlines()
                    test_paths_and_contents[this_file] = "".join(lines)

    p = CollectionPlugin()
    pytest.main(["--collect-only", "-m", "ci", "-s"], plugins=[p])
    ci_paths = [f".{slash}{test}" for test in p.tests]

    # Dedupe just in case
    ci_paths = list(set(ci_paths))

    changed_suite_conftests = [
        f for f in committed_files if re_obj.get("suite_conftest_re").match(f)
    ]
    changed_selectors = [
        f for f in committed_files if re_obj.get("selectors_json_re").match(f)
    ]
    changed_models = [
        f for f in committed_files if re_obj.get("object_model_re").match(f)
    ]
    changed_tests = [f for f in committed_files if re_obj.get("test_re").match(f)]

    if changed_suite_conftests:
        run_list = [
            "." + slash + os.path.join(*suite.split(slash)[-3:-1])
            for suite in changed_suite_conftests
        ]

    if changed_selectors:
        for selector_file in changed_selectors:
            (_, filename) = os.path.split(selector_file)
            model_name = pascalify(filename.split(".")[0])
            for test_name in get_tests_by_model(
                model_name, test_paths_and_contents, run_list
            ):
                run_list.append(test_name)

    if changed_models:
        for model_file in changed_models:
            model_file_contents = "".join([line for line in open(model_file)])
            classes = re_obj.get("class_re").findall(model_file_contents)
            for model_name in classes:
                for test_name in get_tests_by_model(
                    model_name, test_paths_and_contents, run_list
                ):
                    run_list.append(test_name)

    if changed_tests:
        for changed_test in changed_tests:
            found = False
            for file in run_list:
                # Don't add if already exists in suite changes
                pieces = file.split(slash)
                if len(pieces) == 3 and pieces[-1] in changed_test:
                    found = True

                # Don't add if already in list
                if file in changed_test:
                    found = True
            if not found:
                run_list.append(changed_test)

    if not run_list:
        with open(OUTPUT_FILE, "w") as fh:
            fh.write("\n".join(ci_paths))
    else:
        run_list.extend(ci_paths)

        # Dedupe just in case
        run_list = dedupe(run_list, slash)
        run_list = [entry for entry in run_list if os.path.exists(entry.split("::")[0])]
        with open(OUTPUT_FILE, "w") as fh:
            fh.write("\n".join(run_list))