testbed/aws-cloudformation__cfn-lint/src/cfnlint/schema/manager.py

"""
Copyright 2019 Amazon.com, Inc. or its affiliates. All Rights Reserved.
SPDX-License-Identifier: MIT-0
"""

from __future__ import annotations

import filecmp
import fnmatch
import json
import logging
import multiprocessing
import os
import re
import shutil
import sys
import zipfile
from copy import copy
from functools import lru_cache
from typing import Any, Dict, Iterator, Sequence

import jsonpatch
import jsonpointer

from cfnlint.helpers import (
    REGION_PRIMARY,
    REGIONS,
    ToPy,
    get_url_retrieve,
    load_resource,
    url_has_newer_version,
)
from cfnlint.schema._exceptions import ResourceNotFoundError
from cfnlint.schema._getatts import AttributeDict
from cfnlint.schema._schema import Schema
from cfnlint.schema.patch import SchemaPatch

LOGGER = logging.getLogger(__name__)


class _FileLocation:
    def __init__(self, path: list[str]):
        self.path_relative = os.path.join(
            os.path.dirname(__file__),
            "..",
            *path,
        )
        self.module = ".".join(["cfnlint"] + path[:])
        self.path = path


class ProviderSchemaManager:
    def __init__(self) -> None:
        self._root = _FileLocation(
            [
                "data",
                "schemas",
                "providers",
            ]
        )
        self._patches = _FileLocation(
            [
                "data",
                "schemas",
                "patches",
            ]
        )
        self._region_primary = ToPy(REGION_PRIMARY)
        self._registry_schemas: dict[str, Schema] = {}
        self._provider_schema_modules: dict[str, Any] = {}
        self.reset()

    def reset(self) -> None:
        """
        Reset's the cache so specs can be reloaded.
        Important function when processing many templates
        and using spec patching
        """
        self._schemas: dict[str, dict[str, Schema]] = {}
        for region in REGIONS:
            self._schemas[region] = {}
        self._removed_types: list[str] = []
        self.get_resource_schema.cache_clear()
        self.get_resource_types.cache_clear()
        self.get_type_getatts.cache_clear()

    def load_registry_schemas(self, path: str) -> None:
        """Load extra registry schemas from a directory

        Args:
            path (str): the directory to load schema files from
        Returns:
            None: None
        """
        for dirpath, _, filenames in os.walk(path):
            for filename in fnmatch.filter(filenames, "*.json"):
                with open(os.path.join(dirpath, filename), "r", encoding="utf-8") as fh:
                    schema = Schema(json.load(fh))
                    self._registry_schemas[schema.type_name] = schema

    def get_resource_schemas_by_regions(
        self, resource_type: str, regions: Sequence[str]
    ) -> Iterator[tuple[list[str], Schema]]:
        """
        Get unique schemas with their associated regions

        Args:
            resource_type (str): the :: version of the resource type
            regions (Sequence[str]): the regions in which we want schems for
        Returns:
            Iterator[tuple[list[str], Schema]]: the unique schemas with
              their associated regions
        """
        cached_regions: list[str] = []
        cached_schema: Schema | None = None
        for region in regions:
            try:
                schema = self.get_resource_schema(region, resource_type)
            except ResourceNotFoundError:
                continue
            if not schema.is_cached and region != REGION_PRIMARY:
                yield [region], schema
            else:
                cached_regions.append(region)
                cached_schema = schema

        if cached_schema is not None:
            yield cached_regions, cached_schema

    def _normalize_resource_type(self, resource_type: str) -> str:
        """
        Normalize the resource type to the correct format

        Args:
            resource_type (str): the :: version of the resource type
        Returns:
            str: the normalized resource type
        """
        if resource_type.startswith("Custom::"):
            resource_type = "AWS::CloudFormation::CustomResource"
        if resource_type.endswith("::MODULE"):
            resource_type = "Module"

        return resource_type

    @lru_cache(maxsize=None)
    def get_resource_schema(self, region: str, resource_type: str) -> Schema:
        """Get the provider resource schema and cache it to speed up future lookups

        Args:
            region (str): the region in which do get the provider schema for
            resource_type (str): the :: version of the resource type
        Returns:
            dict: returns the schema
        """
        if resource_type not in self._registry_schemas:
            resource_type = self._normalize_resource_type(resource_type)

        if resource_type in self._removed_types:
            raise ResourceNotFoundError(resource_type, region)

        reg = ToPy(region)
        rt = ToPy(resource_type)

        schema = self._schemas[reg.name].get(rt.name)
        if schema is not None:
            return schema

        # dynamically import the modules as needed
        self._provider_schema_modules[reg.name] = __import__(
            f"{self._root.module}.{reg.py}", fromlist=[""]
        )
        # check cfn-lint provided schemas
        if rt.name in self._registry_schemas:
            self._schemas[reg.name][rt.name] = self._registry_schemas[rt.name]
            return self._schemas[reg.name][rt.name]

        # load the schema
        if f"{rt.provider}.json" in self._provider_schema_modules[reg.name].cached:
            schema_cached = copy(
                self.get_resource_schema(
                    region=self._region_primary.name,
                    resource_type=rt.name,
                )
            )
            schema_cached.is_cached = True
            self._schemas[reg.name][rt.name] = schema_cached
            return self._schemas[reg.name][rt.name]
        try:
            self._schemas[reg.name][rt.name] = Schema(
                load_resource(
                    self._provider_schema_modules[reg.name],
                    filename=f"{rt.provider}.json",
                )
            )
            return self._schemas[reg.name][rt.name]
        except Exception as e:
            raise ResourceNotFoundError(rt.name, region) from e

    @lru_cache(maxsize=None)
    def get_resource_types(self, region: str) -> list[str]:
        """Get the resource types for a region

        Args:
            region (str): the region in which to get the resource types for
        Returns:
            list[str]: returns a list of resource types
        """
        reg = ToPy(region)

        if self._region_primary.name not in self._provider_schema_modules:
            self._provider_schema_modules[self._region_primary.name] = __import__(
                f"{self._root.module}.{self._region_primary.py}", fromlist=[""]
            )
        resource_types: list[str] = []
        if reg.name not in self._provider_schema_modules:
            self._provider_schema_modules[region] = __import__(
                f"{self._root.module}.{reg.py}", fromlist=[""]
            )
        resource_types.extend(
            rt
            for rt in self._provider_schema_modules[reg.name].types
            if rt not in self._removed_types
        )
        resource_types.extend(list(self._registry_schemas.keys()))

        return resource_types

    def update(self, force: bool) -> None:
        """Update every regions provider schemas

        Args:
            force (bool): force the schemas to be downloaded
        Returns:
            None: returns when complete
        """
        self._update_provider_schema(self._region_primary.name, force=force)
        # pylint: disable=not-context-manager
        with multiprocessing.Pool() as pool:
            # Patch from registry schema
            provider_pool_tuple = [
                (k, force) for k in REGIONS if k != self._region_primary.name
            ]
            pool.starmap(self._update_provider_schema, provider_pool_tuple)

    def _remove_descriptions(self, spec: Any) -> Any:
        if isinstance(spec, dict):
            r: dict[Any, Any] = {}
            for k, v in spec.items():
                if k != "description":
                    r[k] = self._remove_descriptions(v)

            return r
        elif isinstance(spec, list):
            m: list[Any] = []
            for v in spec:
                m.append(self._remove_descriptions(v))

            return m
        else:
            return spec

    def _update_provider_schema(self, region: str, force: bool = False) -> None:
        """Update the provider schemas from the AWS websites

        Args:
            region (str): the region in which do ge the provider schema for
            force (bool): force the schemas to be downloaded
        Returns:
            None: returns when complete
        """
        # China regions in .com.cn
        suffix = ".cn" if region in ["cn-north-1", "cn-northwest-1"] else ""
        url = f"https://schema.cloudformation.{region}.amazonaws.com{suffix}/CloudformationSchema.zip"
        reg = ToPy(region)
        directory = os.path.join(f"{self._root.path_relative}/{reg.py}/")
        directory_pr = os.path.join(
            f"{self._root.path_relative}/{self._region_primary.py}/"
        )

        multiprocessing_logger = multiprocessing.log_to_stderr()

        multiprocessing_logger.debug("Downloading template %s into %s", url, directory)

        # Check to see if we already have the latest version, and if so stop
        if not (url_has_newer_version(url) or force):
            return

        if not os.path.exists(directory):
            os.mkdir(directory)

        try:
            filehandle = get_url_retrieve(url, caching=True)
            # clean folder
            shutil.rmtree(directory)
            with zipfile.ZipFile(filehandle, "r") as zip_ref:
                zip_ref.extractall(directory)

            filenames = [
                f
                for f in os.listdir(directory)
                if os.path.isfile(os.path.join(directory, f)) and f != "__init__.py"
            ]
            # There is no schema for CDK but its an allowable type
            all_types = ["AWS::CDK::Metadata", "Module"]
            with open(f"{directory}module.json", "w", encoding="utf-8") as fh:
                json.dump(
                    {
                        "additionalProperties": True,
                        "type": "object",
                        "typeName": "Module",
                    },
                    fh,
                    indent=1,
                    separators=(",", ": "),
                    sort_keys=True,
                )
                fh.write("\n")
            for filename in filenames:
                with open(f"{directory}{filename}", "r+", encoding="utf-8") as fh:
                    spec = json.load(fh)
                    all_types.append(spec["typeName"])

            self._patch_region_schemas(region)

            # if the region is not us-east-1 compare the files to those in us-east-1
            # symlink if the files are the same
            if reg.name != self._region_primary.name:
                cached = ["Module"]
                for filename in os.listdir(directory):
                    if filename != "__init__.py":
                        try:
                            if filecmp.cmp(
                                f"{directory}{filename}",
                                f"{directory_pr}{filename}",
                            ):
                                cached.append(filename)
                                os.remove(f"{directory}{filename}")
                        except FileNotFoundError:
                            pass
                        except Exception as e:  # pylint: disable=broad-except
                            # Exceptions will typically be the file
                            # doesn't exist in primary region
                            LOGGER.info(
                                "Issuing comparing %s into %s: %s",
                                f"{directory}{filename}",
                                f"{directory_pr}{filename}",
                                e,
                            )
                with open(f"{directory}__init__.py", encoding="utf-8", mode="w") as f:
                    f.write("from __future__ import annotations\n\n")
                    f.write("# pylint: disable=too-many-lines\ntypes: list[str] = [\n")
                    for rt in sorted(all_types):
                        f.write(f'    "{rt}",\n')
                    f.write(
                        "]\n\n# pylint: disable=too-many-lines\ncached: list[str] = [\n"
                    )
                    for cache_file in sorted(cached):
                        f.write(f'    "{cache_file}",\n')
                    f.write("]\n")
            else:
                with open(f"{directory}__init__.py", encoding="utf-8", mode="w") as f:
                    f.write("from __future__ import annotations\n\n")
                    f.write("# pylint: disable=too-many-lines\ntypes: list[str] = [\n")
                    for rt in sorted(all_types):
                        f.write(f'    "{rt}",\n')
                    f.write("]\ncached: list[str] = []\n")

        except Exception as e:  # pylint: disable=broad-except
            LOGGER.info("Issuing updating schemas for %s: %s", region, e)

    def patch_schemas(self) -> None:

        self._patch_region_schemas(self._region_primary.name)
        # pylint: disable=not-context-manager
        with multiprocessing.Pool() as pool:
            # Patch from registry schema
            provider_pool_tuple = [
                (k,) for k in REGIONS if k != self._region_primary.name
            ]
            pool.starmap(self._patch_region_schemas, provider_pool_tuple)

    def _patch_region_schemas(self, region: str) -> None:
        reg = ToPy(region)
        directory = os.path.join(f"{self._root.path_relative}/{reg.py}/")

        filenames = [
            f
            for f in os.listdir(directory)
            if os.path.isfile(os.path.join(directory, f)) and f != "__init__.py"
        ]

        for filename in filenames:
            with open(f"{directory}{filename}", "r+", encoding="utf-8") as fh:
                spec = json.load(fh)
                try:
                    if "handlers" in spec:
                        del spec["handlers"]
                    if "tagging" in spec and "permissions" in spec.get("tagging", {}):
                        del spec["tagging"]["permissions"]
                        # tagging = spec.get("tagging", {})
                        # if "permissions" in tagging:
                        #    del tagging["permissions"]
                        #    spec["tagging"] = tagging
                    spec = self._remove_descriptions(spec)
                    spec = self._patch_provider_schema(spec, filename, "all")
                    spec = self._patch_provider_schema(spec, filename, region=reg.py)
                except Exception as e:  # pylint: disable=broad-except
                    LOGGER.info(
                        "Issuing patching schema for %s in %s: %s",
                        filename,
                        reg.name,
                        e,
                    )
                # Back to zero to write spec
                fh.seek(0)
                json.dump(
                    spec,
                    fh,
                    indent=1,
                    separators=(",", ": "),
                    sort_keys=True,
                )
                fh.write("\n")
                # Resize doc as needed
                fh.truncate()

    def _patch_provider_schema(
        self, content: Dict, source_filename: str, region: str
    ) -> Dict:
        """Provides the logic to patch a CloudFormation provider schema file.

        Args:
            content: A Dict representing the data that needs to be patched
            source_filename: The source filename for the JSON patches
            region: The region to apply the patch against
        Returns:
            Dict: returns the patched content
        """
        for patch_type in ["providers", "extensions"]:
            source_dir = source_filename.replace("-", "_").replace(".json", "")
            append_dir = os.path.join(
                self._patches.path_relative, patch_type, region, source_dir
            )
            for dirpath, _, filenames in os.walk(append_dir):
                filenames.sort()
                for filename in fnmatch.filter(filenames, "*.json"):
                    file_path = os.path.basename(filename)
                    module = dirpath.replace(f"{append_dir}", f"{region}").replace(
                        os.path.sep, "."
                    )
                    try:
                        jsonpatch.JsonPatch(
                            load_resource(
                                f"{self._patches.module}.{patch_type}.{module}.{source_dir}",
                                file_path,
                            )
                        ).apply(content, in_place=True)
                    except jsonpatch.JsonPatchConflict as e:
                        LOGGER.info(
                            "Patch already applied %s: %s",
                            os.path.join(append_dir, file_path),
                            str(e),
                        )
                    except jsonpatch.JsonPatchTestFailed as e:
                        LOGGER.info(
                            "Patch test failed %s: %s",
                            os.path.join(append_dir, file_path),
                            str(e),
                        )
                    except jsonpatch.JsonPatchException as e:
                        LOGGER.info(
                            "Patch exception raised for %s: %s",
                            os.path.join(append_dir, file_path),
                            str(e),
                        )
                    except jsonpointer.JsonPointerException as e:
                        LOGGER.info(
                            "Patch exception with pointer %s: %s",
                            os.path.join(append_dir, file_path),
                            str(e),
                        )
                    except Exception as e:  # pylint: disable=broad-exception-caught
                        LOGGER.info(
                            "Unknown exception raised applying patch %s: %s",
                            os.path.join(append_dir, file_path),
                            str(e),
                        )

        return content

    def patch(self, filename: str, regions: Sequence[str]):
        try:
            with open(filename, encoding="utf-8") as fp:
                custom_spec_data = json.load(fp)
                schema_patch = SchemaPatch.from_dict(custom_spec_data)
                for region in regions:
                    self._patch(schema_patch, region)
        except IOError as e:
            if e.errno == 2:
                LOGGER.error("Override spec file not found: %s", filename)
                sys.exit(1)
            elif e.errno == 21:
                LOGGER.error(
                    "Override spec file references a directory, not a file: %s",
                    filename,
                )
                sys.exit(1)
            elif e.errno == 13:
                LOGGER.error(
                    "Permission denied when accessing override spec file: %s", filename
                )
                sys.exit(1)
        except ValueError as err:
            LOGGER.error("Override spec file %s is malformed: %s", filename, err)
            sys.exit(1)

    def _patch(self, patch: SchemaPatch, region: str) -> None:
        """Patch the schemas as needed

        Args:
            patch: The patches to be applied to the schemas
        Returns:
            None: Returns when completed
        """

        resource_types = []
        all_resource_types = self.get_resource_types(region)[:]
        # Remove unsupported resource using includes
        if patch.included_resource_types:
            for include in patch.included_resource_types:
                regex = re.compile(include.replace("*", "(.*)") + "$")
                matches = [
                    string for string in all_resource_types if re.match(regex, string)
                ]

                resource_types.extend(matches)
        else:
            resource_types = all_resource_types[:]

        # Remove unsupported resources using the excludes
        for exclude in patch.excluded_resource_types:
            regex = re.compile(exclude.replace("*", "(.*)") + "$")
            matches = [string for string in resource_types if re.match(regex, string)]
            for match in matches:
                resource_types.remove(match)

        # Remove unsupported resources
        removed_types = list(set(all_resource_types) - set(resource_types))
        if removed_types:
            for removed_type in removed_types:
                self._removed_types.append(removed_type)
            self.get_resource_schema.cache_clear()
            self.get_resource_types.cache_clear()

        for resource_type, patches in patch.patches.items():
            try:
                schema = self.get_resource_schema(
                    resource_type=resource_type, region=region
                )
            except ResourceNotFoundError:
                # Resource type doesn't exist in this region
                continue
            schema.patch(patches=patches)

    @lru_cache(maxsize=None)
    def get_type_getatts(self, resource_type: str, region: str) -> AttributeDict:
        """Get the GetAtts for a type in a region

        Args:
            resource_type: The type of the resource. Example: AWS::S3::Bucket
            region: The region to load the resource type from
        Returns:
            Dict(str, Dict): Returns a Dict where the keys are the attributes and the
                value is the CloudFormation schema description of the attribute
        """
        resource_type = self._normalize_resource_type(resource_type)
        self.get_resource_schema(region=region, resource_type=resource_type)
        return self._schemas[region][resource_type].get_atts


PROVIDER_SCHEMA_MANAGER: ProviderSchemaManager = ProviderSchemaManager()