|
| 1 | +# Copyright (c) 2025, Salesforce, Inc. |
| 2 | +# SPDX-License-Identifier: Apache-2 |
| 3 | +# |
| 4 | +# Licensed under the Apache License, Version 2.0 (the "License"); |
| 5 | +# you may not use this file except in compliance with the License. |
| 6 | +# You may obtain a copy of the License at |
| 7 | +# |
| 8 | +# http://www.apache.org/licenses/LICENSE-2.0 |
| 9 | +# |
| 10 | +# Unless required by applicable law or agreed to in writing, software |
| 11 | +# distributed under the License is distributed on an "AS IS" BASIS, |
| 12 | +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. |
| 13 | +# See the License for the specific language governing permissions and |
| 14 | +# limitations under the License. |
| 15 | +from __future__ import annotations |
| 16 | + |
| 17 | +import io |
| 18 | +import os |
| 19 | +from pathlib import Path |
| 20 | +from typing import Optional |
| 21 | + |
| 22 | +from datacustomcode.file.base import BaseDataAccessLayer |
| 23 | + |
| 24 | + |
| 25 | +class FileReaderError(Exception): |
| 26 | + """Base exception for file reader operations.""" |
| 27 | + |
| 28 | + |
| 29 | +class FileNotFoundError(FileReaderError): |
| 30 | + """Raised when a file cannot be found.""" |
| 31 | + |
| 32 | + |
| 33 | +class FileAccessError(FileReaderError): |
| 34 | + """Raised when there's an error accessing a file.""" |
| 35 | + |
| 36 | + |
| 37 | +class DefaultFileReader(BaseDataAccessLayer): |
| 38 | + """Base class for file reading operations. |
| 39 | +
|
| 40 | + This class provides a framework for reading files from various locations |
| 41 | + with configurable search strategies and error handling. |
| 42 | + """ |
| 43 | + |
| 44 | + # Default configuration values |
| 45 | + DEFAULT_CODE_PACKAGE = "payload" |
| 46 | + DEFAULT_FILE_FOLDER = "files" |
| 47 | + DEFAULT_CONFIG_FILE = "config.json" |
| 48 | + |
| 49 | + def __init__( |
| 50 | + self, |
| 51 | + code_package: Optional[str] = None, |
| 52 | + file_folder: Optional[str] = None, |
| 53 | + config_file: Optional[str] = None, |
| 54 | + ): |
| 55 | + """Initialize the file reader with configuration. |
| 56 | +
|
| 57 | + Args: |
| 58 | + code_package: The default code package directory to search |
| 59 | + file_folder: The folder containing files relative to the code package |
| 60 | + config_file: The configuration file to use for path resolution |
| 61 | + """ |
| 62 | + self.code_package = code_package or self.DEFAULT_CODE_PACKAGE |
| 63 | + self.file_folder = file_folder or self.DEFAULT_FILE_FOLDER |
| 64 | + self.config_file = config_file or self.DEFAULT_CONFIG_FILE |
| 65 | + |
| 66 | + def file_open(self, file_name: str) -> io.TextIOWrapper: |
| 67 | + """Open a file for reading. |
| 68 | +
|
| 69 | + Args: |
| 70 | + file_name: The name of the file to open |
| 71 | +
|
| 72 | + Returns: |
| 73 | + A file handle for reading |
| 74 | +
|
| 75 | + Raises: |
| 76 | + FileNotFoundError: If the file cannot be found |
| 77 | + FileAccessError: If there's an error opening the file |
| 78 | + """ |
| 79 | + if not file_name: |
| 80 | + raise ValueError("file_name cannot be empty") |
| 81 | + |
| 82 | + file_path = self._resolve_file_path(file_name) |
| 83 | + |
| 84 | + if not file_path: |
| 85 | + raise FileNotFoundError( |
| 86 | + f"File '{file_name}' not found in any search location" |
| 87 | + ) |
| 88 | + |
| 89 | + try: |
| 90 | + return self._open_file(file_path) |
| 91 | + except (OSError, IOError) as e: |
| 92 | + raise FileAccessError(f"Error opening file '{file_path}': {e}") from e |
| 93 | + |
| 94 | + def _resolve_file_path(self, file_name: str) -> Optional[Path]: |
| 95 | + """Resolve the full path to a file. |
| 96 | +
|
| 97 | + Args: |
| 98 | + file_name: The name of the file to resolve |
| 99 | +
|
| 100 | + Returns: |
| 101 | + The full path to the file, or None if not found |
| 102 | + """ |
| 103 | + # First try the default code package location |
| 104 | + if self._code_package_exists(): |
| 105 | + file_path = self._get_code_package_file_path(file_name) |
| 106 | + if file_path.exists(): |
| 107 | + return file_path |
| 108 | + |
| 109 | + # Fall back to config.json-based location |
| 110 | + config_path = self._find_config_file() |
| 111 | + if config_path: |
| 112 | + file_path = self._get_config_based_file_path(file_name, config_path) |
| 113 | + if file_path.exists(): |
| 114 | + return file_path |
| 115 | + |
| 116 | + return None |
| 117 | + |
| 118 | + def _code_package_exists(self) -> bool: |
| 119 | + """Check if the default code package directory exists. |
| 120 | +
|
| 121 | + Returns: |
| 122 | + True if the code package directory exists |
| 123 | + """ |
| 124 | + return os.path.exists(self.code_package) |
| 125 | + |
| 126 | + def _get_code_package_file_path(self, file_name: str) -> Path: |
| 127 | + """Get the file path relative to the code package. |
| 128 | +
|
| 129 | + Args: |
| 130 | + file_name: The name of the file |
| 131 | +
|
| 132 | + Returns: |
| 133 | + The full path to the file |
| 134 | + """ |
| 135 | + relative_path = f"{self.code_package}/{self.file_folder}/{file_name}" |
| 136 | + return Path.cwd().joinpath(relative_path) |
| 137 | + |
| 138 | + def _find_config_file(self) -> Optional[Path]: |
| 139 | + """Find the configuration file in the current directory tree. |
| 140 | +
|
| 141 | + Returns: |
| 142 | + The path to the config file, or None if not found |
| 143 | + """ |
| 144 | + return self._find_file_in_tree(self.config_file, Path.cwd()) |
| 145 | + |
| 146 | + def _get_config_based_file_path(self, file_name: str, config_path: Path) -> Path: |
| 147 | + """Get the file path relative to the config file location. |
| 148 | +
|
| 149 | + Args: |
| 150 | + file_name: The name of the file |
| 151 | + config_path: The path to the config file |
| 152 | +
|
| 153 | + Returns: |
| 154 | + The full path to the file |
| 155 | + """ |
| 156 | + relative_path = f"{self.file_folder}/{file_name}" |
| 157 | + return config_path.parent.joinpath(relative_path) |
| 158 | + |
| 159 | + def _find_file_in_tree(self, filename: str, search_path: Path) -> Optional[Path]: |
| 160 | + """Find a file within a directory tree. |
| 161 | +
|
| 162 | + Args: |
| 163 | + filename: The name of the file to find |
| 164 | + search_path: The root directory to search from |
| 165 | +
|
| 166 | + Returns: |
| 167 | + The full path to the file, or None if not found |
| 168 | + """ |
| 169 | + for file_path in search_path.rglob(filename): |
| 170 | + return file_path |
| 171 | + return None |
| 172 | + |
| 173 | + def _open_file(self, file_path: Path) -> io.TextIOWrapper: |
| 174 | + """Open a file at the given path. |
| 175 | +
|
| 176 | + Args: |
| 177 | + file_path: The path to the file |
| 178 | +
|
| 179 | + Returns: |
| 180 | + A file handle for reading |
| 181 | + """ |
| 182 | + return open(file_path, "r", encoding="utf-8") |
| 183 | + |
| 184 | + def get_search_locations(self) -> list[Path]: |
| 185 | + """Get all possible search locations for files. |
| 186 | +
|
| 187 | + Returns: |
| 188 | + A list of paths where files might be found |
| 189 | + """ |
| 190 | + locations = [] |
| 191 | + |
| 192 | + # Add code package location |
| 193 | + if self._code_package_exists(): |
| 194 | + locations.append(Path.cwd().joinpath(self.code_package, self.file_folder)) |
| 195 | + |
| 196 | + # Add config-based location |
| 197 | + config_path = self._find_config_file() |
| 198 | + if config_path: |
| 199 | + locations.append(config_path.parent.joinpath(self.file_folder)) |
| 200 | + |
| 201 | + return locations |
| 202 | + |
| 203 | + |
| 204 | +class BaseDataAccessLayer: |
| 205 | + """Default implementation of the file reader. |
| 206 | +
|
| 207 | + This class provides the standard file reading behavior and can be |
| 208 | + easily mocked or subclassed for testing. |
| 209 | + """ |
0 commit comments