deprecation: Remove support for Files

File support was complicating design and increased support load. Also it
is not used often. Focus on databases and data warehouses.

piicatcher/__init__.py

@@ -1,4 +1,4 @@
 # flake8: noqa
 __version__ = "0.13.0"
 
-from piicatcher.api import scan_database, scan_file_object
+from piicatcher.api import scan_database

piicatcher/api.py

@@ -1,5 +1,5 @@
 from argparse import Namespace
-from typing import Any, Dict, List, TextIO, Tuple
+from typing import Any, Dict, Tuple
 
 from piicatcher.explorer.aws import AthenaExplorer
 from piicatcher.explorer.databases import (
@@ -9,29 +9,7 @@
     RedshiftExplorer,
 )
 from piicatcher.explorer.explorer import Explorer
-from piicatcher.explorer.files import IO, Tokenizer
 from piicatcher.explorer.sqlite import SqliteExplorer
-from piicatcher.scanner import NERScanner, RegexScanner
-
-
-def scan_file_object(fd: TextIO) -> List[Any]:
-    """
-
-    Args:
-        fd (file descriptor): A file descriptor open in text mode.
-
-    Returns: A list of PIITypes enum of all the PII types found in the file.
-
-    """
-    scanner = IO("api file object", fd)
-    context = {
-        "tokenizer": Tokenizer(),
-        "regex": RegexScanner(),
-        "ner": NERScanner(),
-    }
-
-    scanner.scan(context)
-    return scanner.get_pii_types()
 
 
 def _scan_db(scanner: Explorer, scan_type: str) -> Dict[Any, Any]:

piicatcher/command_line.py

@@ -7,7 +7,6 @@
 from piicatcher import __version__
 from piicatcher.explorer.aws import cli as aws_cli
 from piicatcher.explorer.databases import cli as db_cli
-from piicatcher.explorer.files import cli as files_cli
 from piicatcher.explorer.metadata import data_logger, scan_logger
 from piicatcher.explorer.snowflake import cli as snowflake_cli
 from piicatcher.explorer.sqlite import cli as sqlite_cli
@@ -87,7 +86,6 @@ def cli(
 
 
 cli.add_command(aws_cli)
-cli.add_command(files_cli)
 cli.add_command(sqlite_cli)
 cli.add_command(db_cli)
 cli.add_command(snowflake_cli)

tests/test_config_file.py

@@ -108,38 +108,6 @@ def test_sqlite(tmp_path, mocker, caplog):
     )
 
 
-def test_files(tmp_path, mocker, caplog):
-    caplog.set_level(logging.DEBUG)
-    config_file = tmp_path / "db_host.ini"
-    config_file.write_text(
-        """
-[files]
-path="file path"
-"""
-    )
-
-    logging.info("Config File: %s" % config_file)
-    explorer = mocker.patch("piicatcher.explorer.files.FileExplorer")
-    runner = CliRunner()
-    result = runner.invoke(cli, ["--config", str(config_file), "files"])
-    assert result.exception is None
-    assert "" == result.stdout
-    assert 0 == result.exit_code
-    explorer.dispatch.assert_called_once_with(
-        Namespace(
-            path="file path",
-            catalog={
-                "host": None,
-                "port": None,
-                "user": None,
-                "password": None,
-                "format": "ascii_table",
-                "file": None,
-            },
-        )
-    )
-
-
 def test_aws(tmp_path, mocker, caplog):
     caplog.set_level(logging.DEBUG)
     config_file = tmp_path / "db_host.ini"