move system.py into misc folder

archivebox/abx/archivebox/base_configset.py

@@ -1,12 +1,9 @@
 __package__ = 'abx.archivebox'
 
 import os
-import re
-import json
 from pathlib import Path
-from typing import Type, Tuple, Callable, ClassVar, Any
+from typing import Type, Tuple, Callable, ClassVar
 
-import toml
 from benedict import benedict
 from pydantic import model_validator, TypeAdapter
 from pydantic_settings import BaseSettings, SettingsConfigDict, PydanticBaseSettingsSource
@@ -17,33 +14,13 @@ from pydantic_pkgr.base_types import func_takes_args_or_kwargs
 import abx
 
 from .base_hook import BaseHook, HookType
-from archivebox.misc import ini_to_toml
+from archivebox.misc import toml_util
 
 
 PACKAGE_DIR = Path(__file__).resolve().parent.parent
 DATA_DIR = Path(os.curdir).resolve()
 
 
-def better_toml_dump_str(val: Any) -> str:
-    try:
-        return toml.encoder._dump_str(val)     # type: ignore
-    except Exception:
-        # if we hit any of toml's numerous encoding bugs,
-        # fall back to using json representation of string
-        return json.dumps(str(val))
-
-class CustomTOMLEncoder(toml.encoder.TomlEncoder):
-    """
-    Custom TomlEncoder to work around https://github.com/uiri/toml's many encoding bugs.
-    More info: https://github.com/fabiocaccamo/python-benedict/issues/439
-    >>> toml.dumps(value, encoder=CustomTOMLEncoder())
-    """
-    def __init__(self, **kwargs):
-        super().__init__(**kwargs)
-        self.dump_funcs[str] = better_toml_dump_str
-        self.dump_funcs[re.RegexFlag] = better_toml_dump_str
-
-
 
 class FlatTomlConfigSettingsSource(TomlConfigSettingsSource):
     """
@@ -155,7 +132,7 @@ class ArchiveBoxBaseConfig(BaseSettings):
             # Convert ArchiveBox.conf in INI format to TOML and save original to .ArchiveBox.bak
             original_ini = ARCHIVEBOX_CONFIG_FILE.read_text()
             ARCHIVEBOX_CONFIG_FILE_BAK.write_text(original_ini)
-            new_toml = ini_to_toml.convert(original_ini)
+            new_toml = toml_util.convert(original_ini)
             ARCHIVEBOX_CONFIG_FILE.write_text(new_toml)
 
             precedence_order = {

archivebox/config/constants.py

@@ -6,6 +6,7 @@ import re
 from typing import Dict
 from pathlib import Path
 import importlib.metadata
+from collections.abc import Mapping
 
 from benedict import benedict
 
@@ -37,7 +38,7 @@ def _detect_installed_version(PACKAGE_DIR: Path):
 VERSION: str = _detect_installed_version(PACKAGE_DIR)
 
 
-class CONSTANTS:
+class ConstantsDict(Mapping):
     PACKAGE_DIR: Path = PACKAGE_DIR     # archivebox source code dir
     DATA_DIR: Path = DATA_DIR           # archivebox user data dir
     ARCHIVE_DIR: Path = ARCHIVE_DIR     # archivebox snapshot data dir
@@ -262,11 +263,24 @@ class CONSTANTS:
         },
     })
 
-    def __getitem__(self, key: str):
-        return getattr(self, key)
+    @classmethod
+    def __getitem__(cls, key: str):
+        return getattr(cls, key)
+    
+    @classmethod
+    def __benedict__(cls):
+        return benedict({key: value for key, value in cls.__dict__.items() if key.isupper() and not key.startswith('_')})
+    
+    @classmethod
+    def __len__(cls):
+        return len(cls.__benedict__())
 
+    @classmethod
+    def __iter__(cls):
+        return iter(cls.__benedict__())
 
-# add all key: values to globals() for easier importing
-globals().update(CONSTANTS.__dict__)
+CONSTANTS = ConstantsDict()
+CONSTANTS_CONFIG = CONSTANTS.__benedict__()
 
-CONSTANTS_CONFIG = CONSTANTS
+# add all key: values to globals() for easier importing
+globals().update(CONSTANTS)

archivebox/config/legacy.py

@@ -353,7 +353,7 @@ def load_config_file(out_dir: str | None=CONSTANTS.DATA_DIR) -> Optional[benedic
 def write_config_file(config: Dict[str, str], out_dir: str | None=CONSTANTS.DATA_DIR) -> benedict:
     """load the ini-formatted config file from OUTPUT_DIR/Archivebox.conf"""
 
-    from ..system import atomic_write
+    from archivebox.misc.system import atomic_write
 
     CONFIG_HEADER = (
     """# This is the config file for your ArchiveBox collection.

archivebox/core/models.py

@@ -22,7 +22,7 @@ from archivebox.config import CONSTANTS
 from abid_utils.models import ABIDModel, ABIDField, AutoDateTimeField
 from queues.tasks import bg_archive_snapshot
 
-from ..system import get_dir_size
+from archivebox.misc.system import get_dir_size
 from ..util import parse_date, base_url
 from ..index.schema import Link
 from ..index.html import snapshot_icons

archivebox/extractors/archive_org.py

@@ -6,7 +6,7 @@ from typing import Optional, List, Dict, Tuple
 from collections import defaultdict
 
 from ..index.schema import Link, ArchiveResult, ArchiveOutput, ArchiveError
-from ..system import run, chmod_file
+from archivebox.misc.system import run, chmod_file
 from ..util import (
     enforce_types,
     is_static_file,

archivebox/extractors/dom.py

@@ -4,7 +4,7 @@ from pathlib import Path
 from typing import Optional
 
 from ..index.schema import Link, ArchiveResult, ArchiveOutput, ArchiveError
-from ..system import run, chmod_file, atomic_write
+from archivebox.misc.system import run, chmod_file, atomic_write
 from ..util import (
     enforce_types,
     is_static_file,

archivebox/extractors/favicon.py

@@ -5,7 +5,7 @@ from pathlib import Path
 from typing import Optional
 
 from ..index.schema import Link, ArchiveResult, ArchiveOutput
-from ..system import chmod_file, run
+from archivebox.misc.system import chmod_file, run
 from ..util import (
     enforce_types,
     domain,

archivebox/extractors/git.py

@@ -5,7 +5,7 @@ from pathlib import Path
 from typing import Optional
 
 from ..index.schema import Link, ArchiveResult, ArchiveOutput, ArchiveError
-from ..system import run, chmod_file
+from archivebox.misc.system import run, chmod_file
 from ..util import (
     enforce_types,
     is_static_file,

archivebox/extractors/headers.py

@@ -5,7 +5,7 @@ from pathlib import Path
 from typing import Optional
 
 from ..index.schema import Link, ArchiveResult, ArchiveOutput
-from ..system import atomic_write
+from archivebox.misc.system import atomic_write
 from ..util import (
     enforce_types,
     get_headers,

archivebox/extractors/htmltotext.py

@@ -12,7 +12,7 @@ from ..config.legacy import (
 )
 from ..index.schema import Link, ArchiveResult, ArchiveError
 from ..logging_util import TimedProgress
-from ..system import atomic_write
+from archivebox.misc.system import atomic_write
 from ..util import (
     enforce_types,
     is_static_file,

archivebox/extractors/media.py

@@ -4,7 +4,7 @@ from pathlib import Path
 from typing import Optional
 
 from ..index.schema import Link, ArchiveResult, ArchiveOutput, ArchiveError
-from ..system import run, chmod_file
+from archivebox.misc.system import run, chmod_file
 from ..util import enforce_types, is_static_file, dedupe
 from ..logging_util import TimedProgress
 

archivebox/extractors/mercury.py

@@ -7,7 +7,7 @@ from typing import Optional, List
 import json
 
 from ..index.schema import Link, ArchiveResult, ArchiveError
-from ..system import run, atomic_write
+from archivebox.misc.system import run, atomic_write
 from ..util import (
     enforce_types,
     is_static_file,

archivebox/extractors/pdf.py

@@ -4,7 +4,7 @@ from pathlib import Path
 from typing import Optional
 
 from ..index.schema import Link, ArchiveResult, ArchiveOutput, ArchiveError
-from ..system import run, chmod_file
+from archivebox.misc.system import run, chmod_file
 from ..util import (
     enforce_types,
     is_static_file,

archivebox/extractors/readability.py

@@ -7,7 +7,7 @@ from typing import Optional
 import json
 
 from ..index.schema import Link, ArchiveResult, ArchiveError
-from ..system import run, atomic_write
+from archivebox.misc.system import run, atomic_write
 from ..util import enforce_types, is_static_file
 from ..logging_util import TimedProgress
 from .title import get_html

archivebox/extractors/screenshot.py

@@ -4,7 +4,7 @@ from pathlib import Path
 from typing import Optional
 
 from ..index.schema import Link, ArchiveResult, ArchiveOutput, ArchiveError
-from ..system import run, chmod_file
+from archivebox.misc.system import run, chmod_file
 from ..util import enforce_types, is_static_file
 from ..logging_util import TimedProgress
 

archivebox/extractors/singlefile.py

@@ -6,7 +6,7 @@ from typing import Optional
 import json
 
 from ..index.schema import Link, ArchiveResult, ArchiveError
-from ..system import run, chmod_file
+from archivebox.misc.system import run, chmod_file
 from ..util import enforce_types, is_static_file, dedupe
 from ..logging_util import TimedProgress
 

archivebox/extractors/wget.py

@@ -7,7 +7,7 @@ from typing import Optional
 from datetime import datetime, timezone
 
 from ..index.schema import Link, ArchiveResult, ArchiveOutput, ArchiveError
-from ..system import run, chmod_file
+from archivebox.misc.system import run, chmod_file
 from ..util import (
     enforce_types,
     without_fragment,

archivebox/index/html.py

@@ -9,7 +9,7 @@ from django.utils.html import format_html, mark_safe   # type: ignore
 from django.core.cache import cache
 
 from .schema import Link
-from ..system import atomic_write
+from archivebox.misc.system import atomic_write
 from ..logging_util import printable_filesize
 from ..util import (
     enforce_types,

archivebox/index/json.py

@@ -11,7 +11,7 @@ from typing import List, Optional, Iterator, Any, Union
 from archivebox.config import VERSION, DATA_DIR, CONSTANTS, SERVER_CONFIG, SHELL_CONFIG
 
 from .schema import Link
-from ..system import atomic_write
+from archivebox.misc.system import atomic_write
 from ..util import enforce_types
 
 

archivebox/index/schema.py

@@ -21,7 +21,7 @@ from archivebox.config.constants import ARCHIVE_DIR, ARCHIVE_DIR_NAME
 
 from plugins_extractor.favicon.apps import FAVICON_CONFIG
 
-from ..system import get_dir_size
+from archivebox.misc.system import get_dir_size
 from ..util import ts_to_date_str, parse_date
 
 

archivebox/logging_util.py

@@ -22,7 +22,7 @@ from rich import print
 from rich.panel import Panel
 
 from archivebox.config import CONSTANTS, DATA_DIR, VERSION, SHELL_CONFIG
-from .system import get_dir_size
+from archivebox.misc.system import get_dir_size
 from .util import enforce_types
 from .misc.logging import ANSI, stderr
 

archivebox/main.py

@@ -30,8 +30,8 @@ from .parsers import (
 )
 from .index.schema import Link
 from .util import enforce_types                         # type: ignore
-from .system import get_dir_size, dedupe_cron_jobs, CRON_COMMENT
-from .system import run as run_shell
+from archivebox.misc.system import get_dir_size, dedupe_cron_jobs, CRON_COMMENT
+from archivebox.misc.system import run as run_shell
 from .index import (
     load_main_index,
     parse_links_from_source,

archivebox/misc/tests.py

@@ -2,7 +2,7 @@ __package__ = 'abx.archivebox'
 
 from django.test import TestCase
 
-from .ini_to_toml import convert, TOML_HEADER
+from .toml_util import convert, TOML_HEADER
 
 TEST_INPUT = """
 [SERVER_CONFIG]

archivebox/misc/ini_to_toml.py → archivebox/misc/toml_util.py

@@ -3,8 +3,12 @@ from typing import Any, List, Callable
 import json
 import ast
 import inspect
+import toml
+import re
 import configparser
 
+from pathlib import Path, PosixPath
+
 from pydantic.json_schema import GenerateJsonSchema
 from pydantic_core import to_jsonable_python
 
@@ -68,8 +72,12 @@ def convert(ini_str: str) -> str:
 
 
 class JSONSchemaWithLambdas(GenerateJsonSchema):
+    """
+    Encode lambda functions in default values properly.
+    Usage:
+    >>> json.dumps(value, encoder=JSONSchemaWithLambdas())
+    """
     def encode_default(self, default: Any) -> Any:
-        """Encode lambda functions in default values properly"""
         config = self._config
         if isinstance(default, Callable):
             return '{{lambda ' + inspect.getsource(default).split('=lambda ')[-1].strip()[:-1] + '}}'
@@ -83,3 +91,24 @@ class JSONSchemaWithLambdas(GenerateJsonSchema):
     # for computed_field properties render them like this instead:
     # inspect.getsource(field.wrapped_property.fget).split('def ', 1)[-1].split('\n', 1)[-1].strip().strip('return '),
 
+
+def better_toml_dump_str(val: Any) -> str:
+    try:
+        return toml.encoder._dump_str(val)     # type: ignore
+    except Exception:
+        # if we hit any of toml's numerous encoding bugs,
+        # fall back to using json representation of string
+        return json.dumps(str(val))
+
+class CustomTOMLEncoder(toml.encoder.TomlEncoder):
+    """
+    Custom TomlEncoder to work around https://github.com/uiri/toml's many encoding bugs.
+    More info: https://github.com/fabiocaccamo/python-benedict/issues/439
+    >>> toml.dumps(value, encoder=CustomTOMLEncoder())
+    """
+    def __init__(self, **kwargs):
+        super().__init__(**kwargs)
+        self.dump_funcs[Path] = lambda x: json.dumps(str(x))
+        self.dump_funcs[PosixPath] = lambda x: json.dumps(str(x))
+        self.dump_funcs[str] = better_toml_dump_str
+        self.dump_funcs[re.RegexFlag] = better_toml_dump_str

archivebox/parsers/__init__.py

@@ -13,7 +13,7 @@ from typing import IO, Tuple, List, Optional
 from datetime import datetime, timezone
 from pathlib import Path 
 
-from ..system import atomic_write
+from archivebox.misc.system import atomic_write
 from ..config.legacy import (
     ANSI,
     OUTPUT_DIR,

archivebox/parsers/pocket_api.py

@@ -12,7 +12,7 @@ from archivebox.config import CONSTANTS
 
 from ..index.schema import Link
 from ..util import enforce_types
-from ..system import atomic_write
+from archivebox.misc.system import atomic_write
 from ..config.legacy import (
     POCKET_CONSUMER_KEY,
     POCKET_ACCESS_TOKENS,

archivebox/parsers/readwise_reader_api.py

@@ -12,7 +12,7 @@ from archivebox.config import CONSTANTS
 
 from ..index.schema import Link
 from ..util import enforce_types
-from ..system import atomic_write
+from archivebox.misc.system import atomic_write
 from ..config.legacy import READWISE_READER_TOKENS