Merge pull request #111 from openzim/handle_descriptions

rgaudin · web-flow · commit cb01a0f34d54 · 2023-08-22T13:38:08.000Z
Add utility function to compute/check ZIM descriptions
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -5,7 +5,12 @@ All notable changes to this project are documented in this file.
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html) (as of version 1.5.0).
 
-## [3.1.1]
+## [Unreleased]
+
+### Added
+- Add utility function to compute/check ZIM descriptions #110
+
+## [3.1.1] - 2023-07-18
 
 ### Changed
 
diff --git a/src/zimscraperlib/constants.py b/src/zimscraperlib/constants.py
@@ -48,6 +48,7 @@
     ),
 }
 
+RECOMMENDED_MAX_TITLE_LENGTH = 30
 MAXIMUM_DESCRIPTION_METADATA_LENGTH = 80
 MAXIMUM_LONG_DESCRIPTION_METADATA_LENGTH = 4000
 
diff --git a/src/zimscraperlib/inputs.py b/src/zimscraperlib/inputs.py
@@ -5,16 +5,18 @@
 import pathlib
 import shutil
 import tempfile
-from typing import Optional, Union
+from typing import Optional, Tuple, Union
 
 from . import logger
+from .constants import MAXIMUM_DESCRIPTION_METADATA_LENGTH as MAX_DESC_LENGTH
+from .constants import MAXIMUM_LONG_DESCRIPTION_METADATA_LENGTH as MAX_LONG_DESC_LENGTH
 from .download import stream_file
 
 
 def handle_user_provided_file(
     source: Optional[Union[pathlib.Path, str]] = None,
     dest: Optional[pathlib.Path] = None,
-    in_dir: pathlib.Path = None,
+    in_dir: Optional[pathlib.Path] = None,
     nocopy: bool = False,
 ) -> Union[pathlib.Path, None]:
     """path to downloaded or copied a user provided file (URL or path)
@@ -49,3 +51,56 @@ def handle_user_provided_file(
         shutil.copy(source, dest)
 
     return dest
+
+
+def compute_descriptions(
+    default_description: str,
+    user_description: Optional[str],
+    user_long_description: Optional[str],
+) -> Tuple[str, Optional[str]]:
+    """Computes short and long descriptions compliant with ZIM standard.
+
+    Based on provided parameters, the function computes a short and a long description
+    which are compliant with the ZIM standard (in terms of length).
+
+    User description(s) are used if set. They are checked to not exceed ZIM standard
+    maximum length ; an error is thrown otherwise ; if ok, they are returned.
+
+    If user_description is not set, the description is computed based on the default
+    description, truncated if needed.
+
+    If user_long_description is not set and default description is too long for the
+    description field, the long_description is computed based on the default description
+    (truncated if needed), otherwise no long description is returned.
+
+    args:
+        default_description:   the description which will be used if user descriptions
+                               are not set (typically fetched online)
+        user_description:      the description set by the user (typically set by a
+                               CLI argument)
+        user_long_description: the long description set by the user (typically set by a
+                               CLI argument)
+
+    Returns a tuple of (description, long_description)
+    """
+
+    if user_description and len(user_description) > MAX_DESC_LENGTH:
+        raise ValueError(
+            f"Description too long ({len(user_description)}>{MAX_DESC_LENGTH})"
+        )
+    if user_long_description and len(user_long_description) > MAX_LONG_DESC_LENGTH:
+        raise ValueError(
+            f"LongDescription too long ({len(user_long_description)}"
+            f">{MAX_LONG_DESC_LENGTH})"
+        )
+
+    if not user_long_description and len(default_description) > MAX_DESC_LENGTH:
+        user_long_description = default_description[0:MAX_LONG_DESC_LENGTH]
+        if len(default_description) > MAX_LONG_DESC_LENGTH:
+            user_long_description = user_long_description[:-1] + "…"
+    if not user_description:
+        user_description = default_description[0:MAX_DESC_LENGTH]
+        if len(default_description) > MAX_DESC_LENGTH:
+            user_description = user_description[:-1] + "…"
+
+    return (user_description, user_long_description)
diff --git a/src/zimscraperlib/zim/metadata.py b/src/zimscraperlib/zim/metadata.py
@@ -8,6 +8,7 @@
     MANDATORY_ZIM_METADATA_KEYS,
     MAXIMUM_DESCRIPTION_METADATA_LENGTH,
     MAXIMUM_LONG_DESCRIPTION_METADATA_LENGTH,
+    RECOMMENDED_MAX_TITLE_LENGTH,
 )
 from ..i18n import is_valid_iso_639_3
 from ..image.probing import is_valid_image
@@ -40,7 +41,7 @@ def validate_standard_str_types(name: str, value: str):
 
 def validate_title(name: str, value: str):
     """ensures Title metadata is within recommended length"""
-    if name == "Title" and len(value) > 30:
+    if name == "Title" and len(value) > RECOMMENDED_MAX_TITLE_LENGTH:
         raise ValueError(f"{name} is too long.")
 
 
diff --git a/tests/inputs/test_inputs.py b/tests/inputs/test_inputs.py
@@ -3,10 +3,17 @@
 # vim: ai ts=4 sts=4 et sw=4 nu
 
 import pathlib
+from typing import Optional
 
 import pytest
 
-from zimscraperlib.inputs import handle_user_provided_file
+from zimscraperlib.constants import (
+    MAXIMUM_DESCRIPTION_METADATA_LENGTH as MAX_DESC_LENGTH,
+)
+from zimscraperlib.constants import (
+    MAXIMUM_LONG_DESCRIPTION_METADATA_LENGTH as MAX_LONG_DESC_LENGTH,
+)
+from zimscraperlib.inputs import compute_descriptions, handle_user_provided_file
 
 
 def test_with_none():
@@ -72,3 +79,169 @@ def test_remote_indir(tmp_path, valid_http_url):
     assert fpath is not None
     assert fpath.exists()
     assert fpath.parent == tmp_path
+
+
+TEXT_NOT_USED = "text not used"
+
+LONG_TEXT = (
+    "Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor "
+    "incididunt ut labore et dolore magna aliqua. At erat pellentesque adipiscing "
+    "commodo elit at imperdiet. Rutrum tellus pellentesque eu tincidunt tortor aliquam"
+    " nulla facilisi. Eget lorem dolor sed viverra ipsum nunc. Ipsum nunc aliquet "
+    "bibendum enim facilisis gravida neque convallis. Aliquam malesuada bibendum arcu "
+    "vitae elementum curabitur. Platea dictumst quisque sagittis purus sit amet "
+    "volutpat. Blandit libero volutpat sed cras ornare. In eu mi bibendum neque "
+    "egestas. Egestas dui id ornare arcu odio. Pulvinar neque laoreet suspendisse "
+    "interdum. Fames ac turpis egestas integer eget aliquet nibh praesent tristique. Et"
+    " egestas quis ipsum suspendisse ultrices gravida dictum fusce. Malesuada fames ac "
+    "turpis egestas. Tincidunt nunc pulvinar sapien et ligula ullamcorper malesuada "
+    "proin libero. In arcu cursus euismod quis viverra. Faucibus in ornare quam viverra"
+    ". Curabitur vitae nunc sed velit dignissim sodales ut eu sem. Velit scelerisque in"
+    " dictum non consectetur a erat nam. Proin fermentum leo vel orci porta non. Fames"
+    " ac turpis egestas sed tempus. Vitae justo eget magna fermentum iaculis eu non. "
+    "Imperdiet massa tincidunt nunc pulvinar sapien et ligula. Laoreet sit amet cursus "
+    "sit amet dictum sit amet. Quis hendrerit dolor magna eget. Orci ac auctor augue "
+    "mauris augue. Consequat interdum varius sit amet mattis. At ultrices mi tempus "
+    "imperdiet nulla malesuada pellentesque elit. Volutpat est velit egestas dui. "
+    "Potenti nullam ac tortor vitae. At tempor commodo ullamcorper a lacus vestibulum "
+    "sed arcu non. Duis ut diam quam nulla. Vestibulum mattis ullamcorper velit sed "
+    "ullamcorper. Sit amet commodo nulla facilisi nullam vehicula. Faucibus purus in "
+    "massa tempor nec feugiat. Sem fringilla ut morbi tincidunt augue interdum velit. "
+    "Etiam dignissim diam quis enim lobortis scelerisque fermentum dui. Nunc vel risus "
+    "commodo viverra maecenas accumsan. Aenean sed adipiscing diam donec adipiscing "
+    "tristique. Maecenas accumsan lacus vel facilisis volutpat est velit egestas. Nulla"
+    " aliquet porttitor lacus luctus accumsan tortor posuere ac. Habitant morbi "
+    "tristique senectus et netus et. Eget mi proin sed libero enim sed faucibus turpis "
+    "in. Vulputate enim nulla aliquet porttitor lacus. Dui ut ornare lectus sit amet "
+    "est. Quam lacus suspendisse faucibus interdum posuere. Sagittis orci a scelerisque"
+    " purus semper eget duis at tellus. Tellus molestie nunc non blandit massa. Feugiat"
+    " vivamus at augue eget arcu dictum varius duis at. Varius morbi enim nunc faucibus"
+    " a pellentesque sit. Id aliquet lectus proin nibh nisl condimentum id venenatis a."
+    " Tortor dignissim convallis aenean et tortor at risus viverra adipiscing. Aliquam "
+    "malesuada bibendum arcu vitae elementum curabitur vitae nunc sed. Habitasse platea"
+    " dictumst quisque sagittis purus sit amet volutpat. Vitae auctor eu augue ut "
+    "lectus. At varius vel pharetra vel turpis nunc eget. Dictum at tempor  commodo "
+    "ullamcorper a lacus vestibulum sed arcu. Pellentesque massa placerat duis "
+    "ultricies. Enim nunc faucibus a pellentesque sit amet porttitor eget dolor. "
+    "Volutpat blandit aliquam etiam erat velit scelerisque in. Amet mattis vulputate "
+    "enim nulla aliquet porttitor. Egestas maecenas pharetra convallis posuere morbi "
+    "leo urna molestie. Duis ut diam quam nulla porttitor massa id. In fermentum "
+    "posuere urna nec tincidunt praesent. Turpis egestas sed tempus urna et pharetra "
+    "pharetra massa. Tellus molestie nunc non blandit massa. Diam phasellus vestibulum "
+    "lorem sed risus ultricies. Egestas erat imperdiet sed euismod nisi porta lorem. "
+    "Quam viverra orci sagittis eu volutpat odio facilisis mauris sit. Ornare aenean "
+    "euismod elementum nisi quis. Laoreet non curabitur gravida arcu ac tortor "
+    "dignissim convallis aenean. Sagittis aliquam malesuada bibendum arcu vitae "
+    "elementum. Sed blandit libero volutpat sed cras ornare. Sagittis eu volutpat odio "
+    "facilisis mauris. Facilisis volutpat est velit egestas dui id ornare arcu odio. "
+    "Eu feugiat pretium  nibh."
+)
+
+
+@pytest.mark.parametrize(
+    "user_description, user_long_description, default_description, raises, "
+    "expected_description, expected_long_description",
+    [
+        # user description set and is short, user long descripion not set, default
+        # description doe not matter
+        (
+            LONG_TEXT[0:MAX_DESC_LENGTH],
+            None,
+            TEXT_NOT_USED,
+            False,
+            LONG_TEXT[0:MAX_DESC_LENGTH],
+            None,
+        ),
+        # user description set and is too long, default description does not matter
+        (LONG_TEXT[0 : MAX_DESC_LENGTH + 1], None, TEXT_NOT_USED, True, None, None),
+        # user description not set and default description is short enough
+        (
+            None,
+            None,
+            LONG_TEXT[0:MAX_DESC_LENGTH],
+            False,
+            LONG_TEXT[0:MAX_DESC_LENGTH],
+            None,
+        ),
+        # user description not set and default description is too long for description
+        # but ok for long description
+        (
+            None,
+            None,
+            LONG_TEXT[0 : MAX_DESC_LENGTH + 1],
+            False,
+            LONG_TEXT[0 : MAX_DESC_LENGTH - 1] + "…",
+            LONG_TEXT[0 : MAX_DESC_LENGTH + 1],
+        ),
+        (
+            None,
+            None,
+            LONG_TEXT[0:MAX_LONG_DESC_LENGTH],
+            False,
+            LONG_TEXT[0 : MAX_DESC_LENGTH - 1] + "…",
+            LONG_TEXT[0:MAX_LONG_DESC_LENGTH],
+        ),
+        # user description not set and default description is too long for description
+        # and long description
+        (
+            None,
+            None,
+            LONG_TEXT[0 : MAX_LONG_DESC_LENGTH + 1],
+            False,
+            LONG_TEXT[0 : MAX_DESC_LENGTH - 1] + "…",
+            LONG_TEXT[0 : MAX_LONG_DESC_LENGTH - 1] + "…",
+        ),
+        # user description set and is short, user long descripion set and is short,
+        # default description does not matter
+        (
+            LONG_TEXT[0:MAX_DESC_LENGTH],
+            LONG_TEXT[0:MAX_LONG_DESC_LENGTH],
+            TEXT_NOT_USED,
+            False,
+            LONG_TEXT[0:MAX_DESC_LENGTH],
+            LONG_TEXT[0:MAX_LONG_DESC_LENGTH],
+        ),
+        # user description set and is short, user long descripion set and is too long,
+        # default description does not matter
+        (
+            LONG_TEXT[0:MAX_DESC_LENGTH],
+            LONG_TEXT[0 : MAX_LONG_DESC_LENGTH + 1],
+            TEXT_NOT_USED,
+            True,
+            None,
+            None,
+        ),
+        # user description not set, user long descripion set and is short,
+        # default description set to something different than long desc
+        (
+            None,
+            LONG_TEXT[0:MAX_LONG_DESC_LENGTH],
+            LONG_TEXT[10:MAX_LONG_DESC_LENGTH],
+            False,
+            LONG_TEXT[10 : MAX_DESC_LENGTH + 9] + "…",
+            LONG_TEXT[0:MAX_LONG_DESC_LENGTH],
+        ),
+    ],
+)
+def test_description(
+    user_description: str,
+    user_long_description: Optional[str],
+    default_description: str,
+    *,
+    raises: bool,
+    expected_description: str,
+    expected_long_description: str,
+):
+    if raises:
+        with pytest.raises(ValueError):
+            compute_descriptions(
+                default_description, user_description, user_long_description
+            )
+        return
+    else:
+        (description, long_description) = compute_descriptions(
+            default_description, user_description, user_long_description
+        )
+
+    assert description == expected_description
+    assert long_description == expected_long_description
diff --git a/tox.ini b/tox.ini
@@ -11,14 +11,15 @@ deps =
     pytest-cov
     -r{toxinidir}/requirements.txt
 commands =
-    pytest --runslow --runinstalled --cov=zimscraperlib --cov-report=term --cov-report term-missing
+    pytest --runslow --runinstalled --cov=zimscraperlib --cov-report=term --cov-report term-missing {posargs}
 
 [testenv:black]
 deps =
     black>=23.1.0,<24
 commands = black --check .
 
 [testenv:isort]
+base_python=py38
 deps =
     isort>=5.12.0,<5.13
 commands = isort --profile black --check src tests

Original file line number	Diff line number	Diff line change
`@@ -48,6 +48,7 @@`
`48`	`48`	`),`
`49`	`49`	`}`
`50`	`50`
	`51`	`+RECOMMENDED_MAX_TITLE_LENGTH = 30`
`51`	`52`	`MAXIMUM_DESCRIPTION_METADATA_LENGTH = 80`
`52`	`53`	`MAXIMUM_LONG_DESCRIPTION_METADATA_LENGTH = 4000`
`53`	`54`