In @integration-tests/.pytest.ini:
- Around line 3-6: Remove the inactive commented pytest options so the config is
clean: either delete the two commented lines containing "; --capture=no" and ";
-rA" if those flags are no longer needed, or, if they should be active,
uncomment the lines to restore "--capture=no" and "-rA"; update the pytest
config where the lines with "--capture=no" and "-rA" appear accordingly.
- Line 26: The pytest marker description for the 'search' marker is
inconsistent; update the marker line that currently reads 'search: search tests'
to follow the project's standard pattern (e.g., 'search: mark tests that
exercise search functionality' or similar) so it matches other marker
descriptions and improves clarity.

In @integration-tests/tests/package_tests/clp_json/test_clp_json.py:
- Around line 56-69: The placeholder test test_clp_json_compression_multifile
currently only calls validate_package_instance and asserts True; replace the
TODO by invoking the clp-json package via the fixt_package_instance to compress
the json-multifile dataset, capture the produced output (files/paths/metadata),
and add assertions that verify compression occurred and is correct (e.g.,
expected compressed file count/names, file sizes reduced, and checksums or
round-trip decompression equality against original data); use
validate_package_instance for setup, then call the package runner/handler on the
json-multifile input, collect outputs and assert expected outcomes instead of
the dummy assert, and log success with logger.info when checks pass.

In @integration-tests/tests/package_tests/clp_text/test_clp_text.py:
- Around line 52-53: The log message in test_clp_text.py uses an f-string
(log_msg = f"{CLP_TEXT_MODE.mode_name} is running successfully." followed by
logger.info(log_msg)) which is inconsistent with test_clp_json.py's hardcoded
log string; pick one style and standardize both tests. Update either
test_clp_text.py to use the same hardcoded string used in test_clp_json.py
(replacing log_msg and the f-string reference to CLP_TEXT_MODE.mode_name) or
change test_clp_json.py to use the CLP_TEXT_MODE.mode_name f-string pattern, and
ensure the logging call remains logger.info(...).

In @integration-tests/tests/utils/asserting_utils.py:
- Line 56: Replace the unnecessary use of object.__setattr__ when setting
instance_validated on package_instance: since PackageInstance is not frozen,
assign directly to the attribute (i.e., use package_instance.instance_validated
= True) wherever object.__setattr__(package_instance, "instance_validated",
True) appears (check functions in asserting_utils.py that manipulate
package_instance).

In @integration-tests/tests/utils/config.py:
- Line 123: Replace the non-idiomatic empty-parentheses dataclass decorators by
removing the parentheses: change each occurrence of @dataclass() to @dataclass
(the three decorator instances currently written as @dataclass() in the file) so
the dataclass declarations use the canonical form; no other behavior or imports
need to change.
- Around line 123-135: The PackageModeConfig dataclass declares component_list:
list[str] which is mutable; if you introduce a default for it later you must
avoid shared-state bugs by using dataclasses.field(default_factory=list) for
component_list in the PackageModeConfig class; either add component_list:
list[str] = field(default_factory=list) (importing field from dataclasses) or
explicitly document/enforce that callers must always pass a list to the
PackageModeConfig constructor so no implicit shared mutable default is used.

In `@integration-tests/tests/fixtures/package_instance.py`:
- Around line 28-42: The teardown always calls stop_clp_package even if
start_clp_package failed, which can raise and hide the original error; modify
the fixture to track whether start_clp_package succeeded (e.g., set a local
started flag after start_clp_package returns and/or after constructing
PackageInstance) and only call stop_clp_package in the finally block when that
flag is true; reference start_clp_package, stop_clp_package, PackageInstance and
fixt_package_test_config so you add the flag in the same scope and conditionally
run stop_clp_package to avoid stopping a package that never started.

In `@integration-tests/tests/fixtures/package_test_config.py`:
- Around line 32-37: The current conversion of base_port_string (from
request.config.getoption("--base-port")) only catches ValueError, so a missing
option causing int(None) raises TypeError and bypasses the handler; update the
error handling in the base_port assignment (where base_port_string, base_port
are used) to catch both ValueError and TypeError (e.g., except (ValueError,
TypeError) as err) and re-raise a ValueError with the same user-friendly message
so omitted or malformed --base-port options produce the intended error message.

In `@integration-tests/tests/utils/config.py`:
- Around line 147-152: Validate the base_port value in the dataclass
__post_init__ before calling _write_temp_config_file: check that self.base_port
is an int within 1–65535 and if not raise a ValueError with a clear message
including the invalid value (e.g., "base_port must be between 1 and 65535, got
X"); perform this check at the start of __post_init__ so invalid configs fail
fast.
- Around line 123-134: Add a __post_init__ to PackageModeConfig that validates
mode_name is a non-empty, non-whitespace string (e.g., strip() and check
truthiness) and raise ValueError with a clear message if invalid; locate the
dataclass PackageModeConfig and implement the check in its __post_init__ (follow
the same pattern used by IntegrationTestLogs and CompressionTestPathConfig) so
instantiation fails fast for empty mode_name.

In `@integration-tests/tests/utils/config.py`:
- Around line 133-134: The dataclass currently declares component_list:
list[str] which allows mutation even with frozen=True; change the type to an
immutable sequence (e.g., component_list: tuple[str, ...]) and update any
constructors/uses to pass/convert tuples (or wrap lists with tuple(...)) so the
frozen dataclass provides true immutability; ensure the dataclass definition and
any code that constructs or iterates over component_list (tests/helpers that
build config objects) are updated to accept the tuple type.

In `@integration-tests/tests/package_tests/clp_json/test_clp_json.py`:
- Around line 79-85: Replace the copied "compress some dataset" TODO with a
clear, accurate task for preparing data for search: update the comment near the
validate_package_instance call (in test_clp_json) to say something like "prepare
dataset for search (e.g., create or compress dataset as needed) and validate
search correctness" so it no longer duplicates the compression-specific TODO
from test_clp_json_compression and clarifies that compression may be a
prerequisite but the goal is to prepare the dataset for subsequent search
verification.