grz-cli 0.1.4__tar.gz → 0.3.0__tar.gz

{grz_cli-0.1.4 → grz_cli-0.3.0}/MANIFEST.in

@@ -1,2 +1,3 @@
-include src/grz_cli/resources/grz-schema.json
+graft src
 include src/grz_cli/resources/thresholds.json
+include LICENSE

{grz_cli-0.1.4/src/grz_cli.egg-info → grz_cli-0.3.0}/PKG-INFO

@@ -1,6 +1,6 @@
-Metadata-Version: 2.2
+Metadata-Version: 2.4
 Name: grz-cli
-Version: 0.1.4
+Version: 0.3.0
 Summary: Tool for validation, encryption and upload of MV submissions to GDCs.
 Author-email: Koray Kirli <koraykirli@gmail.com>, Mathias Lesche <mathias.lesche@tu-dresden.de>, "Florian R. Hölzlwimmer" <git.ich@frhoelzlwimmer.de>, Till Hartmann <till.hartmann@bih-charite.de>, Thomas Sell <thomas.sell@bih-charite.de>
 License: MIT
@@ -19,6 +19,7 @@ Requires-Dist: pyyaml<7,>=6.0.2
 Requires-Dist: tqdm<5,>=4.66.5
 Requires-Dist: pydantic<2.10,>=2.9.2
 Requires-Dist: platformdirs<5,>=4.3.6
+Requires-Dist: grz-pydantic-models>=1.2.0
 Provides-Extra: test
 Requires-Dist: pytest<9,>=8.3.3; extra == "test"
 Requires-Dist: pytest-cov<6,>=5.0.0; extra == "test"

{grz_cli-0.1.4 → grz_cli-0.3.0}/pyproject.toml

@@ -4,7 +4,6 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "grz-cli"
-version = "0.1.4"
 description = "Tool for validation, encryption and upload of MV submissions to GDCs."
 readme = "README.md"
 license = { text = "MIT" }
@@ -15,6 +14,7 @@ authors = [
     { name = "Till Hartmann", email = "till.hartmann@bih-charite.de" },
     { name = "Thomas Sell", email = "thomas.sell@bih-charite.de" },
 ]
+dynamic = ["version"]
 dependencies = [
     "boto3 >=1.35.39,<2",
     "click >=8.1.7,<9",
@@ -24,6 +24,7 @@ dependencies = [
     "tqdm >=4.66.5,<5",
     "pydantic >=2.9.2,<2.10",
     "platformdirs >=4.3.6,<5",
+    "grz-pydantic-models >=1.2.0"
 ]
 classifiers = [
     "Programming Language :: Python :: 3",
@@ -33,6 +34,10 @@ classifiers = [
 requires-python = ">=3.12"
 keywords = ["GRZ", "GDC", "S3"]
 
+
+[tool.setuptools.dynamic]
+version = { attr = "grz_cli.__version__" }
+
 [project.urls]
 "Homepage" = "https://codebase.helmholtz.cloud/grz-mv-genomseq/grz-cli"
 
@@ -193,6 +198,10 @@ python_version = "3.12"
 plugins = "pydantic.mypy"
 packages = "src"
 
+[[tool.mypy.overrides]]
+module = ["grz_pydantic_models.*"]
+follow_untyped_imports = true
+
 [tool.pytest.ini_options]
 addopts = [
     "--import-mode=importlib",

grz_cli-0.3.0/src/grz_cli/__init__.py

@@ -0,0 +1 @@
+__version__ = "0.3.0"

{grz_cli-0.1.4 → grz_cli-0.3.0}/src/grz_cli/cli.py

@@ -2,6 +2,7 @@
 CLI module for handling command-line interface operations.
 """
 
+import importlib
 import logging
 import logging.config
 import sys
@@ -10,6 +11,7 @@ from pathlib import Path
 from tempfile import NamedTemporaryFile
 
 import click
+import grz_pydantic_models
 import platformdirs
 import yaml
 
@@ -107,7 +109,11 @@ class OrderedGroup(click.Group):
     cls=OrderedGroup,
     help="Validate, encrypt, decrypt and upload submissions to a GRZ/GDC.",
 )
-@click.version_option(version="0.1", prog_name="grz-cli")
+@click.version_option(
+    version=importlib.metadata.version("grz-cli"),
+    prog_name="grz-cli",
+    message=f"%(prog)s v%(version)s (metadata schema versions: {', '.join(grz_pydantic_models.get_supported_versions())})",
+)
 @click.option("--log-file", metavar="FILE", type=str, help="Path to log file")
 @click.option(
     "--log-level",
@@ -229,7 +235,7 @@ def decrypt(
         log.error("GRZ private key path is required for decryption.")
         sys.exit(1)
 
-    log.info("Starting encryption...")
+    log.info("Starting decryption...")
 
     submission_dir = Path(submission_dir)
 

{grz_cli-0.1.4 → grz_cli-0.3.0}/src/grz_cli/parser.py

@@ -9,13 +9,8 @@ from itertools import groupby
 from os import PathLike
 from pathlib import Path
 
-from pydantic import ValidationError
-
-from .download import S3BotoDownloadWorker
-from .fastq_validation import validate_paired_end_reads, validate_single_end_reads
-from .file_operations import Crypt4GH, calculate_sha256
-from .models.config import ConfigModel
-from .models.v1_0_0.metadata import (
+from grz_pydantic_models.v1_1_1.metadata import (
+    ChecksumType,
     File,
     FileType,
     GrzSubmissionMetadata,
@@ -23,7 +18,13 @@ from .models.v1_0_0.metadata import (
     SequenceData,
     SequencingLayout,
 )
-from .models.v1_0_0.metadata import File as SubmissionFileMetadata
+from grz_pydantic_models.v1_1_1.metadata import File as SubmissionFileMetadata
+from pydantic import ValidationError
+
+from .download import S3BotoDownloadWorker
+from .fastq_validation import validate_paired_end_reads, validate_single_end_reads
+from .file_operations import Crypt4GH, calculate_sha256
+from .models.config import ConfigModel
 from .progress_logging import FileProgressLogger
 from .states import DecryptionState, EncryptionState, ValidationState
 from .upload import S3BotoUploadWorker
@@ -168,6 +169,51 @@ class Submission:
 
         return retval
 
+    @staticmethod
+    def validate_file_data(metadata: File, local_file_path: Path) -> Generator[str]:
+        """
+        Validates whether the provided file matches this metadata.
+
+        :param metadata: Metadata model object
+        :param local_file_path: Path to the actual file (resolved if symlinked)
+        :return: Generator of errors
+        """
+        # Resolve file path
+        local_file_path = local_file_path.resolve()
+
+        # Check if path exists
+        if not local_file_path.exists():
+            yield f"{str(metadata.file_path)} does not exist!"
+            # Return here as following tests cannot work
+            return
+
+        # Check if path is a file
+        if not local_file_path.is_file():
+            yield f"{str(metadata.file_path)} is not a file!"
+            # Return here as following tests cannot work
+            return
+
+        # Check if the checksum is correct
+        if metadata.checksum_type == "sha256":
+            calculated_checksum = calculate_sha256(local_file_path)
+            if metadata.file_checksum != calculated_checksum:
+                yield (
+                    f"{str(metadata.file_path)}: Checksum mismatch! "
+                    f"Expected: '{metadata.file_checksum}', calculated: '{calculated_checksum}'."
+                )
+        else:
+            yield (
+                f"{str(metadata.file_path)}: Unsupported checksum type: {metadata.checksum_type}. "
+                f"Supported types: {[e.value for e in ChecksumType]}"
+            )
+
+        # Check file size
+        if metadata.file_size_in_bytes != local_file_path.stat().st_size:
+            yield (
+                f"{str(metadata.file_path)}: File size mismatch! "
+                f"Expected: '{metadata.file_size_in_bytes}', observed: '{local_file_path.stat().st_size}'."
+            )
+
     def validate_checksums(self, progress_log_file: str | PathLike) -> Generator[str]:
         """
         Validates the checksum of the files against the metadata and prints the errors.
@@ -185,7 +231,7 @@ class Submission:
             self.__log.debug("Validating '%s'...", str(local_file_path))
 
             # validate the file
-            errors = list(file_metadata.validate_data(local_file_path))
+            errors = list(self.validate_file_data(file_metadata, local_file_path))
             validation_passed = len(errors) == 0
 
             # return log state

{grz_cli-0.1.4 → grz_cli-0.3.0}/src/grz_cli/progress_logging.py

@@ -9,8 +9,9 @@ from collections.abc import Callable
 from os import PathLike
 from pathlib import Path
 
+from grz_pydantic_models.v1_1_1.metadata import File as SubmissionFileMetadata
+
 from .file_operations import read_multiple_json
-from .models.v1_0_0.metadata import File as SubmissionFileMetadata
 from .states import State
 
 Index = tuple[str, float, int]

{grz_cli-0.1.4 → grz_cli-0.3.0/src/grz_cli.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
-Metadata-Version: 2.2
+Metadata-Version: 2.4
 Name: grz-cli
-Version: 0.1.4
+Version: 0.3.0
 Summary: Tool for validation, encryption and upload of MV submissions to GDCs.
 Author-email: Koray Kirli <koraykirli@gmail.com>, Mathias Lesche <mathias.lesche@tu-dresden.de>, "Florian R. Hölzlwimmer" <git.ich@frhoelzlwimmer.de>, Till Hartmann <till.hartmann@bih-charite.de>, Thomas Sell <thomas.sell@bih-charite.de>
 License: MIT
@@ -19,6 +19,7 @@ Requires-Dist: pyyaml<7,>=6.0.2
 Requires-Dist: tqdm<5,>=4.66.5
 Requires-Dist: pydantic<2.10,>=2.9.2
 Requires-Dist: platformdirs<5,>=4.3.6
+Requires-Dist: grz-pydantic-models>=1.2.0
 Provides-Extra: test
 Requires-Dist: pytest<9,>=8.3.3; extra == "test"
 Requires-Dist: pytest-cov<6,>=5.0.0; extra == "test"

{grz_cli-0.1.4 → grz_cli-0.3.0}/src/grz_cli.egg-info/SOURCES.txt

@@ -22,9 +22,6 @@ src/grz_cli.egg-info/requires.txt
 src/grz_cli.egg-info/top_level.txt
 src/grz_cli/models/__init__.py
 src/grz_cli/models/config.py
-src/grz_cli/models/v1_0_0/__init__.py
-src/grz_cli/models/v1_0_0/metadata.py
-src/grz_cli/resources/grz-schema.json
 src/grz_cli/resources/thresholds.json
 tests/test_cli.py
 tests/test_download.py

{grz_cli-0.1.4 → grz_cli-0.3.0}/src/grz_cli.egg-info/requires.txt

@@ -6,6 +6,7 @@ pyyaml<7,>=6.0.2
 tqdm<5,>=4.66.5
 pydantic<2.10,>=2.9.2
 platformdirs<5,>=4.3.6
+grz-pydantic-models>=1.2.0
 
 [test]
 pytest<9,>=8.3.3

{grz_cli-0.1.4 → grz_cli-0.3.0}/tests/test_file_progress_logging.py

@@ -5,9 +5,9 @@ import json
 from pathlib import Path
 
 import pytest
+from grz_pydantic_models.v1_1_1.metadata import File as SubmissionFileMetadata
 
 from grz_cli.file_operations import read_multiple_json
-from grz_cli.models.v1_0_0.metadata import File as SubmissionFileMetadata
 from grz_cli.progress_logging import FileProgressLogger
 
 

{grz_cli-0.1.4 → grz_cli-0.3.0}/tests/test_parser.py

@@ -11,6 +11,7 @@ metadata_missing_read_order = "tests/mock_files/metadata_validation/missing-read
 metadata_missing_vcf_file = "tests/mock_files/metadata_validation/missing-vcf-file.json"
 metadata_missing_fastq_r2 = "tests/mock_files/metadata_validation/missing-fastq-r2.json"
 metadata_no_target_regions = "tests/mock_files/metadata_validation/missing-target-regions.json"
+metadata_incompatible_reference_genomes = "tests/mock_files/metadata_validation/incompatible-reference-genomes.json"
 
 
 def test_submission_metadata(temp_metadata_file_path):
@@ -33,9 +34,12 @@ def test_submission_metadata_fails():
     with pytest.raises(error_types, match="VCF file missing for lab datum"):
         SubmissionMetadata(metadata_missing_vcf_file)
 
-    with pytest.raises(error_types, match="Paired end sequencing layout but number of"):
+    with pytest.raises(error_types, match="Paired end sequencing layout but missing R2 file for flowcell id"):
         SubmissionMetadata(metadata_missing_fastq_r2)
 
+    with pytest.raises(error_types, match="Incompatible reference genomes found"):
+        SubmissionMetadata(metadata_incompatible_reference_genomes)
+
 
 def test_encrypted_submission():
     input_path = "/submission/files/a.fastq"