genelastic 0.7.0__py3-none-any.whl → 0.9.0__py3-none-any.whl

genelastic/import_data/data_file.py

@@ -1,87 +0,0 @@
-"""This module defines the DataFile class, which handles the representation,
-management, and extraction of metadata for a data file within a data bundle.
-
-It includes functionality to construct DataFile instances from paths and
-optional filename patterns, retrieve file paths and metadata, and support
-for extracting metadata from filenames using specified patterns.
-"""
-
-import logging
-import pathlib
-from pathlib import Path
-
-from genelastic.common import AnalysisMetaData
-
-from .filename_pattern import FilenamePattern
-
-logger = logging.getLogger("genelastic")
-
-
-class DataFile:
-    """Class for handling a data file and its metadata."""
-
-    # Initializer
-    def __init__(
-        self,
-        path: Path,
-        bundle_path: Path | None = None,
-        metadata: AnalysisMetaData | None = None,
-    ) -> None:
-        self._path = path
-        self._bundle_path = bundle_path  # The bundle YAML file in which this
-        # file was listed.
-        self._metadata = {} if metadata is None else metadata
-
-    def __repr__(self) -> str:
-        return f"File {self._path}, from bundle {self._bundle_path}, with metadata {self._metadata}"
-
-    # Get path
-    @property
-    def path(self) -> Path:
-        """Retrieve the data file path."""
-        return self._path
-
-    def exists(self) -> bool:
-        """Tests if the associated file exists on disk."""
-        return self._path.is_file()
-
-    # Get bundle path
-    @property
-    def bundle_path(self) -> Path | None:
-        """Retrieve the path to the associated data bundle file."""
-        return self._bundle_path
-
-    # Get metadata
-    @property
-    def metadata(self) -> AnalysisMetaData:
-        """Retrieve a copy of the metadata associated with the data file."""
-        return self._metadata.copy()
-
-    # Factory
-    @classmethod
-    def make_from_bundle(
-        cls,
-        path: Path,
-        bundle_path: Path | None,
-        pattern: FilenamePattern | None = None,
-    ) -> "DataFile":
-        """Construct a DataFile instance from a bundle path, file path,
-        and optional filename pattern.
-        """
-        # Make absolute path
-        if not path.is_absolute() and bundle_path is not None:
-            path = bundle_path.parent / path
-
-        # Extract filename metadata
-        metadata = None
-        if pattern is not None:
-            metadata = pattern.extract_metadata(path.name)
-
-        if metadata:
-            if "ext" not in metadata:
-                metadata["ext"] = pathlib.Path(path).suffixes[0][1:]
-
-            if "cov_depth" in metadata:
-                metadata["cov_depth"] = int(metadata["cov_depth"])
-
-        return cls(path, bundle_path, metadata)

genelastic/import_data/filename_pattern.py

@@ -1,57 +0,0 @@
-"""This module defines the FilenamePattern class, used to define a filename pattern
-and extract metadata from file names using this pattern.
-"""
-
-import re
-
-from genelastic.common import AnalysisMetaData
-
-
-class FilenamePattern:
-    """Class for defining a filename pattern.
-    The pattern is used to extract metadata from filenames
-    and verify filename conformity.
-    """
-
-    # Initializer
-    def __init__(self, pattern: str) -> None:
-        """Initializes a FilenamePattern instance.
-
-        Args:
-            pattern (str): The pattern string used for defining
-             the filename pattern.
-        """
-        self._re = re.compile(pattern)
-
-    def extract_metadata(self, filename: str) -> AnalysisMetaData:
-        """Extracts metadata from the given filename based
-        on the defined pattern.
-
-        Args:
-            filename (str): The filename from which metadata
-            needs to be extracted.
-
-        Returns:
-            dict: A dictionary containing the extracted metadata.
-
-        Raises:
-            RuntimeError: If parsing of filename fails
-            with the defined pattern.
-        """
-        m = self._re.search(filename)
-        if not m:
-            msg = f'Failed parsing filename "{filename}" with pattern "{self._re.pattern}".'
-            raise RuntimeError(msg)
-        return m.groupdict()
-
-    def matches_pattern(self, filename: str) -> bool:
-        """Checks if the given filename matches the defined pattern.
-
-        Args:
-            filename (str): The filename to be checked.
-
-        Returns:
-            bool: True if the filename matches the pattern,
-            False otherwise.
-        """
-        return bool(self._re.match(filename))

genelastic/import_data/tags.py

@@ -1,123 +0,0 @@
-import logging
-import re
-import typing
-
-from genelastic.common import BundleDict
-
-from .constants import DEFAULT_TAG2FIELD, DEFAULT_TAG_PREFIX, DEFAULT_TAG_SUFFIX
-
-logger = logging.getLogger("genelastic")
-
-TagsDefinition: typing.TypeAlias = dict[str, dict[str, str | dict[str, str]]]
-
-
-class Tags:
-    """This class handles the definition of default and custom tags.
-    Tags are used to extract custom metadata from files belonging to an analysis.
-    """
-
-    def __init__(self, documents: typing.Sequence[BundleDict] | None) -> None:
-        """Create a Tag instance."""
-        self._tags: dict[str, dict[str, str]] = DEFAULT_TAG2FIELD
-        self._tag_prefix: str = DEFAULT_TAG_PREFIX
-        self._tag_suffix: str = DEFAULT_TAG_SUFFIX
-
-        redefined_tags = None
-
-        if documents:
-            # Search for tags definition across loaded YAML documents.
-            redefined_tags = self._search_redefined_tags(documents)
-
-        if redefined_tags:
-            self._build_tags(redefined_tags)
-            logger.info(
-                "The following tags will be used to extract metadata from filenames : %s",
-                self._tags,
-            )
-        else:
-            logger.info(
-                "Using the default tags to extract metadata from filenames : %s",
-                self._tags,
-            )
-
-    def _build_tags(self, redefined_tags: TagsDefinition) -> None:
-        # Erase the tags defined by defaults.
-        self._tags = {}
-
-        if "format" in redefined_tags:
-            tag_format = redefined_tags["format"]
-
-            # extra type check for mypy
-            if "prefix" in tag_format and isinstance(tag_format["prefix"], str):
-                self._tag_prefix = tag_format["prefix"]
-
-            # extra type check for mypy
-            if "suffix" in tag_format and isinstance(tag_format["suffix"], str):
-                self._tag_suffix = tag_format["suffix"]
-
-        for tag_name, tag_attrs in redefined_tags["match"].items():
-            if isinstance(tag_attrs, dict):  # extra type check for mypy
-                self._tags[
-                    f"{self._tag_prefix}{tag_name}{self._tag_suffix}"
-                ] = tag_attrs
-
-    @staticmethod
-    def _search_redefined_tags(
-        documents: typing.Sequence[BundleDict],
-    ) -> TagsDefinition | None:
-        documents_with_redefined_tags: list[BundleDict] = [
-            d for d in documents if "tags" in d
-        ]
-        bundle_paths = [d["bundle_file"] for d in documents_with_redefined_tags]
-
-        # If there are more than one 'tags' redefinition across the documents, raise an error.
-        if len(documents_with_redefined_tags) > 1:
-            msg = (
-                f"Only one 'tags' key should be defined across all documents, "
-                f"but multiple were found : {', '.join(bundle_paths)}"
-            )
-            raise RuntimeError(msg)
-
-        if len(documents_with_redefined_tags) == 1:
-            redefined_tags: TagsDefinition = documents_with_redefined_tags[0][
-                "tags"
-            ]
-            return redefined_tags
-
-        return None
-
-    @property
-    def tag_prefix(self) -> str:
-        """Return the tag prefix. Default prefix is '%'."""
-        return self._tag_prefix
-
-    @property
-    def tag_suffix(self) -> str:
-        """Return the tag suffix. There is no suffix by default."""
-        return self._tag_suffix
-
-    @property
-    def items(self) -> typing.ItemsView[str, dict[str, str]]:
-        """Returns the tag items : the key is the tag name,
-        and the value is the tag attributes (a dict containing the 'field' and 'regex' keys).
-        """
-        return self._tags.items()
-
-    @property
-    def search_regex(self) -> str:
-        """Returns a regex to search for a tag inside a string."""
-        return (
-            r"("
-            + re.escape(self._tag_prefix)
-            + r"\w+"
-            + re.escape(self._tag_suffix)
-            + r")"
-        )
-
-    def __len__(self) -> int:
-        """Return the number of registered tags."""
-        return len(self._tags)
-
-    def __getitem__(self, key: str) -> dict[str, str]:
-        """Return a tag by its key."""
-        return self._tags[key]

genelastic/import_data/wet_process.py

@@ -1,28 +0,0 @@
-import copy
-
-from genelastic.common import WetProcessesData
-
-
-class WetProcess:
-    """Class WetProcess that represents a wet process."""
-
-    def __init__(
-        self,
-        proc_id: str,
-        bundle_file: str | None = None,
-        **data: str | float,
-    ) -> None:
-        """Create a WetProcess instance."""
-        self._proc_id = proc_id
-        self._bundle_file = bundle_file
-        self._data: WetProcessesData = data
-
-    @property
-    def id(self) -> str:
-        """Get the wet process ID."""
-        return self._proc_id
-
-    @property
-    def data(self) -> WetProcessesData:
-        """Get data associated to the wet process."""
-        return copy.deepcopy(self._data)

genelastic/import_data/wet_processes.py

@@ -1,53 +0,0 @@
-import logging
-import typing
-
-from genelastic.common import BundleDict
-
-from .wet_process import WetProcess
-
-logger = logging.getLogger("genelastic")
-
-
-class WetProcesses:
-    """Class WetProcesses is a container of WetProces objects."""
-
-    def __init__(self) -> None:
-        """Create an empty container."""
-        self._dict: dict[str, WetProcess] = {}
-
-    def __len__(self) -> int:
-        """Return the number of WetProcess objects inside the container."""
-        return len(self._dict)
-
-    def __getitem__(self, key: str) -> WetProcess:
-        """Return a WetProcess present in the container by its key."""
-        return self._dict[key]
-
-    def add(self, process: WetProcess) -> None:
-        """Add one WetProces object.
-        If a WetProces object with the same ID already exists in the container, the program exits.
-        """
-        if process.id in self._dict:
-            msg = (
-                f"A wet process with the id '{process.id}' is already present."
-            )
-            raise ValueError(msg)
-
-        # Add one WetProcess object.
-        self._dict[process.id] = process
-
-    def get_process_ids(self) -> set[str]:
-        """Get a list of the wet processes IDs."""
-        return set(self._dict.keys())
-
-    @classmethod
-    def from_array_of_dicts(
-        cls, arr: typing.Sequence[BundleDict]
-    ) -> typing.Self:
-        """Build a WetProcesses instance."""
-        wet_processes = cls()
-
-        for d in arr:
-            wet_processes.add(WetProcess(**d))
-
-        return wet_processes

genelastic-0.7.0.dist-info/METADATA

@@ -1,105 +0,0 @@
-Metadata-Version: 2.2
-Name: genelastic
-Version: 0.7.0
-Summary: Generate and store genetic data into an Elasticsearch database.
-Author: CNRGH
-Author-email: Pierrick ROGER <pierrick.roger@cnrgh.fr>, Maxime BLANCHON <maxime.blanchon@cnrgh.fr>
-License: CeCILL
-Keywords: CNRGH,genelastic,generation,storage,elasticsearch,database
-Classifier: Development Status :: 3 - Alpha
-Classifier: Intended Audience :: Science/Research
-Classifier: License :: OSI Approved :: CEA CNRS Inria Logiciel Libre License, version 2.1 (CeCILL-2.1)
-Classifier: Programming Language :: Python :: 3.11
-Requires-Python: >=3.11
-Description-Content-Type: text/markdown
-Requires-Dist: elasticsearch
-Requires-Dist: PyVCF3
-Requires-Dist: schema
-Requires-Dist: PyYAML
-Requires-Dist: biophony>=1.2.1
-Requires-Dist: colorlog
-Provides-Extra: tests
-Requires-Dist: pytest; extra == "tests"
-Requires-Dist: mypy; extra == "tests"
-Requires-Dist: coverage; extra == "tests"
-Requires-Dist: yamllint; extra == "tests"
-Requires-Dist: types-PyYAML; extra == "tests"
-Requires-Dist: ruff<0.9,>=0.8.1; extra == "tests"
-Requires-Dist: pre-commit<4.1,>=4.0.1; extra == "tests"
-Requires-Dist: types-requests; extra == "tests"
-Requires-Dist: ansible<10.5,>=10.4.0; extra == "tests"
-Requires-Dist: ansible-lint<25,>=24.12.2; extra == "tests"
-Provides-Extra: docs
-Requires-Dist: sphinx; extra == "docs"
-Requires-Dist: sphinx-autoapi; extra == "docs"
-Requires-Dist: furo; extra == "docs"
-Provides-Extra: api
-Requires-Dist: flask; extra == "api"
-Requires-Dist: elasticsearch; extra == "api"
-Requires-Dist: environs; extra == "api"
-Requires-Dist: connexion[flask,swagger-ui,uvicorn]; extra == "api"
-Provides-Extra: ui
-Requires-Dist: flask; extra == "ui"
-Requires-Dist: requests; extra == "ui"
-Requires-Dist: environs; extra == "ui"
-
-# genelastic
-
-Storing of genetics data into an Elasticsearch database.
-
-## Prerequisites
-
-- `python` >= 3.11
-- `make`
-
-## Installation
-
-To install dependencies, run the following command:
-
-```bash
-python -m venv .venv
-source .venv/bin/activate
-make install.deps
-```
-
-## Configuration
-
-To start the **API server**, the following environment variables should be defined:
-
-- `GENAPI_ES_URL`: URL of the Elasticsearch server,
-- `GENAPI_ES_ENCODED_API_KEY`: Encoded API key,
-- `GENAPI_ES_INDEX_PREFIX`: Prefix to identify indices of interest,
-- `GENAPI_ES_CERT_FP`: Certificate fingerprint of the Elasticsearch server.
-
-Then, run the following command:
-
-```bash
-make start-api
-```
-
-To start the **UI server**, the following environment variables should be defined:
-
-- `GENUI_API_URL`: URL of the API server.
-
-Then, run the following command:
-
-```bash
-make start-ui
-```
-
-## Developers
-
-This project uses [pre-commit](https://pre-commit.com/) to manage Git hooks scripts. To install project hooks, run:
-
-```bash
-pre-commit install
-```
-
-After that, each commit will succeed only if all hooks (defined in `.pre-commit-config.yaml`) pass.
-
-If necessary (though not recommended),
-you can skip these hooks by using the `--no-verify` / `-n` option when committing:
-
-```bash
-git commit -m "My commit message" --no-verify # This commit will not run installed hooks.
-```

genelastic-0.7.0.dist-info/RECORD

@@ -1,40 +0,0 @@
-genelastic/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-genelastic/api/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-genelastic/api/routes.py,sha256=O_bZGhKex04tWauE1fYRXjNmR2smIA4JZShwcakbJJA,6938
-genelastic/api/server.py,sha256=XkkHpv_28OrfrY6CMDnImBFlDiEQSA8f1BpSSn2LGL4,2248
-genelastic/api/settings.py,sha256=vLfj8ASLqq9f6c3eGg30L_WSBiaOIpnCsk-S-TUZRlk,403
-genelastic/api/extends/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-genelastic/api/extends/example.py,sha256=5vOw2ToYWzOj-vYkOLbqoa1MCnXIvKuv7xW6qBQ-eHU,164
-genelastic/common/__init__.py,sha256=GfM4XaiaHsA2v8N0T3VgnXpZsLXjTc1wIPpRi4BlkZ8,937
-genelastic/common/cli.py,sha256=cDbgc2m20Y1NO0CZR1qaTrCttcf0cjFAP885tuPZsH0,1770
-genelastic/common/elastic.py,sha256=0XpECyNMC9RzQN1X7u-0IXmHp_KbSjkmGizArLjxTS0,7260
-genelastic/common/exceptions.py,sha256=l4sQBW91OuzV2R7DhrVzEMDrfsyKM-iVYufp2Cv6rR8,159
-genelastic/common/types.py,sha256=JeQ6XzIemIF1QgvprGA0UxFHyJSKjEkgMaIumOvpbNg,1013
-genelastic/import_data/__init__.py,sha256=Ir_fI0BrdC44JiiQaWm29_SyHbDe2CQsElKHgQxELCI,602
-genelastic/import_data/analyses.py,sha256=yA4DKtXIUR7JTCp7PEndEjI6FOooW-6lR9vkqTZ0Izs,1861
-genelastic/import_data/analysis.py,sha256=MuHYfIWFFwHfCp9tDkWTtVKkg5AasASwrdOUZ177EeM,7799
-genelastic/import_data/bi_process.py,sha256=CBBGx6oO7KYzUypXzpaopGY4J_nAKcDi2UPJe7B63Bs,650
-genelastic/import_data/bi_processes.py,sha256=Kv93NOntycEiB_BldoMxHNfM6sJvgHQvVqrUSSEb_VA,1365
-genelastic/import_data/cli_gen_data.py,sha256=WYvSgiC30pDvhQuGMxtDKookUF9IQpjQP6UMANvvHSU,3012
-genelastic/import_data/cli_import.py,sha256=Ej-EqaBUNoVHGCefzLsmFbKaXNTaOBs8xJsEdRhbbLo,11622
-genelastic/import_data/cli_info.py,sha256=ue0Pf5cJvrmJ_bBZKG2CXkXDsQO3zQ6CIXhtUwpqXhU,7301
-genelastic/import_data/cli_integrity.py,sha256=lOMA-I1iPXiJD4X80Xg91NKO-fa3Su0uSqBUtdMKU4Q,12282
-genelastic/import_data/cli_validate.py,sha256=AcirmmJQWFDrpiPUNTYnpFjSnvD6YyStYJHIbfi1Ly0,1515
-genelastic/import_data/constants.py,sha256=AKYXdDqWkDzvt-laZqbWlN6C9IoQSSqD70mpSWMU6NQ,760
-genelastic/import_data/data_file.py,sha256=P5oe_yCVQQAALetrUs93Kl1XuqoLQS0iMkeEqPKaY7g,2661
-genelastic/import_data/filename_pattern.py,sha256=3QXOF5ZG9vFA8KG7XXucpA7EIN3dqZmmtXnuFbh0bUQ,1709
-genelastic/import_data/import_bundle.py,sha256=7W0hKwoxwGDtWEGV_NQvyesiNx1lCMpSaqKohpTLmtY,5030
-genelastic/import_data/import_bundle_factory.py,sha256=COdMLTTrx_Y4svZAHVJ5Y0rHM4eQj027rFJiQ66u36M,9546
-genelastic/import_data/logger.py,sha256=X12LBoNTmV6iR8vEPyYIMYSUeQ3LpTvDO69tQYDiOuA,1915
-genelastic/import_data/random_bundle.py,sha256=Fk4oudK42pboGI14aaJjh8tTJxSxYNyF_89BBJ532RY,13821
-genelastic/import_data/tags.py,sha256=N6_dGYqQy2QTN6AEzoEXPxRFmtq3GgTfpId-RjBstJY,4195
-genelastic/import_data/wet_process.py,sha256=2SgN1yZFQA8Hb4ZNNeHpUhmeWlJDApqdjc_oq6ZF0jA,694
-genelastic/import_data/wet_processes.py,sha256=mVUfFG0QUboTEoD3c-Hd1Z2_Tvid_yrKJG_XdC4n6oA,1535
-genelastic/ui/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-genelastic/ui/server.py,sha256=7QX-54tT7CbpKje34kVVXnktI0wuZJOSLUChxeUGd20,2805
-genelastic/ui/settings.py,sha256=hmgDNMVKk_OkKSBKyV8Eth57ULCTINwzeBCFAH7DdeU,218
-genelastic-0.7.0.dist-info/METADATA,sha256=dj4cFcq7LXEv9v-turnGvHnK2KZ4zXrDxG0YKe11zPU,3127
-genelastic-0.7.0.dist-info/WHEEL,sha256=In9FTNxeP60KnTkGw7wk6mJPYd_dQSjEZmXdBdMCI-8,91
-genelastic-0.7.0.dist-info/entry_points.txt,sha256=KHtfefg-7-v0MZ0jAcHrb6z3V6hq7UdCiYYIAlD6hIg,284
-genelastic-0.7.0.dist-info/top_level.txt,sha256=ra4gCsuKH1d0sXygcnwD_u597ir6bYYxWTS7dkA6vdM,11
-genelastic-0.7.0.dist-info/RECORD,,

genelastic-0.7.0.dist-info/WHEEL

@@ -1,5 +0,0 @@
-Wheel-Version: 1.0
-Generator: setuptools (75.8.0)
-Root-Is-Purelib: true
-Tag: py3-none-any
-

genelastic-0.7.0.dist-info/entry_points.txt

@@ -1,6 +0,0 @@
-[console_scripts]
-gnl-data = genelastic.import_data.cli_gen_data:main
-gnl-import = genelastic.import_data.cli_import:main
-gnl-info = genelastic.import_data.cli_info:main
-gnl-integrity = genelastic.import_data.cli_integrity:main
-gnl-validate = genelastic.import_data.cli_validate:main

genelastic-0.7.0.dist-info/top_level.txt

@@ -1 +0,0 @@
-genelastic