Merge pull request #15 from moka-guys/development

v1.2.0 (#15)

Co-Authored-By: Graeme <graeme_c_smith@hotmail.com>
Co-Authored-By: rebeccahaines1 <84131466+rebeccahaines1@users.noreply.github.com>

Author: Graeme-Smith

README.md

@@ -19,6 +19,8 @@ Runs a series of checks on the sample sheet, collects any errors identified. Che
 * Pan numbers are in the list of allowed pan numbers supplied to the script
 * Samplesheet contains any TSO samples
 
+If samplesheet contains an input dev_panno, the package will skip samplesheet chcks for the samplesheet.
+
 ## Usage
 
 ### Python package
@@ -75,10 +77,16 @@ options:
                         Development pan number
   -L LOGDIR, --logdir LOGDIR
                         Directory to save the output logfile to
+  -NSH NO_STREAM_HANDLER, --no_stream_handler NO_STRAM_HANDLER
+                        Provide flag when we don't want a stream handler (prevents
+                        duplication of log messages to terminal if using another
+                        logging instance)
 ```
 
 ### Testing
 
+This repository currently has **92% test coverage**.
+
 Test datasets are stored in [/test/data](../test/data). The script has a full test suite:
 * [test_samplesheet_validator.py](../test/test_samplesheet_validator.py)
 

samplesheet_validator/__main__.py

@@ -1,6 +1,7 @@
 import os
 import argparse
 from .samplesheet_validator import SamplesheetCheck
+from .ss_logger import set_root_logger
 
 
 def get_arguments():
@@ -58,6 +59,16 @@ def get_arguments():
         required=True,
         help="Directory to save the output logfile to",
     )
+    parser.add_argument(
+        "-NSH",
+        "--no_stream_handler",
+        action='store_true',
+        required=False,
+        help=(
+            "Provide flag when we don't want a stream handler (prevents duplication of log messages "
+            "to terminal if using another logging instance)"
+        ),
+    )
     return parser.parse_args()
 
 
@@ -89,6 +100,8 @@ def is_valid_dir(parser: argparse.ArgumentParser, dir: str) -> str:
 
 if __name__ == "__main__":
     parsed_args = get_arguments()
+    if not parsed_args.no_stream_handler:
+        set_root_logger()  # Adds stream handler
     sscheck_obj = SamplesheetCheck(
         parsed_args.samplesheet_path,
         parsed_args.sequencer_ids,

samplesheet_validator/git_tag.py

@@ -0,0 +1,18 @@
+import os
+import subprocess
+
+
+def git_tag() -> str:
+    """
+    Obtain the git tag of the current commit
+        :return (str):  Git tag
+    """
+    filepath = os.path.dirname(os.path.realpath(__file__))
+    cmd = f"git -C {filepath} describe --tags"
+
+    proc = subprocess.Popen(
+        [cmd], stderr=subprocess.PIPE, stdout=subprocess.PIPE, shell=True,
+    )
+    out, _ = proc.communicate()
+    #  Return standard out, removing any new line characters
+    return out.rstrip().decode("utf-8")

samplesheet_validator/samplesheet_validator.py

@@ -14,7 +14,7 @@
 import re
 from typing import Union
 from . import config
-from .ss_logger import SSLogger, shutdown_logs
+from .ss_logger import SSLogger
 from seglh_naming.sample import Sample
 from seglh_naming.samplesheet import Samplesheet
 
@@ -101,7 +101,6 @@ def __init__(
         """
         self.samplesheet_path = samplesheet_path
         self.logdir = logdir
-        self.logger = False
         self.ss_obj = False
         self.pannumbers = []
         self.tso = False
@@ -156,7 +155,6 @@ def ss_checks(self) -> None:
                                     self.check_pannos(sample, column, sample_obj)
                         self.check_tso()
         self.log_summary()
-        shutdown_logs(self.logger)
 
     def check_ss_present(self) -> Union[bool, None]:
         """

samplesheet_validator/ss_logger.py

@@ -8,20 +8,23 @@
 import logging.handlers
 
 
-def shutdown_logs(logger: object) -> None:
+def set_root_logger():
     """
-    To prevent duplicate filehandlers and system handlers close
-    and remove all handlers for a logging object
-        :return (None):
+    Set up root logger and add stream handler - we only want to add stream handler once
+    else it will duplicate log messages to the terminal
     """
-    for handler in logger.handlers[:]:
-        logger.removeHandler(handler)
-        handler.close()
+    logger = logging.getLogger()
+    logger.setLevel(logging.DEBUG)
+    stream_handler = logging.StreamHandler(sys.stdout)
+    stream_handler.setLevel(logging.DEBUG)
+    stream_handler.setFormatter(logging.Formatter(config.LOGGING_FORMATTER))
+    stream_handler.name = "stream_handler"
+    logger.addHandler(stream_handler)
 
 
 class SSLogger:
     """
-    Creates a python logging object with a file handler, syslog handler and stream handler
+    Creates a python logging object with a file handler and syslog handler
 
     Attributes
         timestamp (str):                        Timestamp from config
@@ -35,8 +38,6 @@ class SSLogger:
             Get file handler for the logger
         _get_syslog_handler()
             Get syslog handler for the logger
-        _get_stream_handler()
-            Get stream handler for the logger (sends to stdout)
     """
 
     def __init__(self, logfile_path: str):
@@ -58,7 +59,6 @@ def get_logger(self) -> logging.Logger:
         logger.filepath = self.logfile_path
         logger.setLevel(logging.DEBUG)
         logger.addHandler(self._get_file_handler())
-        logger.addHandler(self._get_stream_handler())
         logger.addHandler(self._get_syslog_handler())
         logger.timestamp = self.timestamp
         logger.log_msgs = config.LOG_MSGS
@@ -85,14 +85,3 @@ def _get_syslog_handler(self) -> logging.handlers.SysLogHandler:
         syslog_handler.setFormatter(self.logging_formatter)
         syslog_handler.name = "syslog_handler"
         return syslog_handler
-
-    def _get_stream_handler(self) -> logging.StreamHandler:
-        """
-        Get stream handler for the logger (sends to stdout)
-            :return stream_handler (logging.StreamHandler): StreamHandler
-        """
-        stream_handler = logging.StreamHandler(sys.stdout)
-        stream_handler.setLevel(logging.DEBUG)
-        stream_handler.setFormatter(self.logging_formatter)
-        stream_handler.name = "stream_handler"
-        return stream_handler

setup.py

@@ -1,10 +1,11 @@
 from setuptools import setup
+from samplesheet_validator.git_tag import git_tag
 
 setup(
     name="samplesheet_validator",
-    version="1.1.0",
+    version=git_tag(),
     description="Python library for samplesheet validation",
-    url="http://github.com/moka-guys/samplesheet_validator",
+    url="https://github.com/moka-guys/samplesheet_validator",
     author="Rachel Duffin",
     author_email="rachel.duffin2@nhs.net",
     license="Apache License, Version 2.0",

test/test_samplesheet_validator.py

@@ -6,11 +6,21 @@
 import os
 import argparse
 import pytest
-from samplesheet_validator.ss_logger import shutdown_logs
 from samplesheet_validator import samplesheet_validator
 from samplesheet_validator.__main__ import is_valid_dir, is_valid_file
 
 
+def shutdown_logs(logger: object) -> None:
+    """
+    To prevent duplicate filehandlers and system handlers close
+    and remove all handlers for a logging object
+        :return (None):
+    """
+    for handler in logger.handlers[:]:
+        logger.removeHandler(handler)
+        handler.close()
+
+
 def get_sscheck_obj(samplesheet: str) -> object:
     """
     Function to retrieve a samplesheet check object and carry out the