feat: explicit check for local file existence in arg parsing

pydeeptools/deeptools/alignmentSieve2.py

@@ -5,6 +5,7 @@
 
 from deeptools import parserCommon
 from deeptools.hp import r_alignmentsieve
+from deeptools.parserCommon import existingFile
 
 
 def parseArguments():
@@ -18,6 +19,7 @@ def parseArguments():
     required.add_argument('--bam', '-b',
                           metavar='FILE1',
                           help='An indexed BAM file.',
+                          type=existingFile,
                           required=True)
 
     required.add_argument('--outFile', '-o',
@@ -122,6 +124,7 @@ def parseArguments():
                            metavar="BED file",
                            nargs="+",
                            default="None",
+                           type=existingFile,
                            required=False)
 
     filtering.add_argument('--ignoreDuplicates',

pydeeptools/deeptools/bamCompare2.py

@@ -4,6 +4,7 @@
 
 from deeptools import parserCommon
 from deeptools.hp import r_bamcompare
+from deeptools.parserCommon import existingFile
 
 
 def parseArguments():
@@ -51,12 +52,14 @@ def getRequiredArgs():
                           metavar='BAM file',
                           help='Sorted BAM file 1. Usually the BAM file '
                           'for the treatment.',
+                          type=existingFile,
                           required=True)
 
     required.add_argument('--bamfile2', '-b2',
                           metavar='BAM file',
                           help='Sorted BAM file 2. Usually the BAM '
                           'file for the control.',
+                          type=existingFile,
                           required=True)
 
     return parser

pydeeptools/deeptools/bamCoverage2.py

@@ -4,6 +4,7 @@
 
 from deeptools import parserCommon
 from deeptools.hp import r_bamcoverage
+from deeptools.parserCommon import existingFile
 
 
 def parseArguments():
@@ -45,6 +46,7 @@ def get_required_args():
     required.add_argument('--bam', '-b',
                           help='BAM file to process',
                           metavar='BAM file',
+                          type=existingFile,
                           required=True)
 
     return parser

pydeeptools/deeptools/bamPEFragmentSize.py

@@ -8,7 +8,7 @@
 
 from deeptools import matplotlib_defaults  # noqa: F401
 from deeptools.getFragmentAndReadSize import get_read_and_fragment_length
-from deeptools.parserCommon import writableFile
+from deeptools.parserCommon import existingFile, writableFile
 
 
 def parse_arguments():
@@ -30,6 +30,7 @@ def parse_arguments():
         help="List of BAM files to process",
         nargs="+",
         metavar="bam files",
+        type=existingFile,
     )
 
     parser.add_argument(
@@ -114,6 +115,7 @@ def parse_arguments():
         "-bl",
         help="A BED file containing regions that should be excluded from all analyses. Currently this works by rejecting genomic chunks that happen to overlap an entry. Consequently, for BAM files, if a read partially overlaps a blacklisted region or a fragment spans over it, then the read/fragment might still be considered.",
         metavar="BED file",
+        type=existingFile,
         required=False,
     )
     parser.add_argument(

pydeeptools/deeptools/bigwigAverage.py

@@ -4,6 +4,7 @@
 import numpy as np
 
 from deeptools import parserCommon, writeBedGraph_bam_and_bw
+from deeptools.parserCommon import existingFile
 
 debug = 0
 
@@ -30,6 +31,7 @@ def parse_arguments(args=None):
         metavar="Bigwig files",
         help="Bigwig files separated by space.",
         nargs="+",
+        type=existingFile,
         required=True,
     )
 

pydeeptools/deeptools/bigwigCompare.py

@@ -2,6 +2,7 @@
 
 from deeptools import parserCommon, writeBedGraph_bam_and_bw
 from deeptools.getRatio import getRatio
+from deeptools.parserCommon import existingFile
 
 debug = 0
 
@@ -26,12 +27,14 @@ def parse_arguments(args=None):
                         metavar='Bigwig file',
                         help='Bigwig file 1. Usually the file for the '
                         'treatment.',
+                        type=existingFile,
                         required=True)
 
     parser.add_argument('--bigwig2', '-b2',
                         metavar='Bigwig file',
                         help='Bigwig file 2. Usually the file for the '
                         'control.',
+                        type=existingFile,
                         required=True)
 
     parser.add_argument('--scaleFactors',

pydeeptools/deeptools/computeMatrix2.py

@@ -5,7 +5,7 @@
 
 from deeptools import parserCommon
 from deeptools.hp import r_computematrix
-from deeptools.parserCommon import numberOfProcessors, writableFile
+from deeptools.parserCommon import existingFile, numberOfProcessors, writableFile
 
 
 def parse_arguments(args=None):
@@ -86,6 +86,7 @@ def computeMatrixRequiredArgs(args=None):
         "the regions to plot. If multiple bed files are given, each one is considered a "
         "group that can be plotted separately.",
         nargs="+",
+        type=existingFile,
         required=True,
     )
     required.add_argument(
@@ -99,6 +100,7 @@ def computeMatrixRequiredArgs(args=None):
         "http://genome.ucsc.edu/goldenPath/help/bigWig.html ",
         metavar="File",
         nargs="+",
+        type=existingFile,
         required=True,
     )
     return parser
@@ -369,7 +371,8 @@ def computeMatrixOptArgs(case=["scale-regions", "reference-point"][0]):
         "-bl",
         help="A BED file (optionally gzip-compressed) containing regions that should be excluded from all analyses. Filtering is performed at base-pair resolution, so only the portion of a region that overlaps a blacklisted region is excluded.",
         metavar="BED file",
-        default='none',
+        default='None',
+        type=existingFile,
         required=False,
     )
 

pydeeptools/deeptools/computeMatrixOperations.py

@@ -10,6 +10,7 @@
 import numpy as np
 
 from deeptools import heatmapper
+from deeptools.parserCommon import existingFile
 
 
 def parse_arguments():
@@ -153,6 +154,7 @@ def bindArgs():
     required.add_argument('--matrixFile', '-m',
                           help='Matrix files from the computeMatrix tool.',
                           nargs='+',
+                          type=existingFile,
                           required=True)
 
     required.add_argument('--outFileName', '-o',
@@ -168,6 +170,7 @@ def infoArgs():
 
     required.add_argument('--matrixFile', '-m',
                           help='Matrix file from the computeMatrix tool.',
+                          type=existingFile,
                           required=True)
 
     return parser
@@ -259,6 +262,7 @@ def sortArgs():
 
     required.add_argument('--matrixFile', '-m',
                           help='Matrix file from the computeMatrix tool.',
+                          type=existingFile,
                           required=True)
 
     required.add_argument('--outFileName', '-o',
@@ -276,7 +280,8 @@ def sortArgs():
                                'column with group labels. Note that these should be '
                                'sorted such that all group entries are together.',
                           required=True,
-                          nargs='+')
+                          nargs='+',
+                          type=existingFile)
 
     optional = parser.add_argument_group('Optional arguments')
 

pydeeptools/deeptools/estimateReadFiltering.py

@@ -5,6 +5,7 @@
 
 from deeptools import bamHandler, parserCommon, utilities
 from deeptools.mapReduce import mapReduce
+from deeptools.parserCommon import existingFile
 from deeptools.utilities import smartLabels
 
 
@@ -40,6 +41,7 @@ def parseArguments():
                           metavar='FILE1 FILE2',
                           help='List of indexed bam files separated by spaces.',
                           nargs='+',
+                          type=existingFile,
                           required=True)
 
     general = parser.add_argument_group('General arguments')
@@ -147,6 +149,7 @@ def parseArguments():
                            help="A BED or GTF file containing regions that should be excluded from all analyses. Currently this works by rejecting genomic chunks that happen to overlap an entry. Consequently, for BAM files, if a read partially overlaps a blacklisted region or a fragment spans over it, then the read/fragment might still be considered. Please note that you should adjust the effective genome size, if relevant.",
                            metavar="BED file",
                            nargs="+",
+                           type=existingFile,
                            required=False)
 
     return parser

pydeeptools/deeptools/estimateScaleFactor.py

@@ -2,7 +2,7 @@
 import sys
 from importlib.metadata import version
 
-from deeptools.parserCommon import numberOfProcessors
+from deeptools.parserCommon import existingFile, numberOfProcessors
 from deeptools.SES_scaleFactor import estimateScaleFactor
 
 debug = 0
@@ -22,6 +22,7 @@ def parseArguments(args=None):
                         metavar='list of bam files',
                         help='List of indexed BAM files, space delineated',
                         nargs='+',
+                        type=existingFile,
                         required=True)
 
     parser.add_argument('--ignoreForNormalization', '-ignore',

pydeeptools/deeptools/multiBamSummary2.py

@@ -6,6 +6,7 @@
 
 from deeptools import parserCommon
 from deeptools.hp import r_mbams
+from deeptools.parserCommon import existingFile
 from deeptools.utilities import smartLabels
 
 
@@ -108,6 +109,7 @@ def bamcorrelate_args(case="bins"):
         metavar="FILE1 FILE2",
         help="List of indexed bam files separated by spaces.",
         nargs="+",
+        type=existingFile,
         required=True,
     )
 
@@ -189,6 +191,7 @@ def bamcorrelate_args(case="bins"):
             "the regions specified in these files.",
             metavar="FILE1.bed FILE2.bed",
             nargs="+",
+            type=existingFile,
             required=True,
         )
 

pydeeptools/deeptools/multiBigwigSummary.py

@@ -7,6 +7,7 @@
 
 import deeptools.getScorePerBigWigBin as score_bw
 from deeptools import parserCommon
+from deeptools.parserCommon import existingFile
 from deeptools.utilities import smartLabels
 
 
@@ -128,6 +129,7 @@ def multiBigwigSummaryArgs(case="bins"):
         metavar="FILE1 FILE2",
         help="List of bigWig files, separated by spaces.",
         nargs="+",
+        type=existingFile,
         required=True,
     )
 
@@ -216,6 +218,7 @@ def multiBigwigSummaryArgs(case="bins"):
             help="Limits the analysis to the regions specified in this file.",
             metavar="file1.bed file2.bed",
             nargs="+",
+            type=existingFile,
             required=True,
         )
 

pydeeptools/deeptools/parserCommon.py

@@ -333,6 +333,7 @@ def getParentArgParse(args=None, binSize=True, blackList=True):
                               help="A BED or GTF file (optionally gzip-compressed) containing regions that should be excluded from all analyses. Filtering is performed at base-pair resolution, so only the portion of a read/fragment that overlaps a blacklisted region is excluded. Please note that you should adjust the effective genome size, if relevant.",
                               metavar="BED file",
                               nargs="+",
+                              type=existingFile,
                               required=False)
 
     optional.add_argument('--numberOfProcessors', '-p',
@@ -414,6 +415,17 @@ def writableFile(string):
     return string
 
 
+def existingFile(string):
+    """
+    Simple function that checks if a path exists (skip URL, skip 'None' as placeholders for rust code)
+    """
+    if string == "None" or string.startswith(("http://", "https://", "ftp://")):
+        return string
+    if not Path(string).is_file():
+        raise argparse.ArgumentTypeError(f"{string} file does not exist")
+    return string
+
+
 """
 Arguments used by heatmapper and profiler
 """
@@ -424,7 +436,7 @@ def heatmapperMatrixArgs(args=None):
     required = parser.add_argument_group('Required arguments')
     required.add_argument('--matrixFile', '-m',
                           help='Matrix file from the computeMatrix tool.',
-                          type=Path,
+                          type=existingFile,
                           )
 
     required.add_argument('--outFileName', '-out', '-o',

pydeeptools/deeptools/plotCorrelation.py

@@ -7,7 +7,7 @@
 
 from deeptools import matplotlib_defaults  # noqa: F401
 from deeptools.correlation import Correlation
-from deeptools.parserCommon import writableFile
+from deeptools.parserCommon import existingFile, writableFile
 
 
 def parse_arguments(args=None):
@@ -48,6 +48,7 @@ def plot_correlation_args():
     required.add_argument('--corData', '-in',
                           metavar='FILE',
                           help='Compressed matrix of values generated by multiBigwigSummary or multiBamSummary',
+                          type=existingFile,
                           required=True)
 
     required.add_argument('--corMethod', '-c',

pydeeptools/deeptools/plotCoverage.py

@@ -11,6 +11,7 @@
     matplotlib_defaults,  # noqa: F401
     parserCommon,
 )
+from deeptools.parserCommon import existingFile
 from deeptools.utilities import smartLabels
 
 
@@ -69,6 +70,7 @@ def required_args():
                           metavar='FILE1 FILE2',
                           help='List of indexed BAM files separated by spaces.',
                           nargs='+',
+                          type=existingFile,
                           required=True)
 
     optional = parser.add_argument_group('Optional arguments')
@@ -122,7 +124,8 @@ def required_args():
                           '--outRawCounts or many tens of thousands of regions, as per-base '
                           'coverage is used!',
                           metavar='FILE1.bed FILE2.bed',
-                          nargs='+')
+                          nargs='+',
+                          type=existingFile)
 
     optional.add_argument('--outRawCounts',
                           help='Save raw counts (coverages) to file.',

pydeeptools/deeptools/plotEnrichment.py

@@ -14,6 +14,7 @@
 from deeptools.countReadsPerBin import CountReadsPerBin as cr
 from deeptools.getFragmentAndReadSize import get_read_and_fragment_length
 from deeptools.mapReduce import blSubtract, getUserRegion, mapReduce
+from deeptools.parserCommon import existingFile
 from deeptools.utilities import getCommonChrNames, getTLen, mungeChromosome, smartLabels
 
 
@@ -57,6 +58,7 @@ def plot_enrichment_args():
                           metavar='file1.bam file2.bam',
                           help='List of indexed bam files separated by spaces.',
                           nargs='+',
+                          type=existingFile,
                           required=True)
 
     required.add_argument('--BED',
@@ -67,6 +69,7 @@ def plot_enrichment_args():
                           'and "peak" for BED files.',
                           metavar='FILE1.bed FILE2.bed',
                           nargs='+',
+                          type=existingFile,
                           required=True)
 
     optional = parser.add_argument_group('Optional arguments')