Improve docs re chimeric/unmapped/unpaired read pairs

tests/count_help

@@ -71,15 +71,6 @@ Options:
     --mapping-quality=MAPPING_QUALITY
                         Minimum mapping quality for a read to be retained
                         [default=0]
-    --unmapped-reads=UNMAPPED_READS
-                        How to handle unmapped reads. Options are 'discard',
-                        'use' or 'correct' [default=discard]
-    --chimeric-pairs=CHIMERIC_PAIRS
-                        How to handle chimeric read pairs. Options are
-                        'discard', 'use' or 'correct' [default=use]
-    --unpaired-reads=UNPAIRED_READS
-                        How to handle unpaired reads. Options are 'discard',
-                        'use' or 'correct' [default=use]
     --ignore-umi        Ignore UMI and dedup only on position
     --ignore-tlen       Option to dedup paired end reads based solely on
                         read1, whether or not the template length is the same
@@ -90,6 +81,17 @@ Options:
     -o, --out-sam       Output alignments in sam format [default=False]
     --no-sort-output    Don't Sort the output
 
+  Dedup and Count SAM/BAM options:
+    --unmapped-reads=UNMAPPED_READS
+                        How to handle unmapped reads. Options are 'discard' or
+                        'use' [default=discard]
+    --chimeric-pairs=CHIMERIC_PAIRS
+                        How to handle chimeric read pairs. Options are
+                        'discard' or 'use'  [default=use]
+    --unpaired-reads=UNPAIRED_READS
+                        How to handle unpaired reads. Options are 'discard'or
+                        'use' [default=use]
+
   input/output options:
     -I FILE, --stdin=FILE
                         file to read stdin from [default = stdin].

tests/dedup_help

@@ -81,15 +81,6 @@ Options:
     --mapping-quality=MAPPING_QUALITY
                         Minimum mapping quality for a read to be retained
                         [default=0]
-    --unmapped-reads=UNMAPPED_READS
-                        How to handle unmapped reads. Options are 'discard',
-                        'use' or 'correct' [default=discard]
-    --chimeric-pairs=CHIMERIC_PAIRS
-                        How to handle chimeric read pairs. Options are
-                        'discard', 'use' or 'correct' [default=use]
-    --unpaired-reads=UNPAIRED_READS
-                        How to handle unpaired reads. Options are 'discard',
-                        'use' or 'correct' [default=use]
     --ignore-umi        Ignore UMI and dedup only on position
     --ignore-tlen       Option to dedup paired end reads based solely on
                         read1, whether or not the template length is the same
@@ -100,6 +91,17 @@ Options:
     -o, --out-sam       Output alignments in sam format [default=False]
     --no-sort-output    Don't Sort the output
 
+  Dedup and Count SAM/BAM options:
+    --unmapped-reads=UNMAPPED_READS
+                        How to handle unmapped reads. Options are 'discard' or
+                        'use' [default=discard]
+    --chimeric-pairs=CHIMERIC_PAIRS
+                        How to handle chimeric read pairs. Options are
+                        'discard' or 'use'  [default=use]
+    --unpaired-reads=UNPAIRED_READS
+                        How to handle unpaired reads. Options are 'discard'or
+                        'use' [default=use]
+
   input/output options:
     -I FILE, --stdin=FILE
                         file to read stdin from [default = stdin].

tests/group_help

@@ -84,15 +84,6 @@ Options:
     --mapping-quality=MAPPING_QUALITY
                         Minimum mapping quality for a read to be retained
                         [default=0]
-    --unmapped-reads=UNMAPPED_READS
-                        How to handle unmapped reads. Options are 'discard',
-                        'use' or 'correct' [default=discard]
-    --chimeric-pairs=CHIMERIC_PAIRS
-                        How to handle chimeric read pairs. Options are
-                        'discard', 'use' or 'correct' [default=use]
-    --unpaired-reads=UNPAIRED_READS
-                        How to handle unpaired reads. Options are 'discard',
-                        'use' or 'correct' [default=use]
     --ignore-umi        Ignore UMI and dedup only on position
     --ignore-tlen       Option to dedup paired end reads based solely on
                         read1, whether or not the template length is the same
@@ -103,6 +94,17 @@ Options:
     -o, --out-sam       Output alignments in sam format [default=False]
     --no-sort-output    Don't Sort the output
 
+  Group SAM/BAM options:
+    --unmapped-reads=UNMAPPED_READS
+                        How to handle unmapped reads. Options are 'discard',
+                        'use' or 'output' [default=discard]
+    --chimeric-pairs=CHIMERIC_PAIRS
+                        How to handle chimeric read pairs. Options are
+                        'discard', 'use' or 'output' [default=use]
+    --unpaired-reads=UNPAIRED_READS
+                        How to handle unpaired reads. Options are 'discard',
+                        'use' or 'output' [default=use]
+
   input/output options:
     -I FILE, --stdin=FILE
                         file to read stdin from [default = stdin].

umi_tools/Utilities.py

@@ -258,8 +258,6 @@ class method (:func:`cachedmethod`) calls.
 import regex
 from umi_tools import __version__
 
-from builtins import bytes, chr
-
 
 class DefaultOptions:
     stdlog = sys.stdout
@@ -599,6 +597,8 @@ def Start(parser=None,
           add_extract_options=False,
           add_group_dedup_options=True,
           add_sam_options=True,
+          add_dedup_count_sam_options=False,
+          add_group_sam_options=False,
           add_umi_grouping_options=True,
           return_parser=False):
     """set up an experiment.
@@ -885,27 +885,6 @@ def Start(parser=None,
         group.add_option("--output-unmapped", dest="output_unmapped", action="store_true",
                          default=False, help=optparse.SUPPRESS_HELP)
 
-        group.add_option("--unmapped-reads", dest="unmapped_reads",
-                         type="choice",
-                         choices=("discard", "use", "output"),
-                         default="discard",
-                         help=("How to handle unmapped reads. Options are "
-                               "'discard', 'use' or 'correct' [default=%default]"))
-
-        group.add_option("--chimeric-pairs", dest="chimeric_pairs",
-                         type="choice",
-                         choices=("discard", "use", "output"),
-                         default="use",
-                         help=("How to handle chimeric read pairs. Options are "
-                               "'discard', 'use' or 'correct' [default=%default]"))
-
-        group.add_option("--unpaired-reads", dest="unpaired_reads",
-                         type="choice",
-                         choices=("discard", "use", "output"),
-                         default="use",
-                         help=("How to handle unpaired reads. Options are "
-                               "'discard', 'use' or 'correct' [default=%default]"))
-
         group.add_option("--ignore-umi", dest="ignore_umi",
                          action="store_true", help="Ignore UMI and dedup"
                          " only on position", default=False)
@@ -941,6 +920,56 @@ def Start(parser=None,
 
         parser.add_option_group(group)
 
+    if add_dedup_count_sam_options:
+        group = OptionGroup(parser, "Dedup and Count SAM/BAM options")
+
+        group.add_option("--unmapped-reads", dest="unmapped_reads",
+                         type="choice",
+                         choices=("discard", "use"),
+                         default="discard",
+                         help=("How to handle unmapped reads. Options are "
+                               "'discard' or 'use' [default=%default]"))
+
+        group.add_option("--chimeric-pairs", dest="chimeric_pairs",
+                         type="choice",
+                         choices=("discard", "use"),
+                         default="use",
+                         help=("How to handle chimeric read pairs. Options are "
+                               "'discard' or 'use'  [default=%default]"))
+
+        group.add_option("--unpaired-reads", dest="unpaired_reads",
+                         type="choice",
+                         choices=("discard", "use"),
+                         default="use",
+                         help=("How to handle unpaired reads. Options are "
+                               "'discard'or 'use' [default=%default]"))
+        parser.add_option_group(group)
+
+    if add_group_sam_options:
+        group = OptionGroup(parser, "Group SAM/BAM options")
+
+        group.add_option("--unmapped-reads", dest="unmapped_reads",
+                         type="choice",
+                         choices=("discard", "use", "output"),
+                         default="discard",
+                         help=("How to handle unmapped reads. Options are "
+                               "'discard', 'use' or 'output' [default=%default]"))
+
+        group.add_option("--chimeric-pairs", dest="chimeric_pairs",
+                         type="choice",
+                         choices=("discard", "use", "output"),
+                         default="use",
+                         help=("How to handle chimeric read pairs. Options are "
+                               "'discard', 'use' or 'output' [default=%default]"))
+
+        group.add_option("--unpaired-reads", dest="unpaired_reads",
+                         type="choice",
+                         choices=("discard", "use", "output"),
+                         default="use",
+                         help=("How to handle unpaired reads. Options are "
+                               "'discard', 'use' or 'output' [default=%default]"))
+        parser.add_option_group(group)
+
     if add_pipe_options:
         group = OptionGroup(parser, "input/output options")
         group.add_option("-I", "--stdin", dest="stdin", type="string",
@@ -1178,7 +1207,7 @@ def validateExtractOptions(options):
                              "(starting with 'umi_') %s, %s" % (
                                  options.pattern, options.pattern2))
 
-    return(extract_cell, extract_umi)
+    return (extract_cell, extract_umi)
 
 
 def validateSamOptions(options, group=False):

umi_tools/count.py

@@ -91,7 +91,7 @@ def main(argv=None):
     parser.add_option_group(group)
 
     # add common options (-h/--help, ...) and parse command line
-    (options, args) = U.Start(parser, argv=argv, add_group_dedup_options=False)
+    (options, args) = U.Start(parser, argv=argv, add_group_dedup_options=False, add_dedup_count_sam_options=True)
 
     options.per_gene = True  # hardcodes counting to per-gene only
 

umi_tools/dedup.py

@@ -201,7 +201,7 @@ def main(argv=None):
     parser.add_option_group(group)
 
     # add common options (-h/--help, ...) and parse command line
-    (options, args) = U.Start(parser, argv=argv)
+    (options, args) = U.Start(parser, argv=argv, add_dedup_count_sam_options=True)
 
     U.validateSamOptions(options, group=False)
 

umi_tools/group.py

@@ -150,7 +150,7 @@ def main(argv=None):
     parser.add_option_group(group)
 
     # add common options (-h/--help, ...) and parse command line
-    (options, args) = U.Start(parser, argv=argv)
+    (options, args) = U.Start(parser, argv=argv, add_group_sam_options=True)
 
     U.validateSamOptions(options, group=True)