Merging in version 1.2 from the devel branch

.gitignore

@@ -3,6 +3,7 @@
 
 # Setuptools distribution folder.
 /dist/
+/build/
 
 # Python egg metadata, regenerated from source files by setuptools.
 /*.egg-info

CHANGES.md

@@ -14,3 +14,9 @@
 * View only the first and last N characters of accessions with option -aa.
 * Read from stdin with the magic file name "-".
 
+## v1.2.0
+
+* The option `-l` colors the alignment but does not break it into blocks. Suitable for piping to `less -RS`,
+  as suggested by Mark McMullan <[email protected]>.
+* More indices indicated below alignments, and with an up-arrow as a tick mark.
+* Added option `-sm` to allow restricting output to sequences with accessions containing a given string. 

README.md

@@ -1,5 +1,6 @@
 [![PyPI version](https://badge.fury.io/py/alv.svg)](https://badge.fury.io/py/alv) 
 [![Build Status](https://travis-ci.org/arvestad/alv.svg?branch=master)](https://travis-ci.org/arvestad/alv)
+
 # alv: a command-line alignment viewer
 
 View you DNA or protein multiple-sequence alignments right at your command line. No need to launch a
@@ -56,14 +57,16 @@ Run `python setup.py develop test` for development install and to execute tests.
 
 All of the sequences in PFAM's seed alignment for PF00005
 
-![PF00005 seed MSA](https://github.com/arvestad/alv/blob/master/doc/screenshot_PF00005.png)
+![PF00005 seed MSA](https://github.com/arvestad/alv/raw/master/doc/screenshot_PF00005.png)
+
+### Yeast sequences from PF00005
 
-### Ten peptide sequences from PF00005
+Using the option `-sm YEAST`, we reduce the alignment to the ones with a matching accession.
 
-![MSA from PF00005](https://github.com/arvestad/alv/blob/master/doc/screenshot_1.png)
+![MSA from PF00005](https://github.com/arvestad/alv/raw/master/doc/PF00005_yeast.png)
 
 ### Seven coding DNA sequences
 
 `alv` is autodetecting that the given DNA sequences are coding and therefore colors codons instead
 of nucleotides.
-![Sample screenshot](https://github.com/arvestad/alv/blob/master/doc/screenshot_2.png)
+![Sample screenshot](https://github.com/arvestad/alv/raw/master/doc/screenshot_2.png)

TODO.md

@@ -1,13 +1,12 @@
 # Planned fixes and features
 
+* Better screenshots
 * Review how colors work in different terminals. The color scheme working for me looked strange when Vilde was
   showing me her work.
-* Make it possible to color alignments without breaking alignment into blocks. I.e., a "pure"
-  version of `alv -k -w 99999 msa.fa | less -SR`. Suggested by Mark McMullan <[email protected]>.
 * Add support for restricting to a sub-alignment.
 * Add option --glimpse.
 * Explicitly choose parts of an alignment to view/color.
-
+* Pypi.org does not handle MarkDown in the README. Should look for a solution. 
 
 # Considered features
 

alv/alignment.py

@@ -22,6 +22,12 @@ def __init__(self, alignment):
     def _update_seq_index(self):
         self.seq_indices = { r.id : i for i, r in enumerate(self.al)} # Get a dictionary mapping accession to row index in alignment
 
+    def al_width(self):
+        '''
+        The number of columns in the alignment.
+        '''
+        return self.al.get_alignment_length()
+
     def trim_accessions(self, start, stop):
         for record in self.al:
             acc = record.id
@@ -61,32 +67,33 @@ def sort_by_identity(self, acc):
                                        reverse=True))
         return sorted_accessions
 
-    def accession_widths(self):
+    def accession_widths(self, accessions=None):
         '''
         Compute the space needed for all accessions, plus one for
         a delimiter.
         '''
         max_accession_length = 5        # initial guess, or minimum
         for record in self.al:
-            if len(record.id) > max_accession_length:
-                  max_accession_length = len(record.id)
+            if not accessions or record.id in accessions:
+                if len(record.id) > max_accession_length:
+                    max_accession_length = len(record.id)
         return max_accession_length
 
 
-    def block_width(self, terminal_width, args):
+    def block_width(self, terminal_width, args_width):
         '''
         For wide alignments, we need to break it up in blocks.
         This method calculates how many characters to output in a block.
         
         Take the margin size (for accessions) into account and avoid ending up
         with blocks of size 1.
         '''
-        if args.width == 0:
+        if args_width == 0:
             al_width = self.al.get_alignment_length()
             left_margin = 1 + self.accession_widths() # Add 1 for a space to the right of the accessions
             return self._compute_block_width(terminal_width, al_width, left_margin)
         else:
-            return args.width
+            return args_width
 
     def _compute_block_width(self, terminal_width, al_width, left_margin):
         '''
@@ -110,7 +117,8 @@ def blocks(self, block_width):
             raise AlvEmptyAlignment()
         else:
             for start in range(0, al_width, block_width):
-                yield AlignmentBlock(start, start + block_width)
+                end = min(al_width, start + block_width)
+                yield AlignmentBlock(start, end)
 
     def apply_painter(self, acc, block, painter):
         '''

alv/alignmentterminal.py

@@ -15,14 +15,15 @@ def __init__(self, args):
             self.order = args.sorting_order.split(',')
             if len(self.order) == 0:
                 raise Exception('Bad order specification: no accessions in input')
+        if args.select_matching:
+            self.selection = args.select_matching
+        else:
+            self.selection = False
 
-    def output_alignment(self, al, painter, args):
+    def output_alignment(self, al, painter, width):
         '''
         Output alignment al to stdout in blocks of width at most w with colors from painter.
         '''
-        self.left_margin = 1 + al.accession_widths()
-        assert self.left_margin < self.width - 10
-
         if self.sorting == 'alpha':
             accessions = al.sorted_accessions()
         elif self.sorting == 'fixed':
@@ -31,19 +32,79 @@ def output_alignment(self, al, painter, args):
             accessions = al.sort_by_identity(self.order)
         else:
             accessions = al.accessions()
-        accessions = list(accessions)
+
+        if self.selection:
+            chosen_accessions = []
+            for acc in accessions:
+                if self.selection in acc:
+                    chosen_accessions.append(acc)
+        else:
+            chosen_accessions = list(accessions)
+
+        self.left_margin = 1 + al.accession_widths(chosen_accessions)
+        assert self.left_margin < self.width - 10
 
-        columns_per_block = al.block_width(self.width, args)
+        columns_per_block = al.block_width(self.width, width)
         for block in al.blocks(columns_per_block):
-            for acc in accessions:
+            for acc in chosen_accessions:
                 colored_subseq = al.apply_painter(acc, block, painter)
                 print("{0:{width}}{1}".format(acc, colored_subseq, width=self.left_margin))
-            print(' ' * self.left_margin, block.start, sep='')
+            print(make_tick_string(self.left_margin, block.start, block.end, 20, 7))
 
+#            print(' ' * self.left_margin, '↑',  block.start, sep='') # print index of first column
 
-    def print_one_sequence_block(self, record, left_margin, start, block_width):
-        colored_string = colorize_sequence_string(rec.seq[start : start + block_width])
-        print("{0:{width}}{1}".format(rec.id, colored_string, width=left_margin))
+
+    # def print_one_sequence_block(self, record, left_margin, start, block_width):
+    #     colored_string = colorize_sequence_string(rec.seq[start : start + block_width])
+    #     print("{0:{width}}{1}".format(rec.id, colored_string, width=left_margin))
 
 
 
+def calc_tick_indices(start, end, distance, min_distance):
+    '''
+    Return a list of indices for which we want a tick mark at the bottom of the alignment.
+    The goal is to have an index for the starting position of a block (leftmost column number),
+    and then a tick mark on even multiples of 20 (or what is given by 'distance'), for example:
+        53   60                  80                 100
+    Care is needed so that space is left between first and second indices, and min_distance indicates
+    how much.
+    '''
+    first_even_pos = (start // distance + 1) * distance
+    if first_even_pos - start < min_distance:
+        first_even_pos += distance # Compensate a bit
+    positions = range(first_even_pos, end, distance)
+    return positions
+
+def make_one_tick(position, space):
+    '''
+    Return a string which is 'space' wide and contains a number (the position)
+    followed by an up-arrow.
+    '''
+
+    return '{0:>{width}}↑'.format(position, width=space-1) 
+
+def make_tick_string(left_margin, start, end, distance, min_distance):
+    '''
+    Construct the index bar which is printed at the bottom of an alignment block.
+
+    left_margin is how much space is allowed for accessions.
+    start is the column number of the beginning of an alignment block.
+    end is the last column of an alignment block.
+    distance is the desired distance between up-arrows
+    min_distance is the space we allow for position numbers plus an up-arrow
+    '''
+    even_indices = calc_tick_indices(start, end, distance, min_distance)
+
+    # Initial space
+    index_bar = ' ' * (left_margin - min_distance + 1) # Account for space needed by indices
+
+    # Add first column index
+    index_bar += make_one_tick(start, min(left_margin+1, min_distance))
+
+    last_pos = start
+    for pos in even_indices:
+        spacer = pos - last_pos - 1
+        index_block = '{0:>{width}}↑'.format(pos, width=spacer)
+        index_bar += index_block
+        last_pos = pos
+    return index_bar

alv/version.py

@@ -1 +1 @@
-__version__ = '1.1.0'
+__version__ = '1.2.0'