biopython/Bio/AlignIO/PhylipIO.py

# Copyright 2006-2016 by Peter Cock.  All rights reserved.
# Revisions copyright 2011 Brandon Invergo. All rights reserved.
#
# This file is part of the Biopython distribution and governed by your
# choice of the "Biopython License Agreement" or the "BSD 3-Clause License".
# Please see the LICENSE file that should have been included as part of this
# package.
"""AlignIO support for "phylip" format from Joe Felsenstein's PHYLIP tools.

You are expected to use this module via the Bio.AlignIO functions (or the
Bio.SeqIO functions if you want to work directly with the gapped sequences).

Support for "relaxed phylip" format is also provided. Relaxed phylip differs
from standard phylip format in the following ways:

 - No whitespace is allowed in the sequence ID.
 - No truncation is performed. Instead, sequence IDs are padded to the longest
   ID length, rather than 10 characters. A space separates the sequence
   identifier from the sequence.

Relaxed phylip is supported by RAxML and PHYML.

Note
====

In TREE_PUZZLE (Schmidt et al. 2003) and PHYML (Guindon and Gascuel 2003)
a dot/period (".") in a sequence is interpreted as meaning the same
character as in the first sequence.  The PHYLIP documentation from 3.3 to 3.69
http://evolution.genetics.washington.edu/phylip/doc/sequence.html says:

"a period was also previously allowed but it is no longer allowed,
because it sometimes is used in different senses in other programs"

Biopython 1.58 or later treats dots/periods in the sequence as invalid, both
for reading and writing. Older versions did nothing special with a dot/period.
"""

import string

from Bio.Align import MultipleSeqAlignment
from Bio.Seq import Seq
from Bio.SeqRecord import SeqRecord

from .Interfaces import AlignmentIterator
from .Interfaces import SequentialAlignmentWriter

_PHYLIP_ID_WIDTH = 10
_NO_DOTS = "PHYLIP format no longer allows dots in sequence"


class PhylipWriter(SequentialAlignmentWriter):
    """Phylip alignment writer."""

    def write_alignment(self, alignment, id_width=_PHYLIP_ID_WIDTH):
        """Use this to write (another) single alignment to an open file.

        This code will write interlaced alignments (when the sequences are
        longer than 50 characters).

        Note that record identifiers are strictly truncated to id_width,
        defaulting to the value required to comply with the PHYLIP standard.

        For more information on the file format, please see:
        http://evolution.genetics.washington.edu/phylip/doc/sequence.html
        http://evolution.genetics.washington.edu/phylip/doc/main.html#inputfiles
        """
        handle = self.handle

        if len(alignment) == 0:
            raise ValueError("Must have at least one sequence")
        length_of_seqs = alignment.get_alignment_length()
        for record in alignment:
            if length_of_seqs != len(record.seq):
                raise ValueError("Sequences must all be the same length")
        if length_of_seqs <= 0:
            raise ValueError("Non-empty sequences are required")

        # Check for repeated identifiers...
        # Apply this test *after* cleaning the identifiers
        names = []
        seqs = []
        for record in alignment:
            """
            Quoting the PHYLIP version 3.6 documentation:

            The name should be ten characters in length, filled out to
            the full ten characters by blanks if shorter. Any printable
            ASCII/ISO character is allowed in the name, except for
            parentheses ("(" and ")"), square brackets ("[" and "]"),
            colon (":"), semicolon (";") and comma (","). If you forget
            to extend the names to ten characters in length by blanks,
            the program [i.e. PHYLIP] will get out of synchronization
            with the contents of the data file, and an error message will
            result.

            Note that Tab characters count as only one character in the
            species names. Their inclusion can cause trouble.
            """
            name = sanitize_name(record.id, id_width)
            if name in names:
                raise ValueError(
                    "Repeated name %r (originally %r), possibly due to truncation"
                    % (name, record.id)
                )
            names.append(name)
            sequence = str(record.seq)
            if "." in sequence:
                # Do this check here (once per record, not once per block)
                raise ValueError(_NO_DOTS)
            seqs.append(sequence)

        # From experimentation, the use of tabs is not understood by the
        # EMBOSS suite.  The nature of the expected white space is not
        # defined in the PHYLIP documentation, simply "These are in free
        # format, separated by blanks".  We'll use spaces to keep EMBOSS
        # happy.
        handle.write(" %i %s\n" % (len(alignment), length_of_seqs))
        block = 0
        while True:
            for name, sequence in zip(names, seqs):
                if block == 0:
                    # Write name (truncated/padded to id_width characters)
                    # Now truncate and right pad to expected length.
                    handle.write(name[:id_width].ljust(id_width))
                else:
                    # write indent
                    handle.write(" " * id_width)
                # Write five chunks of ten letters per line...
                for chunk in range(5):
                    i = block * 50 + chunk * 10
                    seq_segment = sequence[i : i + 10]
                    # TODO - Force any gaps to be '-' character?
                    # TODO - How to cope with '?' or '.' in the sequence?
                    handle.write(f" {seq_segment}")
                    if i + 10 > length_of_seqs:
                        break
                handle.write("\n")
            block += 1
            if block * 50 >= length_of_seqs:
                break
            handle.write("\n")


class PhylipIterator(AlignmentIterator):
    """Reads a Phylip alignment file returning a MultipleSeqAlignment iterator.

    Record identifiers are limited to at most 10 characters.

    It only copes with interlaced phylip files!  Sequential files won't work
    where the sequences are split over multiple lines.

    For more information on the file format, please see:
    http://evolution.genetics.washington.edu/phylip/doc/sequence.html
    http://evolution.genetics.washington.edu/phylip/doc/main.html#inputfiles
    """

    # Default truncation length
    id_width = _PHYLIP_ID_WIDTH

    _header = None  # for caching lines between __next__ calls

    def _is_header(self, line):
        line = line.strip()
        parts = [x for x in line.split() if x]
        if len(parts) != 2:
            return False  # First line should have two integers
        try:
            number_of_seqs = int(parts[0])
            length_of_seqs = int(parts[1])
            return True
        except ValueError:
            return False  # First line should have two integers

    def _split_id(self, line):
        """Extract the sequence ID from a Phylip line (PRIVATE).

        Returning a tuple containing: (sequence_id, sequence_residues)

        The first 10 characters in the line are are the sequence id, the
        remainder are sequence data.
        """
        seq_id = line[: self.id_width].strip()
        seq = line[self.id_width :].strip().replace(" ", "")
        return seq_id, seq

    def __next__(self):
        """Parse the next alignment from the handle."""
        handle = self.handle

        if self._header is None:
            line = handle.readline()
        else:
            # Header we saved from when we were parsing
            # the previous alignment.
            line = self._header
            self._header = None

        if not line:
            raise StopIteration
        line = line.strip()
        parts = [x for x in line.split() if x]
        if len(parts) != 2:
            raise ValueError("First line should have two integers")
        try:
            number_of_seqs = int(parts[0])
            length_of_seqs = int(parts[1])
        except ValueError:
            raise ValueError("First line should have two integers") from None

        assert self._is_header(line)

        if (
            self.records_per_alignment is not None
            and self.records_per_alignment != number_of_seqs
        ):
            raise ValueError(
                "Found %i records in this alignment, told to expect %i"
                % (number_of_seqs, self.records_per_alignment)
            )

        ids = []
        seqs = []

        # By default, expects STRICT truncation / padding to 10 characters.
        # Does not require any whitespace between name and seq.
        for i in range(number_of_seqs):
            line = handle.readline().rstrip()
            sequence_id, s = self._split_id(line)
            ids.append(sequence_id)
            if "." in s:
                raise ValueError(_NO_DOTS)
            seqs.append([s])

        # Look for further blocks
        line = ""
        while True:
            # Skip any blank lines between blocks...
            while "" == line.strip():
                line = handle.readline()
                if not line:
                    break  # end of file
            if not line:
                break  # end of file

            if self._is_header(line):
                # Looks like the start of a concatenated alignment
                self._header = line
                break

            # print("New block...")
            for i in range(number_of_seqs):
                s = line.strip().replace(" ", "")
                if "." in s:
                    raise ValueError(_NO_DOTS)
                seqs[i].append(s)
                line = handle.readline()
                if (not line) and i + 1 < number_of_seqs:
                    raise ValueError("End of file mid-block")
            if not line:
                break  # end of file

        records = (
            SeqRecord(Seq("".join(s)), id=i, name=i, description=i)
            for (i, s) in zip(ids, seqs)
        )
        return MultipleSeqAlignment(records)


# Relaxed Phylip
class RelaxedPhylipWriter(PhylipWriter):
    """Relaxed Phylip format writer."""

    def write_alignment(self, alignment):
        """Write a relaxed phylip alignment."""
        # Check inputs
        for name in (s.id.strip() for s in alignment):
            if any(c in name for c in string.whitespace):
                raise ValueError(f"Whitespace not allowed in identifier: {name}")

        # Calculate a truncation length - maximum length of sequence ID plus a
        # single character for padding
        # If no sequences, set id_width to 1. super(...) call will raise a
        # ValueError
        if len(alignment) == 0:
            id_width = 1
        else:
            id_width = max(len(s.id.strip()) for s in alignment) + 1
        super().write_alignment(alignment, id_width)


class RelaxedPhylipIterator(PhylipIterator):
    """Relaxed Phylip format Iterator."""

    def _split_id(self, line):
        """Extract the sequence ID from a Phylip line (PRIVATE).

        Returns a tuple containing: (sequence_id, sequence_residues)

        For relaxed format split at the first whitespace character.
        """
        seq_id, sequence = line.split(None, 1)
        sequence = sequence.strip().replace(" ", "")
        return seq_id, sequence


class SequentialPhylipWriter(SequentialAlignmentWriter):
    """Sequential Phylip format Writer."""

    def write_alignment(self, alignment, id_width=_PHYLIP_ID_WIDTH):
        """Write a Phylip alignment to the handle."""
        handle = self.handle

        if len(alignment) == 0:
            raise ValueError("Must have at least one sequence")
        length_of_seqs = alignment.get_alignment_length()
        for record in alignment:
            if length_of_seqs != len(record.seq):
                raise ValueError("Sequences must all be the same length")
        if length_of_seqs <= 0:
            raise ValueError("Non-empty sequences are required")

        # Check for repeated identifiers...
        # Apply this test *after* cleaning the identifiers
        names = []
        for record in alignment:
            # Either remove the banned characters, or map them to something
            # else like an underscore "_" or pipe "|" character...
            name = sanitize_name(record.id, id_width)
            if name in names:
                raise ValueError(
                    "Repeated name %r (originally %r), possibly due to truncation"
                    % (name, record.id)
                )
            names.append(name)

        # From experimentation, the use of tabs is not understood by the
        # EMBOSS suite.  The nature of the expected white space is not
        # defined in the PHYLIP documentation, simply "These are in free
        # format, separated by blanks".  We'll use spaces to keep EMBOSS
        # happy.
        handle.write(" %i %s\n" % (len(alignment), length_of_seqs))
        for name, record in zip(names, alignment):
            sequence = str(record.seq)
            if "." in sequence:
                raise ValueError(_NO_DOTS)
            handle.write(name[:id_width].ljust(id_width))
            # Write the entire sequence to one line (see sequential format
            # notes in the SequentialPhylipIterator docstring
            handle.write(sequence)
            handle.write("\n")


class SequentialPhylipIterator(PhylipIterator):
    """Sequential Phylip format Iterator.

    The sequential format carries the same restrictions as the normal
    interleaved one, with the difference being that the sequences are listed
    sequentially, each sequence written in its entirety before the start of
    the next. According to the PHYLIP documentation for input file
    formatting, newlines and spaces may optionally be entered at any point
    in the sequences.
    """

    _header = None  # for caching lines between __next__ calls

    def __next__(self):
        """Parse the next alignment from the handle."""
        handle = self.handle

        if self._header is None:
            line = handle.readline()
        else:
            # Header we saved from when we were parsing
            # the previous alignment.
            line = self._header
            self._header = None

        if not line:
            raise StopIteration
        line = line.strip()
        parts = [x for x in line.split() if x]
        if len(parts) != 2:
            raise ValueError("First line should have two integers")
        try:
            number_of_seqs = int(parts[0])
            length_of_seqs = int(parts[1])
        except ValueError:
            raise ValueError("First line should have two integers") from None

        assert self._is_header(line)

        if (
            self.records_per_alignment is not None
            and self.records_per_alignment != number_of_seqs
        ):
            raise ValueError(
                "Found %i records in this alignment, told to expect %i"
                % (number_of_seqs, self.records_per_alignment)
            )

        ids = []
        seqs = []

        # By default, expects STRICT truncation / padding to 10 characters.
        # Does not require any whitespace between name and seq.
        for i in range(number_of_seqs):
            line = handle.readline().rstrip()
            sequence_id, s = self._split_id(line)
            ids.append(sequence_id)
            while len(s) < length_of_seqs:
                # The sequence may be split into multiple lines
                line = handle.readline().strip()
                if not line:
                    break
                if line == "":
                    continue
                s = "".join([s, line.strip().replace(" ", "")])
                if len(s) > length_of_seqs:
                    raise ValueError(
                        "Found a record of length %i, "
                        "should be %i" % (len(s), length_of_seqs)
                    )
            if "." in s:
                raise ValueError(_NO_DOTS)
            seqs.append(s)
        while True:
            # Find other alignments in the file
            line = handle.readline()
            if not line:
                break
            if self._is_header(line):
                self._header = line
                break

        records = (
            SeqRecord(Seq(s), id=i, name=i, description=i) for (i, s) in zip(ids, seqs)
        )
        return MultipleSeqAlignment(records)


def sanitize_name(name, width=None):
    """Sanitise sequence identifier for output.

    Removes the banned characters "[]()" and replaces the characters ":;"
    with "|". The name is truncated to "width" characters if specified.
    """
    name = name.strip()
    for char in "[](),":
        name = name.replace(char, "")
    for char in ":;":
        name = name.replace(char, "|")
    if width is not None:
        name = name[:width]
    return name