Logo Search packages:      
Sourcecode: python-biopython version File versions  Download package

__init__.py

# Copyright 2000 by Jeffrey Chang, Brad Chapman.  All rights reserved.
# Copyright 2006 by PeterC.  All rights reserved.
# This code is part of the Biopython distribution and governed by its
# license.  Please see the LICENSE file that should have been included
# as part of this package.

"""Code to work with GenBank formatted files.

Classes:
Iterator              Iterate through a file of GenBank entries
Dictionary            Access a GenBank file using a dictionary interface.
ErrorFeatreParser     Catch errors caused during parsing.
FeatureParser         Parse GenBank data in Seq and SeqFeature objects.
RecordParser          Parse GenBank data into a Record object.
NCBIDictionary        Access GenBank using a dictionary interface.

_BaseGenBankConsumer  A base class for GenBank consumer that implements
                      some helpful functions that are in common between
                      consumers.
_FeatureConsumer      Create SeqFeature objects from info generated by
                      the Scanner
_RecordConsumer       Create a GenBank record object from Scanner info.
_PrintingConsumer     A debugging consumer.

_Scanner              Set up a GenBank parser to parse a record.

ParserFailureError    Exception indicating a failure in the parser (ie.
                      scanner or consumer)
LocationParserError   Exception indiciating a problem with the spark based
                      location parser.

Functions:
index_file            Get a GenBank file ready to be used as a Dictionary.
search_for            Do a query against GenBank.
download_many         Download many GenBank records.

"""
import cStringIO

# other Biopython stuff
from Bio.expressions import genbank
from Bio.ParserSupport import AbstractConsumer
import utils

from Bio import Mindy
from Bio.Mindy import SimpleSeqRecord
from Bio import db
from Bio import EUtils
from Bio.EUtils import DBIds, DBIdsClient

#Constants used to parse GenBank header lines
GENBANK_INDENT = 12
GENBANK_SPACER = " " * GENBANK_INDENT

#Constants for parsing GenBank feature lines
FEATURE_KEY_INDENT = 5
FEATURE_QUALIFIER_INDENT = 21
FEATURE_KEY_SPACER = " " * FEATURE_KEY_INDENT
FEATURE_QUALIFIER_SPACER = " " * FEATURE_QUALIFIER_INDENT
        
00061 class Dictionary:
    """Access a GenBank file using a dictionary-like interface.
    """
00064     def __init__(self, indexname, parser = None):
        """Initialize and open up a GenBank dictionary.

        Each entry is a full GenBank record (i.e. from the LOCUS line
        to the // at  the end of the sequence).

        Most GenBank files have only one such "entry", in which case
        using this dictionary class is rather unnecessary.

        Arguments:
        o indexname - The name of the index for the file. This should have been
        created using the index_file function.
        o parser - An optional argument specifying a parser object that
        the records should be run through before returning the output. If
        parser is None then the unprocessed contents of the file will be
        returned.
        """
        self._index = Mindy.open(indexname)
        self._parser = parser
    
    def __len__(self):
        return len(self.keys())

    def __getitem__(self, key):
        # first try to retrieve by the base id
        try:
            seqs = self._index.lookup(id = key)
        # if we can't do that, we have to try and fetch by alias
        except KeyError:
            seqs = self._index.lookup(aliases = key)

        if len(seqs) == 1:
            seq = seqs[0]
        else:
            raise KeyError("Multiple sequences found for %s" % key)

        if self._parser:
            handle = cStringIO.StringIO(seq.text)
            return self._parser.parse(handle)
        else:
            return seq.text

    def keys(self):
        primary_key_retriever = self._index['id']
        return primary_key_retriever.keys()
        
00110 class Iterator:
    """Iterator interface to move over a file of GenBank entries one at a time.
    """
00113     def __init__(self, handle, parser = None):
        """Initialize the iterator.

        Arguments:
        o handle - A handle with GenBank entries to iterate through.
        o parser - An optional parser to pass the entries through before
        returning them. If None, then the raw entry will be returned.
        """
        from Martel import RecordReader
        from Bio import File
        if isinstance(handle, File.UndoHandle):
            _handle = handle
        else:
            _handle = File.UndoHandle(handle)
        # skip ahead until we find first record
        while _handle.peekline() and _handle.peekline().find("LOCUS") < 0:
            _handle.readline()

        self._reader = RecordReader.StartsWith(_handle, "LOCUS")          
        self._parser = parser

00134     def next(self):
        """Return the next GenBank record from the handle.

        Will return None if we ran out of records.
        """
        from Bio import File
        data = self._reader.next()

        if self._parser is not None:
            if data:
                return self._parser.parse(File.StringHandle(data))

        return data
    
    def __iter__(self):
        return iter(self.next, None)

00151 class ParserFailureError(Exception):
    """Failure caused by some kind of problem in the parser.
    """
    pass

00156 class LocationParserError(Exception):
    """Could not Properly parse out a location from a GenBank file.
    """
    pass
                                                          
00161 class FeatureParser:
    """Parse GenBank files into Seq + Feature objects.
    """
00164     def __init__(self, debug_level = 0, use_fuzziness = 1, 
                 feature_cleaner = utils.FeatureValueCleaner()):
        """Initialize a GenBank parser and Feature consumer.

        Arguments:
        o debug_level - An optional argument that species the amount of
        debugging information the parser should spit out. By default we have
        no debugging info (the fastest way to do things), but if you want
        you can set this as high as two and see exactly where a parse fails.
        o use_fuzziness - Specify whether or not to use fuzzy representations.
        The default is 1 (use fuzziness).
        o feature_cleaner - A class which will be used to clean out the
        values of features. This class must implement the function 
        clean_value. GenBank.utils has a "standard" cleaner class, which
        is used by default.
        """
        self._scanner = _Scanner(debug_level)
        self.use_fuzziness = use_fuzziness
        self._cleaner = feature_cleaner

00184     def parse(self, handle):
        """Parse the specified handle.
        """
        self._consumer = _FeatureConsumer(self.use_fuzziness, 
                                          self._cleaner)
        self._scanner.feed(handle, self._consumer)
        return self._consumer.data

00192 class RecordParser:
    """Parse GenBank files into Record objects
    """
00195     def __init__(self, debug_level = 0):
        """Initialize the parser.

        Arguments:
        o debug_level - An optional argument that species the amount of
        debugging information the parser should spit out. By default we have
        no debugging info (the fastest way to do things), but if you want
        you can set this as high as two and see exactly where a parse fails.
        """
        self._scanner = _Scanner(debug_level)

00206     def parse(self, handle):
        """Parse the specified handle into a GenBank record.
        """
        self._consumer = _RecordConsumer()
        self._scanner.feed(handle, self._consumer)
        return self._consumer.data

00213 class _BaseGenBankConsumer(AbstractConsumer):
    """Abstract GenBank consumer providing useful general functions.

    This just helps to eliminate some duplication in things that most
    GenBank consumers want to do.
    """
    # Special keys in GenBank records that we should remove spaces from
    # For instance, \translation keys have values which are proteins and
    # should have spaces and newlines removed from them. This class
    # attribute gives us more control over specific formatting problems.
    remove_space_keys = ["translation"]

    def __init__(self):
        pass

00228     def _split_keywords(self, keyword_string):
        """Split a string of keywords into a nice clean list.
        """
        # process the keywords into a python list
        if keyword_string[-1] == '.':
            keywords = keyword_string[:-1]
        else:
            keywords = keyword_string
        keyword_list = keywords.split(';')
        clean_keyword_list = [x.strip() for x in keyword_list]
        return clean_keyword_list

00240     def _split_accessions(self, accession_string):
        """Split a string of accession numbers into a list.
        """
        # first replace all line feeds with spaces
        accession = accession_string.replace("\n", " ")

        accession_list = accession.split(' ')
        clean_accession_list = [x.strip() for x in accession_list]

        return clean_accession_list

00251     def _split_taxonomy(self, taxonomy_string):
        """Split a string with taxonomy info into a list.
        """
        if taxonomy_string[-1] == '.':
            tax_info = taxonomy_string[:-1]
        else:
            tax_info = taxonomy_string
        tax_list = tax_info.split(';')
        new_tax_list = []
        for tax_item in tax_list:
            new_items = tax_item.split("\n")
            new_tax_list.extend(new_items)
        while '' in new_tax_list:
            new_tax_list.remove('')
        clean_tax_list = [x.strip() for x in new_tax_list]

        return clean_tax_list

00269     def _clean_location(self, location_string):
        """Clean whitespace out of a location string.

        The location parser isn't a fan of whitespace, so we clean it out
        before feeding it into the parser.
        """
        import string
        location_line = location_string
        for ws in string.whitespace:
            location_line = location_line.replace(ws, '')

        return location_line

00282     def _remove_newlines(self, text):
        """Remove any newlines in the passed text, returning the new string.
        """
        # get rid of newlines in the qualifier value
        newlines = ["\n", "\r"]
        for ws in newlines:
            text = text.replace(ws, "")

        return text

00292     def _normalize_spaces(self, text):
        """Replace multiple spaces in the passed text with single spaces.
        """
        # get rid of excessive spaces
        text_parts = text.split(" ")
        text_parts = filter(None, text_parts)
        return ' '.join(text_parts)

00300     def _remove_spaces(self, text):
        """Remove all spaces from the passed text.
        """
        return text.replace(" ", "")

00305     def _convert_to_python_numbers(self, start, end):
        """Convert a start and end range to python notation.

        In GenBank, starts and ends are defined in "biological" coordinates,
        where 1 is the first base and [i, j] means to include both i and j.

        In python, 0 is the first base and [i, j] means to include i, but
        not j. 

        So, to convert "biological" to python coordinates, we need to 
        subtract 1 from the start, and leave the end and things should
        be converted happily.
        """
        new_start = start - 1
        new_end = end

        return new_start, new_end

00323 class _FeatureConsumer(_BaseGenBankConsumer):
    """Create a SeqRecord object with Features to return.

    Attributes:
    o use_fuzziness - specify whether or not to parse with fuzziness in
    feature locations.
    o feature_cleaner - a class that will be used to provide specialized
    cleaning-up of feature values.
    """
    def __init__(self, use_fuzziness, feature_cleaner = None):
        from Bio.SeqRecord import SeqRecord
        _BaseGenBankConsumer.__init__(self)
        self.data = SeqRecord(None, id = None)

        self._use_fuzziness = use_fuzziness
        self._feature_cleaner = feature_cleaner

        self._seq_type = ''
        self._seq_data = []
        self._current_ref = None
        self._cur_feature = None
        self._cur_qualifier_key = None
        self._cur_qualifier_value = None

00347     def locus(self, locus_name):
        """Set the locus name is set as the name of the Sequence.
        """
        self.data.name = locus_name

    def size(self, content):
        pass

00355     def residue_type(self, type):
        """Record the sequence type so we can choose an appropriate alphabet.
        """
        self._seq_type = type

    def data_file_division(self, division):
        self.data.annotations['data_file_division'] = division

    def date(self, submit_date):
        self.data.annotations['date'] = submit_date 

00366     def definition(self, definition):
        """Set the definition as the description of the sequence.
        """
        self.data.description = definition

00371     def accession(self, acc_num):
        """Set the accession number as the id of the sequence.

        If we have multiple accession numbers, the first one passed is
        used.
        """
        new_acc_nums = self._split_accessions(acc_num)

        # if we haven't set the id information yet, add the first acc num
        if self.data.id is None:
            if len(new_acc_nums) > 0:
                self.data.id = new_acc_nums[0]

    def nid(self, content):
        self.data.annotations['nid'] = content

    def pid(self, content):
        self.data.annotations['pid'] = content

00390     def version(self, version_id):
        """Set the version to overwrite the id.

        Since the verison provides the same information as the accession
        number, plus some extra info, we set this as the id if we have
        a version.
        """
        self.data.id = version_id

    def db_source(self, content):
        self.data.annotations['db_source'] = content.rstrip()

    def gi(self, content):
        self.data.annotations['gi'] = content

    def keywords(self, content):
        self.data.annotations['keywords'] = self._split_keywords(content)

    def segment(self, content):
        self.data.annotations['segment'] = content

    def source(self, content):
        if content[-1] == '.':
            source_info = content[:-1]
        else:
            source_info = content
        self.data.annotations['source'] = source_info

    def organism(self, content):
        self.data.annotations['organism'] = content

    def taxonomy(self, content):
        self.data.annotations['taxonomy'] = self._split_taxonomy(content)
        
00424     def reference_num(self, content):
        """Signal the beginning of a new reference object.
        """
        from Bio.SeqFeature import Reference
        # if we have a current reference that hasn't been added to
        # the list of references, add it.
        if self._current_ref is not None:
            self.data.annotations['references'].append(self._current_ref)
        else:
            self.data.annotations['references'] = []

        self._current_ref = Reference()

00437     def reference_bases(self, content):
        """Attempt to determine the sequence region the reference entails.

        Possible types of information we may have to deal with:
        
        (bases 1 to 86436)
        (sites)
        (bases 1 to 105654; 110423 to 111122)
        1  (residues 1 to 182)
        """
        # first remove the parentheses or other junk
        ref_base_info = content[1:-1]

        all_locations = []
        # parse if we've got 'bases' and 'to'
        if ref_base_info.find('bases') != -1 and \
            ref_base_info.find('to') != -1:
            # get rid of the beginning 'bases'
            ref_base_info = ref_base_info[5:]
            locations = self._split_reference_locations(ref_base_info)
            all_locations.extend(locations)
        elif (ref_base_info.find("residues") >= 0 and
              ref_base_info.find("to") >= 0):
            residues_start = ref_base_info.find("residues")
            # get only the information after "residues"
            ref_base_info = ref_base_info[(residues_start + len("residues ")):]
            locations = self._split_reference_locations(ref_base_info)
            all_locations.extend(locations)

        # make sure if we are not finding information then we have
        # the string 'sites' or the string 'bases'
        elif (ref_base_info == 'sites' or
              ref_base_info.strip() == 'bases'):
            pass
        # otherwise raise an error
        else:
            raise ValueError("Could not parse base info %s in record %s" %
                             (ref_base_info, self.data.id))

        self._current_ref.location = all_locations

00478     def _split_reference_locations(self, location_string):
        """Get reference locations out of a string of reference information
        
        The passed string should be of the form:

            1 to 20; 20 to 100

        This splits the information out and returns a list of location objects
        based on the reference locations.
        """
        from Bio import SeqFeature
        # split possibly multiple locations using the ';'
        all_base_info = location_string.split(';')

        new_locations = []
        for base_info in all_base_info:
            start, end = base_info.split('to')
            new_start, new_end = \
              self._convert_to_python_numbers(int(start.strip()),
                                              int(end.strip()))
            this_location = SeqFeature.FeatureLocation(new_start, new_end)
            new_locations.append(this_location)
        return new_locations

    def authors(self, content):
        self._current_ref.authors = content

    def consrtm(self, content):
        self._current_ref.consrtm = content

    def title(self, content):
        self._current_ref.title = content

    def journal(self, content):
        self._current_ref.journal = content

    def medline_id(self, content):
        self._current_ref.medline_id = content

    def pubmed_id(self, content):
        self._current_ref.pubmed_id = content

    def remark(self, content):
        self._current_ref.comment = content

    def comment(self, content):
        self.data.annotations['comment'] = "\n".join(content)

00526     def features_line(self, content):
        """Get ready for the feature table when we reach the FEATURE line.
        """
        self.start_feature_table()

00531     def start_feature_table(self):
        """Indicate we've got to the start of the feature table.
        """
        # make sure we've added on our last reference object
        if self._current_ref is not None:
            self.data.annotations['references'].append(self._current_ref)
            self._current_ref = None

00539     def _add_feature(self):
        """Utility function to add a feature to the SeqRecord.

        This does all of the appropriate checking to make sure we haven't
        left any info behind, and that we are only adding info if it
        exists.
        """
        if self._cur_feature:
            # if we have a left over qualifier, add it to the qualifiers
            # on the current feature
            self._add_qualifier()

            self._cur_qualifier_key = ''
            self._cur_qualifier_value = ''
            self.data.features.append(self._cur_feature)
            
    def feature_key(self, content):
        from Bio import SeqFeature
        # if we already have a feature, add it on
        self._add_feature()

        # start a new feature
        self._cur_feature = SeqFeature.SeqFeature()
        self._cur_feature.type = content

        # assume positive strand to start with if we have DNA or cDNA
        # (labelled as mRNA). The complement in the location will 
        # change this later if something is on the reverse strand
        if self._seq_type.find("DNA") >= 0 or self._seq_type.find("mRNA") >= 0:
            self._cur_feature.strand = 1

00570     def location(self, content):
        """Parse out location information from the location string.

        This uses Andrew's nice spark based parser to do the parsing
        for us, and translates the results of the parse into appropriate
        Location objects.
        """
        from Bio.GenBank import LocationParser
        # --- first preprocess the location for the spark parser
        
        # we need to clean up newlines and other whitespace inside
        # the location before feeding it to the parser.
        # locations should have no whitespace whatsoever based on the
        # grammer
        location_line = self._clean_location(content)

        # Older records have junk like replace(266,"c") in the
        # location line. Newer records just replace this with
        # the number 266 and have the information in a more reasonable
        # place. So we'll just grab out the number and feed this to the
        # parser. We shouldn't really be losing any info this way.
        if location_line.find('replace') != -1:
            comma_pos = location_line.find(',')
            location_line = location_line[8:comma_pos]
        
        # feed everything into the scanner and parser
        try:
            parse_info = \
                       LocationParser.parse(LocationParser.scan(location_line))
        # spark raises SystemExit errors when parsing fails
        except SystemExit:
            raise LocationParserError(location_line)

        # print "parse_info:", repr(parse_info)
        
        # add the parser information the current feature
        self._set_location_info(parse_info, self._cur_feature)

00608     def _set_function(self, function, cur_feature):
        """Set the location information based on a function.

        This handles all of the location functions like 'join', 'complement'
        and 'order'.

        Arguments:
        o function - A LocationParser.Function object specifying the
        function we are acting on.
        o cur_feature - The feature to add information to.
        """
        from Bio import SeqFeature
        from Bio.GenBank import LocationParser
        assert isinstance(function, LocationParser.Function), \
               "Expected a Function object, got %s" % function
        
        if function.name == "complement":
            # mark the current feature as being on the opposite strand
            cur_feature.strand = -1
            # recursively deal with whatever is left inside the complement
            for inner_info in function.args:
                self._set_location_info(inner_info, cur_feature)
        # deal with functions that have multipe internal segments that
        # are connected somehow.
        # join and order are current documented functions.
        # one-of is something I ran across in old files. Treating it
        # as a sub sequence feature seems appropriate to me.
        # bond is some piece of junk I found in RefSeq files. I have
        # no idea how to interpret it, so I jam it in here
        elif (function.name == "join" or function.name == "order" or
              function.name == "one-of" or function.name == "bond"):
            self._set_ordering_info(function, cur_feature)
        elif (function.name == "gap"):
            assert len(function.args) == 1, \
              "Unexpected number of arguments in gap %s" % function.args
            # make the cur information location a gap object
            position = self._get_position(function.args[0].local_location)
            cur_feature.location = SeqFeature.PositionGap(position)
        else:
            raise ValueError("Unexpected function name: %s" % function.name)

00649     def _set_ordering_info(self, function, cur_feature):
        """Parse a join or order and all of the information in it.

        This deals with functions that order a bunch of locations,
        specifically 'join' and 'order'. The inner locations are
        added as subfeatures of the top level feature
        """
        from Bio import SeqFeature
        # for each inner element, create a sub SeqFeature within the
        # current feature, then get the information for this feature
        for inner_element in function.args:
            new_sub_feature = SeqFeature.SeqFeature()
            # inherit the type from the parent
            new_sub_feature.type = cur_feature.type 
            # add the join or order info to the location_operator
            cur_feature.location_operator = function.name
            new_sub_feature.location_operator = function.name
            # inherit references and strand from the parent feature
            new_sub_feature.ref = cur_feature.ref
            new_sub_feature.ref_db = cur_feature.ref_db
            new_sub_feature.strand = cur_feature.strand

            # set the information for the inner element
            self._set_location_info(inner_element, new_sub_feature)

            # now add the feature to the sub_features
            cur_feature.sub_features.append(new_sub_feature)

        # set the location of the top -- this should be a combination of
        # the start position of the first sub_feature and the end position
        # of the last sub_feature

        # these positions are already converted to python coordinates 
        # (when the sub_features were added) so they don't need to
        # be converted again
        feature_start = cur_feature.sub_features[0].location.start
        feature_end = cur_feature.sub_features[-1].location.end
        cur_feature.location = SeqFeature.FeatureLocation(feature_start,
                                                          feature_end)

00689     def _set_location_info(self, parse_info, cur_feature):
        """Set the location information for a feature from the parse info.

        Arguments:
        o parse_info - The classes generated by the LocationParser.
        o cur_feature - The feature to add the information to.
        """
        from Bio.GenBank import LocationParser
        # base case -- we are out of information
        if parse_info is None:
            return
        # parse a location -- this is another base_case -- we assume
        # we have no information after a single location
        elif isinstance(parse_info, LocationParser.AbsoluteLocation):
            self._set_location(parse_info, cur_feature)
            return
        # parse any of the functions (join, complement, etc)
        elif isinstance(parse_info, LocationParser.Function):
            self._set_function(parse_info, cur_feature)
        # otherwise we are stuck and should raise an error
        else:
            raise ValueError("Could not parse location info: %s"
                             % parse_info)

00713     def _set_location(self, location, cur_feature):
        """Set the location information for a feature.

        Arguments:
        o location - An AbsoluteLocation object specifying the info
        about the location.
        o cur_feature - The feature to add the information to.
        """
        # check to see if we have a cross reference to another accession
        # ie. U05344.1:514..741
        if location.path is not None:
            cur_feature.ref = location.path.accession
            cur_feature.ref_db = location.path.database
        # now get the actual location information
        cur_feature.location = self._get_location(location.local_location)

00729     def _get_location(self, range_info):
        """Return a (possibly fuzzy) location from a Range object.

        Arguments:
        o range_info - A location range (ie. something like 67..100). This
        may also be a single position (ie 27).

        This returns a FeatureLocation object.
        If parser.use_fuzziness is set at one, the positions for the
        end points will possibly be fuzzy.
        """
        from Bio import SeqFeature
        from Bio.GenBank import LocationParser
        # check if we just have a single base
        if not(isinstance(range_info, LocationParser.Range)):
            pos = self._get_position(range_info)
            # move the single position back one to be consistent with how
            # python indexes numbers (starting at 0)
            pos.position = pos.position  - 1
            return SeqFeature.FeatureLocation(pos, pos)
        # otherwise we need to get both sides of the range
        else:
            # get *Position objects for the start and end
            start_pos = self._get_position(range_info.low)
            end_pos = self._get_position(range_info.high)

            start_pos.position, end_pos.position = \
              self._convert_to_python_numbers(start_pos.position,
                                              end_pos.position)

            return SeqFeature.FeatureLocation(start_pos, end_pos)

00761     def _get_position(self, position):
        """Return a (possibly fuzzy) position for a single coordinate.

        Arguments:
        o position - This is a LocationParser.* object that specifies
        a single coordinate. We will examine the object to determine
        the fuzziness of the position.

        This is used with _get_location to parse out a location of any
        end_point of arbitrary fuzziness.
        """
        from Bio import SeqFeature
        from Bio.GenBank import LocationParser
        # case 1 -- just a normal number
        if (isinstance(position, LocationParser.Integer)):
            final_pos = SeqFeature.ExactPosition(position.val) 
        # case 2 -- we've got a > sign
        elif isinstance(position, LocationParser.LowBound):
            final_pos = SeqFeature.AfterPosition(position.base.val)
        # case 3 -- we've got a < sign
        elif isinstance(position, LocationParser.HighBound):
            final_pos = SeqFeature.BeforePosition(position.base.val)
        # case 4 -- we've got 100^101
        elif isinstance(position, LocationParser.Between):
            final_pos = SeqFeature.BetweenPosition(position.low.val,
                                                 position.high.val)
        # case 5 -- we've got (100.101)
        elif isinstance(position, LocationParser.TwoBound):
            final_pos = SeqFeature.WithinPosition(position.low.val,
                                                position.high.val)
        # case 6 -- we've got a one-of(100, 110) location
        elif isinstance(position, LocationParser.Function) and \
                        position.name == "one-of":
            # first convert all of the arguments to positions
            position_choices = []
            for arg in position.args:
                # we only handle AbsoluteLocations with no path
                # right now. Not sure if other cases will pop up
                assert isinstance(arg, LocationParser.AbsoluteLocation), \
                  "Unhandled Location type %r" % arg
                assert arg.path is None, "Unhandled path in location"
                position = self._get_position(arg.local_location)
                position_choices.append(position)
            final_pos = SeqFeature.OneOfPosition(position_choices)
        # if it is none of these cases we've got a problem!
        else:
            raise ValueError("Unexpected LocationParser object %r" %
                             position)

        # if we are using fuzziness return what we've got
        if self._use_fuzziness:
            return final_pos
        # otherwise return an ExactPosition equivalent
        else:
            return SeqFeature.ExactPosition(final_pos.location)

00817     def _add_qualifier(self):
        """Add a qualifier to the current feature without loss of info.

        If there are multiple qualifier keys with the same name we
        would lose some info in the dictionary, so we append a unique
        number to the end of the name in case of conflicts.
        """
        # if we've got a key from before, add it to the dictionary of
        # qualifiers
        if self._cur_qualifier_key:
            key = self._cur_qualifier_key
            value = "".join(self._cur_qualifier_value)
            if self._feature_cleaner is not None:
                value = self._feature_cleaner.clean_value(key, value)
            # if the qualifier name exists, append the value
            if key in self._cur_feature.qualifiers:
                self._cur_feature.qualifiers[key].append(value)
            # otherwise start a new list of the key with its values
            else:
                self._cur_feature.qualifiers[key] = [value]

00838     def feature_qualifier_name(self, content_list):
        """When we get a qualifier key, use it as a dictionary key.
        
        We receive a list of keys, since you can have valueless keys such as
        /pseudo which would be passed in with the next key (since no other
        tags separate them in the file)
        """
        for content in content_list:
            # add a qualifier if we've got one
            self._add_qualifier()

            # remove the / and = from the qualifier if they're present
            qual_key = content.replace('/', '')
            qual_key = qual_key.replace('=', '')
            qual_key = qual_key.strip()
            
            self._cur_qualifier_key = qual_key
            self._cur_qualifier_value = []
        
    def feature_qualifier_description(self, content):
        # get rid of the quotes surrounding the qualifier if we've got 'em
        qual_value = content.replace('"', '')
        
        self._cur_qualifier_value.append(qual_value)

00863     def contig_location(self, content):
        """Deal with a location of CONTIG information.
        """
        from Bio import SeqFeature
        # add a last feature if is hasn't been added,
        # so that we don't overwrite it
        self._add_feature()

        # make a feature to add the information to
        self._cur_feature = SeqFeature.SeqFeature()
        self._cur_feature.type = "contig"
        
        # now set the location on the feature using the standard
        # location handler
        self.location(content)
        # add the contig information to the annotations and get rid
        # of the feature to prevent it from being added to the feature table
        self.data.annotations["contig"] = self._cur_feature
        self._cur_feature = None

    def origin_name(self, content):
        pass

    def base_count(self, content):
        pass

    def base_number(self, content):
        pass

00892     def sequence(self, content):
        """Add up sequence information as we get it.

        To try and make things speedier, this puts all of the strings
        into a list of strings, and then uses string.join later to put
        them together. Supposedly, this is a big time savings
        """
        new_seq = content.replace(' ', '')
        new_seq = new_seq.upper()

        self._seq_data.append(new_seq)

00904     def record_end(self, content):
        """Clean up when we've finished the record.
        """
        from Bio import Alphabet
        from Bio.Alphabet import IUPAC
        from Bio.Seq import Seq
        # add the last feature in the table which hasn't been added yet
        self._add_feature()

        # add the sequence information
        # first, determine the alphabet
        # we default to an generic alphabet if we don't have a
        # seq type or have strange sequence information.
        seq_alphabet = Alphabet.generic_alphabet

        if self._seq_type:
            # mRNA is really also DNA, since it is actually cDNA
            if self._seq_type.find('DNA') != -1 or \
               self._seq_type.find('mRNA') != -1:
                seq_alphabet = IUPAC.ambiguous_dna
            # are there every really RNA sequences in GenBank?
            elif self._seq_type.find('RNA') != -1:
                seq_alphabet = IUPAC.ambiguous_rna
            elif self._seq_type == "PROTEIN":
                seq_alphabet = IUPAC.protein  # or extended protein?
            # work around ugly GenBank records which have circular or
            # linear but no indication of sequence type
            elif self._seq_type in ["circular", "linear"]:
                pass
            # we have a bug if we get here
            else:
                raise ValueError("Could not determine alphabet for seq_type %s"
                                 % self._seq_type)

        # now set the sequence
        sequence = "".join(self._seq_data)
        self.data.seq = Seq(sequence, seq_alphabet)

00942 class _RecordConsumer(_BaseGenBankConsumer):
    """Create a GenBank Record object from scanner generated information.
    """
    def __init__(self):
        _BaseGenBankConsumer.__init__(self)
        import Record
        self.data = Record.Record()

        self._seq_data = []
        self._cur_reference = None
        self._cur_feature = None
        self._cur_qualifier = None
        
    def locus(self, content):
        self.data.locus = content

    def size(self, content):
        self.data.size = content

    def residue_type(self, content):
        self.data.residue_type = content

    def data_file_division(self, content):
        self.data.data_file_division = content

    def date(self, content):
        self.data.date = content

    def definition(self, content):
        self.data.definition = content

    def accession(self, content):
        new_accessions = self._split_accessions(content)
        self.data.accession.extend(new_accessions)

    def nid(self, content):
        self.data.nid = content

    def pid(self, content):
        self.data.pid = content

    def version(self, content):
        self.data.version = content

    def db_source(self, content):
        self.data.db_source = content.rstrip()

    def gi(self, content):
        self.data.gi = content

    def keywords(self, content):
        self.data.keywords = self._split_keywords(content)

    def segment(self, content):
        self.data.segment = content

    def source(self, content):
        self.data.source = content

    def organism(self, content):
        self.data.organism = content

    def taxonomy(self, content):
        self.data.taxonomy = self._split_taxonomy(content)

01007     def reference_num(self, content):
        """Grab the reference number and signal the start of a new reference.
        """
        # check if we have a reference to add
        if self._cur_reference is not None:
            self.data.references.append(self._cur_reference)

        self._cur_reference = Record.Reference()
        self._cur_reference.number = content

    def reference_bases(self, content):
        self._cur_reference.bases = content

    def authors(self, content):
        self._cur_reference.authors = content

    def consrtm(self, content):
        self._cur_reference.consrtm = content

    def title(self, content):
        self._cur_reference.title = content

    def journal(self, content):
        self._cur_reference.journal = content

    def medline_id(self, content):
        self._cur_reference.medline_id = content
        
    def pubmed_id(self, content):
        self._cur_reference.pubmed_id = content

    def remark(self, content):
        self._cur_reference.remark = content
        
    def comment(self, content):
        self.data.comment = "\n".join(content)

01044     def primary_ref_line(self,content):
        """Data for the PRIMARY line"""
        self.data.primary.append(content)

    def primary(self,content):
        pass
    
01051     def features_line(self, content):
        """Get ready for the feature table when we reach the FEATURE line.
        """
        self.start_feature_table()

01056     def start_feature_table(self):
        """Signal the start of the feature table.
        """
        # we need to add on the last reference
        if self._cur_reference is not None:
            self.data.references.append(self._cur_reference)

01063     def feature_key(self, content):
        """Grab the key of the feature and signal the start of a new feature.
        """
        # first add on feature information if we've got any
        self._add_feature()

        self._cur_feature = Record.Feature()
        self._cur_feature.key = content

01072     def _add_feature(self):
        """Utility function to add a feature to the Record.

        This does all of the appropriate checking to make sure we haven't
        left any info behind, and that we are only adding info if it
        exists.
        """
        if self._cur_feature is not None:
            # if we have a left over qualifier, add it to the qualifiers
            # on the current feature
            if self._cur_qualifier is not None:
                self._cur_feature.qualifiers.append(self._cur_qualifier)

            self._cur_qualifier = None
            self.data.features.append(self._cur_feature)

    def location(self, content):
        self._cur_feature.location = self._clean_location(content)

01091     def feature_qualifier_name(self, content_list):
        """Deal with qualifier names
        
        We receive a list of keys, since you can have valueless keys such as
        /pseudo which would be passed in with the next key (since no other
        tags separate them in the file)
        """
        for content in content_list:
            # the record parser keeps the /s -- add them if we don't have 'em
            if content.find("/") != 0:
                content = "/%s" % content
            # add on a qualifier if we've got one
            if self._cur_qualifier is not None:
                self._cur_feature.qualifiers.append(self._cur_qualifier)

            self._cur_qualifier = Record.Qualifier()
            self._cur_qualifier.key = content

    def feature_qualifier_description(self, content):
        # if we have info then the qualifier key should have a ='s
        if self._cur_qualifier.key.find("=") == -1:
            self._cur_qualifier.key = "%s=" % self._cur_qualifier.key
        cur_content = self._remove_newlines(content)
        # remove all spaces from the value if it is a type where spaces
        # are not important
        for remove_space_key in self.__class__.remove_space_keys:
            if self._cur_qualifier.key.find(remove_space_key) >= 0:
                cur_content = self._remove_spaces(cur_content)
        self._cur_qualifier.value = self._normalize_spaces(cur_content)

    def base_count(self, content):
        self.data.base_counts = content

    def origin_name(self, content):
        self.data.origin = content

01127     def contig_location(self, content):
        """Signal that we have contig information to add to the record.
        """
        self.data.contig = self._clean_location(content) 

01132     def sequence(self, content):
        """Add sequence information to a list of sequence strings.

        This removes spaces in the data and uppercases the sequence, and
        then adds it to a list of sequences. Later on we'll join this
        list together to make the final sequence. This is faster than
        adding on the new string every time.
        """
        new_seq = content.replace(' ', '')
        self._seq_data.append(new_seq.upper())

01143     def record_end(self, content):
        """Signal the end of the record and do any necessary clean-up.
        """
        # add together all of the sequence parts to create the
        # final sequence string
        self.data.sequence = "".join(self._seq_data)
        # add on the last feature
        self._add_feature()

def _strip_and_combine(line_list):
    """Combine multiple lines of content separated by spaces.

    This function is used by the EventGenerator callback function to
    combine multiple lines of information. The lines are first
    stripped to remove whitepsace, and then combined so they are separated
    by a space. This is a simple minded way to combine lines, but should
    work for most cases.
    """
    # first strip out extra whitespace
    stripped_line_list = [x.strip() for x in line_list]

    # now combine everything with spaces
    return ' '.join(stripped_line_list)

01167 class _Scanner:
    """Does the parsing of a GenBank file.  Earlier versions of this
    class used Martel to do this, but there where significant speed
    and memory limitations to this.
    """
01172     def __init__(self, debug = 0):
        """Initialize the scanner.

        Arguments:
        o debug - The level of debugging that the parser should
        display. Level 0 is no debugging, Level 2 displays the most
        debugging info (but is slower).
        """

        #Following dictionary maps GenBank lines to the associated
        #consumer methods - the special cases like LOCUS where one
        #genbank line triggers several consumer calls have to be
        #handled individually.
        self._consumer_dict = {
            'DEFINITION' : 'definition',
            'ACCESSION'  : 'accession',
            'NID'        : 'nid',
            'PID'        : 'pid',
            'DBSOURCE'   : 'db_source',
            'KEYWORDS'   : 'keywords',
            'SEGMENT'    : 'segment',
            'SOURCE'     : 'source',
            'AUTHORS'    : 'authors',
            'CONSRTM'    : 'consrtm',
            'TITLE'      : 'title',
            'JOURNAL'    : 'journal',
            'MEDLINE'    : 'medline_id',
            'PUBMED'     : 'pubmed_id',
            'REMARK'     : 'remark'}
        #We have to handle the following specially:
        #ORIGIN (locus, size, residue_type, data_file_division and date)
        #COMMENT (comment)
        #VERSION (version and gi)
        #REFERENCE (eference_num and reference_bases)
        #ORGANISM (organism and taxonomy)

        #Save the requested debug level
        self._debug = debug

01211     def _feed_locus(self, line, consumer):
        """Sub function for the '_feed_header' (and thus 'feed') method.
        
        Should not be invoked directly.  This deals with the possible
        variations of the LOCUS line which is passed in as a string
        argument. No return value"""
        #####################################
        # LOCUS line                        #
        #####################################
        assert line[0:GENBANK_INDENT] == 'LOCUS       ', \
               'LOCUS line does not start correctly:\n' + line

        #Have to break up the locus line, and handle the different bits of it.
        #There are at least two different versions of the locus line...
        if line[29:33] in [' bp ', ' aa '] :
            #Old...
            #
            #    Positions  Contents
            #    ---------  --------
            #    00:06      LOCUS
            #    06:12      spaces
            #    12:??      Locus name
            #    ??:??      space
            #    ??:29      Length of sequence, right-justified
            #    29:33      space, bp, space
            #    33:41      strand type
            #    41:42      space
            #    42:51      Blank (implies linear), linear or circular
            #    51:52      space
            #    52:55      The division code (e.g. BCT, VRL, INV)
            #    55:62      space
            #    62:73      Date, in the form dd-MMM-yyyy (e.g., 15-MAR-1991)
            #
            assert line[29:33] in [' bp ', ' aa '] , \
                   'LOCUS line does not contain size units at expected position:\n' + line
            assert line[41:42] == ' ', \
                   'LOCUS line does not contain space at position 42:\n' + line
            assert line[42:51].strip() in ['','linear','circular'], \
                   'LOCUS line does not contain valid entry (linear, circular, ...):\n' + line
            assert line[51:52] == ' ', \
                   'LOCUS line does not contain space at position 52:\n' + line
            assert line[55:62] == '       ', \
                   'LOCUS line does not contain spaces from position 56 to 62:\n' + line
            assert line[64:65] == '-', \
                   'LOCUS line does not contain - at position 65 in date:\n' + line
            assert line[68:69] == '-', \
                   'LOCUS line does not contain - at position 69 in date:\n' + line

            name_and_length_str = line[GENBANK_INDENT:29]
            while name_and_length_str.find('  ')<>-1 :
                name_and_length_str = name_and_length_str.replace('  ',' ')
            name_and_length = name_and_length_str.split(' ')
            assert len(name_and_length)<=2, \
                   'Cannot parse the name and length in the LOCUS line:\n' + line
            assert len(name_and_length)<>1, \
                   'Name and length collide in the LOCUS line:\n' + line
                   #Should be possible to split them based on position, if
                   #a clear definition of the standard exists THAT AGREES with
                   #existing files.
            consumer.locus(name_and_length[0])
            consumer.size(name_and_length[1])
            #consumer.residue_type(line[33:41].strip())
            consumer.residue_type(line[33:51].strip())
            consumer.data_file_division(line[52:55])
            consumer.date(line[62:73])
        elif line[40:44] in [' bp ', ' aa '] :
            #New...
            #
            #    Positions  Contents
            #    ---------  --------
            #    00:06      LOCUS
            #    06:12      spaces
            #    12:??      Locus name
            #    ??:??      space
            #    ??:40      Length of sequence, right-justified
            #    40:44      space, bp, space
            #    44:47      Blank, ss-, ds-, ms-
            #    47:54      Blank, DNA, RNA, tRNA, mRNA, uRNA, snRNA
            #    54:55      space
            #    55:63      Blank (implies linear), linear or circular
            #    63:64      space
            #    64:67      The division code (e.g. BCT, VRL, INV)
            #    67:68      space
            #    68:79      Date, in the form dd-MMM-yyyy (e.g., 15-MAR-1991)
            #
            assert line[40:44] in [' bp ', ' aa '] , \
                   'LOCUS line does not contain size units at expected position:\n' + line
            assert line[44:47] in ['   ', 'ss-', 'ds-', 'ms-'], \
                    'LOCUS line does not have valid strand type (Single stranded, ...):\n' + line
            assert line[47:54].strip() in ['','DNA','RNA','tRNA','mRNA','uRNA','snRNA'], \
                   'LOCUS line does not contain valid sequence type (DNA, RNA, ...):\n' + line
            assert line[54:55] == ' ', \
                   'LOCUS line does not contain space at position 55:\n' + line
            assert line[55:63].strip() in ['','linear','circular'], \
                   'LOCUS line does not contain valid entry (linear, circular, ...):\n' + line
            assert line[63:64] == ' ', \
                   'LOCUS line does not contain space at position 64:\n' + line
            assert line[67:68] == ' ', \
                   'LOCUS line does not contain space at position 68:\n' + line
            assert line[70:71] == '-', \
                   'LOCUS line does not contain - at position 71 in date:\n' + line
            assert line[74:75] == '-', \
                   'LOCUS line does not contain - at position 75 in date:\n' + line

            name_and_length_str = line[GENBANK_INDENT:40]
            while name_and_length_str.find('  ')<>-1 :
                name_and_length_str = name_and_length_str.replace('  ',' ')
            name_and_length = name_and_length_str.split(' ')
            assert len(name_and_length)<=2, \
                   'Cannot parse the name and length in the LOCUS line:\n' + line
            assert len(name_and_length)<>1, \
                   'Name and length collide in the LOCUS line:\n' + line
                   #Should be possible to split them based on position, if
                   #a clear definition of the stand exists THAT AGREES with
                   #existing files.
            consumer.locus(name_and_length[0])
            consumer.size(name_and_length[1])
            consumer.residue_type(line[44:63].strip())
            consumer.data_file_division(line[64:67])
            consumer.date(line[68:79])
        elif line[GENBANK_INDENT:].strip().count(" ")==0 : 
            #Truncated LOCUS line, as produced by some EMBOSS tools - see bug 1762
            #
            #e.g.
            #
            #    "LOCUS       U00096"
            #
            #rather than:
            #
            #    "LOCUS       U00096               4639675 bp    DNA     circular BCT"
            #
            #    Positions  Contents
            #    ---------  --------
            #    00:06      LOCUS
            #    06:12      spaces
            #    12:??      Locus name
            if line[GENBANK_INDENT:].strip() <> "" :
                consumer.locus(line[GENBANK_INDENT:].strip())
            else :
                #Must just have just "LOCUS       "
                #We can cope with this, but is it a legitimate GenBank file?
                print "Warning: Minimal LOCUS line found - is this correct?\n" + line
                pass
        else :
            assert False, \
                   'Did not recognise the LOCUS line layout:\n' + line
        
01358     def _feed_header(self, handle, consumer):
        """Sub function for the 'feed' method.

        Should not be invoked directly.  Returns a string, the last line
        read from the GenBank which marks the end of the header (normally
        be the FEATURE line )."""
        
        # skip ahead until we find first record
        line = handle.readline()
        while line.find('LOCUS       ') <> 0:
            assert line, \
                   'Unexpected end of file while looking for LOCUS line'
            if self._debug : print "Ignoring line:\n" + line
            line = handle.readline()

        if self._debug : print "Starting LOCUS"
        self._feed_locus(line, consumer)
        if self._debug : print "Finished LOCUS"

        #############################################################
        #now read in and cache JUST the header lines, up to and
        #excluding the features and nucleotide sequence..
        line = handle.readline()
        while True :
            assert line, \
                   'Unexpected end of file during GenBank header section'
            assert line[0:GENBANK_INDENT] <> GENBANK_SPACER, \
                   'Unexpected continuation of an entry:\n' + line

            if line.find("FEATURES             ") == 0 :
                #We are about to start the features...
                return line
            elif line.find('BASE COUNT')==0 :
                #This means we have finished the header, and
                #there is no list of features...
                return line
            elif line.find('ORIGIN')==0 :
                #This means we have finished the header, and
                #there is no list of features...
                return line
            elif line.find('CONTIG')==0 :
                #This means we have finished the header, and
                #there is no list of features...
                return line

            line_type = line[0:GENBANK_INDENT].strip()
            data = line[GENBANK_INDENT:]
            if data[-1:]=='\n' : data = data[:-1]
            if data[-1:]=='\r' : data = data[:-1]

            if line_type == 'VERSION' :
                #Need to call consumer.version(), and maybe also consumer.gi() as well.
                #e.g.
                # VERSION     AC007323.5  GI:6587720
                while data.find('  ')<>-1:
                    data = data.replace('  ',' ')
                if data.find(' GI:')==-1 :
                    consumer.version(data)
                else :
                    if self._debug : print "Version [" + data.split(' GI:')[0] + "], gi [" + data.split(' GI:')[1] + "]"
                    consumer.version(data.split(' GI:')[0])
                    consumer.gi(data.split(' GI:')[1])
                #Read in the next line!
                line = handle.readline()
            elif line_type == 'REFERENCE' :
                if self._debug >1 : print "Found reference [" + data + "]"
                #Need to call consumer.reference_num() and consumer.reference_bases()
                #e.g.
                # REFERENCE   1  (bases 1 to 86436)
                #
                #Note that this can be multiline, see Bug 1968, e.g.
                #
                # REFERENCE   42 (bases 1517 to 1696; 3932 to 4112; 17880 to 17975; 21142 to
                #             28259)
                #
                #For such cases we will call the consumer once only.
                data = data.strip()

                #Read in the next line, and see if its more of the reference:
                while True:
                    line = handle.readline()
                    if line[0:GENBANK_INDENT] == GENBANK_SPACER :
                        #Add this continuation to the data string
                        data = data + " " + line[GENBANK_INDENT:]
                        if data[-1:]=='\n' : data = data[:-1]
                        if data[-1:]=='\r' : data = data[:-1]
                        if self._debug >1 : print "Extended reference text [" + data + "]"
                    else :
                        #End of the reference, leave this text in the variable "line"
                        break

                #We now have all the reference line(s) stored in a string, data,
                #which we pass to the consumer
                while data.find('  ')<>-1:
                    data = data.replace('  ',' ')
                if data.find(' ')==-1 :
                    if self._debug >2 : print 'Reference number \"' + data + '\"'
                    consumer.reference_num(data)
                else :
                    if self._debug >2 : print 'Reference number \"' + data[:data.find(' ')] + '\", \"' + data[data.find(' ')+1:] + '\"'
                    consumer.reference_num(data[:data.find(' ')])
                    consumer.reference_bases(data[data.find(' ')+1:])
            elif line_type == 'ORGANISM' :
                #The first line is the organism, but subsequent lines go to the taxonomy consumer
                consumer.organism(data)
                data = ""
                while True :
                    line = handle.readline()
                    if line[0:GENBANK_INDENT] == GENBANK_SPACER :
                        data = data + ' ' + line[GENBANK_INDENT:]
                        if data[-1:]=='\n' : data = data[:-1]
                        if data[-1:]=='\r' : data = data[:-1]
                    else :
                        #We now have all the data for this taxonomy:
                        consumer.taxonomy(data.strip())
                        #End of continuation - return to top of loop!
                        break
            elif line_type == 'COMMENT' :
                if self._debug > 1 : print "Found comment"
                #This can be multiline, and should call consumer.comment() once
                #with a list where each entry is a line.
                list=[]
                list.append(data)
                while True:
                    line = handle.readline()
                    if line[0:GENBANK_INDENT] == GENBANK_SPACER :
                        data = line[GENBANK_INDENT:]
                        if data[-1:]=='\n' : data = data[:-1]
                        if data[-1:]=='\r' : data = data[:-1]
                        list.append(data)
                        if self._debug > 2 : print "Comment continuation [" + data + "]"
                    else :
                        #End of the comment
                        break
                consumer.comment(list)
                list=[]
            elif line_type in self._consumer_dict :
                #Its a semi-automatic entry!
                #Now, this may be a multi line entry...
                while True :
                    line = handle.readline()
                    if line[0:GENBANK_INDENT] == GENBANK_SPACER :
                        data = data + ' ' + line[GENBANK_INDENT:]
                        if data[-1:]=='\n' : data = data[:-1]
                        if data[-1:]=='\r' : data = data[:-1]
                    else :
                        #We now have all the data for this entry:
                        getattr(consumer, self._consumer_dict[line_type])(data)
                        #End of continuation - return to top of loop!
                        break
            else :
                #Its an unknown line type, the NCBI might have added
                #something new... see bug 1946
                #Print a warning, and ignore the line (and any continuation)
                print 'WARNING - Ignoring an unknown line type, ' + line_type + ' found:\n' + line
                while True :
                    line = handle.readline()
                    if line[0:GENBANK_INDENT] == GENBANK_SPACER :
                        #More of the unknown line type, could print another
                        #warning...
                        pass
                    else :
                        #Got to the end of it.
                        break
        #############################################################
        if self._debug : print "Found start of features..."
        assert line.find("FEATURES             ") == 0, \
               'Internal error - expected to be on the FEATURES line, not:\n' + line
        #############################################################

01528     def feed(self, handle, consumer):
        """Feed a set of data into the scanner.

        Arguments:
        o handle - A handle with the information to parse.
        o consumer - The consumer that should be informed of events.
        """

        #This will look for the LOCUS line, and then
        #continue until the end of the header is found.
        #This is usually the FEATURES line.
        line = self._feed_header(handle, consumer)

        if line[-1:]=='\n' : line = line[:-1]
        if line[-1:]=='\r' : line = line[:-1]

        if self._debug > 1 :
            print "self._feed_header returned :\n" + line

        if line.find("FEATURES             ") <> 0 :
            #It is abnormal, but allowed, that there is no list
            #of features... e.g. output from EMBOSS seqret, See bug 1762
            if self._debug : print "WARNING - Header was not followed by FEATURES, but by:\n" + line
        else :
            #At this point, the FEATURES line has normally just been
            #read in. Its doesn't actually have any usefull content
            #except as a "section divider"
            consumer.features_line(line)
            #Now parse the features table...
            line = handle.readline()
            if line[-1:]=='\n' : line = line[:-1]
            if line[-1:]=='\r' : line = line[:-1]
            while True :
                assert line, \
                       'Unexpected blank line/end of file during GenBank features section'
                assert line[0:FEATURE_QUALIFIER_INDENT]<>FEATURE_QUALIFIER_SPACER, \
                       'Unexpected continuation of feature:\n' + line
                if line[0:FEATURE_KEY_INDENT]<>FEATURE_KEY_SPACER :
                    #This should be the BASE COUNT, ORIGIN or CONTIG line now
                    break
                
                #So, start of a new feature then.
                #This line should have a feature_key and a location string
                #(note the location could span several lines!)

                #Extract the key...
                feature_key = line[FEATURE_KEY_INDENT:FEATURE_QUALIFIER_INDENT].strip()
                consumer.feature_key(feature_key)

                #Extract the location...
                feature_location = line[FEATURE_QUALIFIER_INDENT:].strip()
                while feature_location[-1:]=="," :
                    #Still more to come!
                    line = handle.readline()
                    assert line[0:FEATURE_QUALIFIER_INDENT]==FEATURE_QUALIFIER_SPACER, \
                           'Expected continuation of location, not:\n' + line
                    line = line[FEATURE_QUALIFIER_INDENT:]
                    if line[-1:]=='\n' : line = line[:-1]
                    if line[-1:]=='\r' : line = line[:-1]
                    feature_location = feature_location + line
                #############################################################
                if self._debug > 1 : print "Starting " + feature_key + " feature at location " + feature_location
                #############################################################
                consumer.location(feature_location)
                #We have dealt with that line (key and location, plus any lines continuing of the location).
                #The next line could be the first of one or more qualifiers to the feature:
                #
                #Can have single line entries like this:
                #                     /pseudo                (no information)
                #                     /name=Fred             (unquoted)
                #                     /name="Fred Jones"     (quoted)
                #
                #Can have quoted multiline entries like this:
                #                     /name="Fred
                #                     Jones"
                #
                #Or even like this (see bug 1903):
                #                     /name="
                #                     Fred Jones"
                #
                #We can also have unquoted multiline entries (bug 1758):
                #                     /transl_except=(pos:complement(18111697..18111699),
                #                     aa:OTHER)
                #
                #Note that you can get text of the form /word= which are INSIDE a
                #quoted entry, which does not mean its a new feature key!
                #(Any examples to hand?)
                #
                #Also note that the NCBI have sometimes used unescaped quotes inside
                #the text, for example it was reported on the mailing list that one
                #version of NC_000913 had the following:
                #                     /note="2'-(5"-phosphoribosyl)-3'-dephospho-CoA...
                #The 02-DEC-2005 version of the citX gene had been changed to:
                #                     /note="2'-(5'-phosphoribosyl)-3'-dephospho-CoA...
                #
                #We have also seen blank lines inside a multiline feature (see bug 1942)
                line = handle.readline()
                if line[-1:]=='\n' : line = line[:-1]
                if line[-1:]=='\r' : line = line[:-1]
                while line[0:FEATURE_QUALIFIER_INDENT]==FEATURE_QUALIFIER_SPACER :
                    #This SHOULD be the start of a new qualifier for the current feature
                    line = line[FEATURE_QUALIFIER_INDENT:]
                    if line[0:1]<>'/' :
                        #This is an unquoted multiline feature key, as reported in bug 1758
                        print "WARNING - Unquoted multiline '%s' entry for %s feature with location %s" \
                              % (qualifier_name, feature_key, feature_location)
                        #This will append the line to the description parsed so far:
                        consumer.feature_qualifier_description(line)    
                    elif line.find("=") == -1 :
                        #Can have qualifiers with no data, like /psuedo
                        qualifier_name = line[1:]
                        consumer.feature_qualifier_name([qualifier_name])
                        #There is no description in this case
                    else :
                        #Expect format /name=... or /name="...
                        #or in odd cases, /name="\n... see bug 1903
                        qualifier_name = line[1:line.find('=')]
                        consumer.feature_qualifier_name([qualifier_name])
                        qualifier_description = line[line.find("=")+1:]
                        #Try and get all of the description now, so we can make a single
                        #call to the consumer.  Making less function calls should be
                        #slightly faster.
                        if qualifier_description=='\"' \
                        or ( qualifier_description[0:1]=='\"' and qualifier_description[-1:]<>'\"' ) :
                            #There should now be one or more lines continuing the description
                            while True :
                                line = handle.readline()
                                if line[-1:]=='\n' : line = line[:-1]
                                if line[-1:]=='\r' : line = line[:-1]
                                while line=="" :
                                    #See bug 1942, blank line in feature with malformed /note entry
                                    print "WARNING - Blank line in '%s' entry for %s feature with location %s" \
                                          % (qualifier_name, feature_key, feature_location)
                                    #Ignore it, carry on
                                    line = handle.readline()
                                    if not line : break
                                    if line[-1:]=='\n' : line = line[:-1]
                                    if line[-1:]=='\r' : line = line[:-1]
                                assert line[0:FEATURE_QUALIFIER_INDENT]==FEATURE_QUALIFIER_SPACER, \
                                       "Expected qualifier description continuation in " \
                                       + "'%s' entry for %s feature with location %s, not:\n%s" \
                                       % (qualifier_name, feature_key, feature_location, line)
                                #Note, for backwards compatibility we do not remove the FEATURE_QUALIFIER_SPACER
                                #from the description
                                qualifier_description = qualifier_description + '\n' + line
                                if qualifier_description[-1:]=='\"' :
                                    #That should be the end of the description continuation
                                    break
                        consumer.feature_qualifier_description(qualifier_description)
                    #We have dealt with that line (qualifier and description, plus any line continuing the description).
                    #There could be another qualifier for this feature...
                    line = handle.readline()
                    if line[-1:]=='\n' : line = line[:-1]
                    if line[-1:]=='\r' : line = line[:-1]
                #We have dealt with all the qualifiers (if any) for this feature
                #Now move on to the next line...

            #############################################################
            if self._debug : print "Finished features"
            #############################################################
            #We should now be on the BASE COUNT, ORIGIN or CONTIG line
        
        if line.find('BASE COUNT')==0 :
            #Hmm.
            line = line[10:].strip()
            if self._debug : print "base_count = " + line
            consumer.base_count(line)
            line = handle.readline()
            if line[-1:]=='\n' : line = line[:-1]
            if line[-1:]=='\r' : line = line[:-1]

        if line.find("ORIGIN")==0 :
            #############################################################
            if self._debug : print "Starting sequence..."
            #############################################################
            #May need to call consumer for origin_name if part of ORIGIN line...
            line = line[6:].strip()
            if line :
                if self._debug : print "origin_name = " + line
                consumer.origin_name(line)
            #Now just consume the sequence lines until reach the // marker
            #or a CONTIG line
            while True :
                line = handle.readline()
                if line[-1:]=='\n' : line = line[:-1]
                if line[-1:]=='\r' : line = line[:-1]
                assert line, \
                       'Unexpected blank line/end of file during GenBank ORIGIN section'
                if line=='//' :
                    break
                if line.find('CONTIG')==0 :
                    break
                assert line[9:10]==' ', \
                       'Sequence line mal-formed, \n' + line
                consumer.base_number(line[0:10].strip())
                consumer.sequence(line[10:])
        
        if line.find("CONTIG")==0 :
            #############################################################
            if self._debug : print "Starting Contig..."
            #############################################################
            line = line[6:].strip()
            contig_location = line + '\n'
            while True :
                line = handle.readline()
                if line[-1:]=='\n' : line = line[:-1]
                if line[-1:]=='\r' : line = line[:-1]
                assert line, \
                       'Unexpected blank line/end of file during GenBank CONTIG section'
                if line=='//' :
                    consumer.contig_location(contig_location)
                    break
                elif line[:GENBANK_INDENT]==GENBANK_SPACER :
                    contig_location = contig_location + line
                else:
                    consumer.contig_location(contig_location)
                    assert False, \
                           'Expected CONTIG continuation line or // marker, got\n' + line
                    break

        if line=='//' :
            #############################################################
            if self._debug : print "Found record end..."
            #############################################################
            consumer.record_end(line)
            return

        assert False, \
               'Unexpected line near end of file:\n' + line
        return

def index_file(filename, indexname, rec2key = None, use_berkeley = 0):
    """Index a GenBank file to prepare it for use as a dictionary.

    Arguments:
    filename - The name of the GenBank file to be indexed.
    indexname - The name of the index to create
    rec2key - A reference to a function object which, when called with a 
    SeqRecord object, will return a key to be used for the record. If no 
    function is specified then the records will be indexed by the 'id'
    attribute of the SeqRecord (the versioned GenBank id).
    use_berkeley - specifies whether to use the BerkeleyDB indexer, which 
    uses the bsddb3 wrappers around the embedded database Berkeley DB. By
    default, the standard flat file (non-Berkeley) indexes are used.
    """
    if rec2key:
        indexer = SimpleSeqRecord.FunctionIndexer(rec2key)
    else:
        indexer = SimpleSeqRecord.SimpleIndexer()

    if use_berkeley:
        SimpleSeqRecord.create_berkeleydb([filename], indexname, indexer)
    else:
        SimpleSeqRecord.create_flatdb([filename], indexname, indexer)

01783 class NCBIDictionary:
    """Access GenBank using a read-only dictionary interface.
    """
    VALID_DATABASES = ['nucleotide', 'protein']
    VALID_FORMATS = ['genbank', 'fasta']
01788     def __init__(self, database, format, parser = None):
        """Initialize an NCBI dictionary to retrieve sequences.

        Create a new Dictionary to access GenBank.  Valid values for
        database are 'nucleotide' and 'protein'.
        Valid values for format are 'genbank' (for nucleotide genbank and
        protein genpept) and 'fasta'.
        dely and retmax are old options kept only for compatibility -- do not
        bother to set them.
        parser is an optional parser object
        to change the results into another form.  If unspecified, then
        the raw contents of the file will be returned.
        """
        self.parser = parser
        if database not in self.__class__.VALID_DATABASES:
            raise ValueError("Invalid database %s, should be one of %s" %
                    (database, self.__class__.VALID_DATABASES))
        if format not in self.__class__.VALID_FORMATS:
            raise ValueError("Invalid format %s, should be one of %s" %
                    (format, self.__class__.VALID_FORMATS))

        if format == 'fasta':
            self.db = db["fasta-sequence-eutils"]
        elif format == 'genbank':
            if database == 'nucleotide':
                self.db = db["nucleotide-genbank-eutils"]
            elif database == 'protein':
                self.db = db["protein-genbank-eutils"]

    def __len__(self):
        raise NotImplementedError, "GenBank contains lots of entries"
    def clear(self):
        raise NotImplementedError, "This is a read-only dictionary"
    def __setitem__(self, key, item):
        raise NotImplementedError, "This is a read-only dictionary"
    def update(self):
        raise NotImplementedError, "This is a read-only dictionary"
    def copy(self):
        raise NotImplementedError, "You don't need to do this..."
    def keys(self):
        raise NotImplementedError, "You don't really want to do this..."
    def items(self):
        raise NotImplementedError, "You don't really want to do this..."
    def values(self):
        raise NotImplementedError, "You don't really want to do this..."
    
01834     def has_key(self, id):
        """S.has_key(id) -> bool"""
        try:
            self[id]
        except KeyError:
            return 0
        return 1

    def get(self, id, failobj=None):
        try:
            return self[id]
        except KeyError:
            return failobj
        raise "How did I get here?"

01849     def __getitem__(self, id):
        """Return the GenBank entry specified by the GenBank ID.
        
        Raises a KeyError if there's an error.
        """
        handle = self.db[id]
        # Parse the record if a parser was passed in.
        if self.parser is not None:
            return self.parser.parse(handle)
        return handle.read()

def search_for(search, database='nucleotide',
               reldate=None, mindate=None, maxdate=None,
               start_id = 0, max_ids = 50000000):
    """search_for(search[, reldate][, mindate][, maxdate]
    [, batchsize][, delay][, callback_fn][, start_id][, max_ids]) -> ids

    Search GenBank and return a list of the GenBank identifiers (gi's)
    that match the criteria.  search is the search string used to
    search the database.  Valid values for database are
    'nucleotide', 'protein', 'popset' and 'genome'.  reldate is
    the number of dates prior to the current date to restrict the
    search.  mindate and maxdate are the dates to restrict the search,
    e.g. 2002/01/01.  start_id is the number to begin retrieval on.
    max_ids specifies the maximum number of id's to retrieve.
    
    batchsize, delay and callback_fn are old parameters for
    compatibility -- do not set them.
    """
    # deal with dates
    date_restrict = None
    if reldate:
        date_restrict = EUtils.WithinNDays(reldate)
    elif mindate:
        date_restrict = EUtils.DateRange(mindate, maxdate)

    eutils_client = DBIdsClient.DBIdsClient()
    db_ids = eutils_client.search(search, database, daterange = date_restrict,
            retstart = start_id, retmax = max_ids)
    ids = []
    for db_id in db_ids:
        ids.append(db_id.dbids.ids[0])
    return ids

def download_many(ids, database = 'nucleotide'):
    """download_many(ids, database) -> handle of results

    Download many records from GenBank.  ids is a list of gis or
    accessions.  

    callback_fn, broken_fn, delay, faildelay, batchsize, parser are old
    parameter for compatibility. They should not be used.
    """
    db_ids = DBIds(database, ids)
    if database in ['nucleotide']:
        format = 'gb'
    elif database in ['protein']:
        format = 'gp'
    else:
        raise ValueError("Unexpected database: %s" % database)

    eutils_client = DBIdsClient.from_dbids(db_ids)
    result_handle = eutils_client.efetch(retmode = "text", rettype = format)
    return cStringIO.StringIO(result_handle.read())


Generated by  Doxygen 1.6.0   Back to index