Merge pull request #47 from andrewjpage/wrap_header_lines

Wrap publication lines
sanger-pathogens · Dec 10, 2015 · f861acd · f861acd
2 parents 5f41452 + d22f31f
commit f861acd
Show file tree

Hide file tree

Showing 4 changed files with 82 additions and 9 deletions.
diff --git a/README.md b/README.md
@@ -1,9 +1,9 @@
 # gff3toembl
-Converts gff3 files to emble files for uploading to EBI.
+Converts PROKKA GFF3 files to EMBL files for uploading annotated assemblies to EBI.
 
 [![Build Status](https://travis-ci.org/sanger-pathogens/gff3toembl.svg?branch=master)](https://travis-ci.org/sanger-pathogens/gff3toembl)
 
-NB this implements some EBI specific conventions and is not a generic conversion tool.
+NB this implements some EBI specific conventions and is not a generic conversion tool. Its also not a validator, so you need to pass in parameters which are acceptable to EMBL.
 
 ## Installation
 gff3toembl only works with Python 2.7 and there are known issues on other versions of Python.

diff --git a/gff3toembl/EMBLContig.py b/gff3toembl/EMBLContig.py
@@ -277,32 +277,49 @@ def __init__(self,
     source_attributes = self.build_source_attributes(organism, taxon_id, sequence_name)
     self.source_feature = EMBLFeature(feature_type='source', start=1, end=sequence_length,
                                       strand='+', feature_attributes=source_attributes)
-
     self.header_template = """\
 ID   XXX; XXX; {genome_type}; genomic DNA; STD; {classification}; {sequence_length} BP.
 XX
 AC   XXX;
 XX
 AC * _{sequence_identifier}
 XX
-PR   Project:{project};
+{project_line}
 XX
 DE   XXX;
 XX
 RN   [1]
-RA   {authors};
-RT   "{title}";
-RL   {publication}.
+{publication_authors}
+{publication_title}
+{publication_name}
 XX
 FH   Key             Location/Qualifiers
 FH
 """
 
+  def header_attribute_formatter(self, key, header_text, quote_character, final_character):
+    wrapper = TextWrapper()
+    wrapper.initial_indent=key + '   '
+    wrapper.subsequent_indent=key + '   '
+    wrapper.width=79
+    attribute_text_template='{attribute_quote_character}{attribute_header_text}{attribute_quote_character}{attribute_final_character}'
+    attribute_text=attribute_text_template.format(attribute_header_text = header_text, 
+                                                  attribute_quote_character = quote_character, 
+                                                  attribute_final_character = final_character)
+    return wrapper.fill(attribute_text)
+
   def remove_non_word_characters(self, sequence_identifier):
     return re.sub(r'\W+', '', sequence_identifier)
 
   def format(self):
-    return self.header_template.format(**self.__dict__) + self.source_feature.format()
+    project_line = self.header_attribute_formatter("PR", "Project:" + self.project, '', ';' )                              
+    publication_authors = self.header_attribute_formatter("RA", self.authors,'',';' )
+    publication_title   = self.header_attribute_formatter("RT", self.title,'"',';' )
+    publication_name    = self.header_attribute_formatter("RL", self.publication,'','.' )
+    return self.header_template.format(project_line        = project_line, 
+                                       publication_authors = publication_authors, 
+                                       publication_title   = publication_title, 
+                                       publication_name    = publication_name,  **self.__dict__) + self.source_feature.format()
 
   def build_source_attributes(self, organism, taxon_id, sequence_name):
     def empty_string_if_none(value):

diff --git a/gff3toembl/tests/EMBLContig_test.py b/gff3toembl/tests/EMBLContig_test.py
@@ -294,6 +294,9 @@ def test_format(self):
     for calculated_line,expected_line in zip(calculated_header.split('\n'), expected_header.split('\n')):
       self.assertEqual(calculated_line, expected_line)
     self.assertEqual(len(calculated_header), len(expected_header))
+
+
+
 
   def test_format_long_organism_name(self):
     header = EMBLHeader()
@@ -336,6 +339,59 @@ def test_format_long_organism_name(self):
 FT                   /mol_type="genomic DNA"
 FT                   /db_xref="taxon:5678"
 FT                   /note="chromX"
+"""
+
+    calculated_header = header.format()
+    for calculated_line,expected_line in zip(calculated_header.split('\n'), expected_header.split('\n')):
+      self.assertEqual(calculated_line, expected_line)
+    self.assertEqual(len(calculated_header), len(expected_header))
+
+
+  def test_format_long_publication(self):
+    header = EMBLHeader()
+
+    header.authors="Jagger M., Richards K., Watts C., Wood R., Jones B., Stewart I., Wyman B., Taylor M."
+    header.classification="UNC"
+    header.genome_type="circular"
+    header.organism="organism"
+    header.project="PRJ1234"
+    header.publication="The Rolling Stones, 12 X 5, The Rolling Stones No. 2, Out of Our Heads, Aftermath, Between the Buttons, Their Satanic Majesties Request, Beggars Banquet"
+    header.sequence_identifier="contig123"
+    header.sequence_length=1234
+    header.sequence_name="chromX"
+    header.taxon_id=5678
+    header.title="Let It Bleed, Sticky Fingers, Exile on Main St., Goats Head Soup, It's Only Rock 'n Roll, Black and Blue, Some Girls, Emotional Rescue, Tattoo You, Undercover"
+    source_attributes = {"organism": header.organism, "db_xref": "taxon:5678", "note": "chromX"}
+    header.source_feature = EMBLFeature('source', 1, 1234, '+', source_attributes)
+
+    expected_header = """\
+ID   XXX; XXX; circular; genomic DNA; STD; UNC; 1234 BP.
+XX
+AC   XXX;
+XX
+AC * _contig123
+XX
+PR   Project:PRJ1234;
+XX
+DE   XXX;
+XX
+RN   [1]
+RA   Jagger M., Richards K., Watts C., Wood R., Jones B., Stewart I., Wyman B.,
+RA   Taylor M.;
+RT   "Let It Bleed, Sticky Fingers, Exile on Main St., Goats Head Soup, It's
+RT   Only Rock 'n Roll, Black and Blue, Some Girls, Emotional Rescue, Tattoo
+RT   You, Undercover";
+RL   The Rolling Stones, 12 X 5, The Rolling Stones No. 2, Out of Our Heads,
+RL   Aftermath, Between the Buttons, Their Satanic Majesties Request, Beggars
+RL   Banquet.
+XX
+FH   Key             Location/Qualifiers
+FH
+FT   source          1..1234
+FT                   /organism="organism"
+FT                   /mol_type="genomic DNA"
+FT                   /db_xref="taxon:5678"
+FT                   /note="chromX"
 """
 
     calculated_header = header.format()

diff --git a/setup.py b/setup.py
@@ -8,7 +8,7 @@ def read(fname):
 
 setup(
     name='gff3toembl',
-    version='1.0.1',
+    version='1.0.2',
     description='Convert a GFF3 file to EMBL format for submission',
     long_description=read('README.md'),
     packages = find_packages(),