fix EBI-ENA submission with already submitted samples

qiita_ware/commands.py

@@ -207,7 +207,7 @@ def submit_EBI(artifact_id, action, send, test=False, test_size=False):
                  ebi_submission.submission_xml_fp]
     total_size = sum([stat(tr).st_size for tr in to_review if tr is not None])
     # note that the max for EBI is 10M but let's play it safe
-    max_size = 8.5e+6 if not test_size else 6000
+    max_size = 10e+6 if not test_size else 5000
     if total_size > max_size:
         LogEntry.create(
             'Runtime', 'The submission: %d is larger than allowed (%d), will '
@@ -228,10 +228,12 @@ def submit_EBI(artifact_id, action, send, test=False, test_size=False):
         cols_to_drop = cols_to_drop - {'taxon_id', 'scientific_name',
                                        'description'}
         all_samples = ebi_submission.sample_template.ebi_sample_accessions
-        samples = {k: all_samples[k] for k in ebi_submission.samples}
-        ebi_submission.write_xml_file(
-            ebi_submission.generate_sample_xml(samples, cols_to_drop),
-            ebi_submission.sample_xml_fp)
+        samples = {k: all_samples[k] for k in ebi_submission.samples
+                   if all_samples[k] is None}
+        if samples:
+            ebi_submission.write_xml_file(
+                ebi_submission.generate_sample_xml(samples, cols_to_drop),
+                ebi_submission.sample_xml_fp)
 
         # now let's recalculate the size to make sure it's fine
         new_total_size = sum([stat(tr).st_size

qiita_ware/ebi.py

@@ -393,7 +393,7 @@ def generate_sample_xml(self, samples=None, ignore_columns=None):
             The list of samples to be included in the sample xml. If not
             provided or an empty list is provided, all the samples are used
         ignore_columns : list of str, optional
-            The list of columns to ignore during submission; helful for when
+            The list of columns to ignore during submission; helpful for when
             the submissions are too large
 
         Returns
@@ -411,14 +411,18 @@ def generate_sample_xml(self, samples=None, ignore_columns=None):
         for sample_name in sorted(samples):
             sample_info = dict(self.samples[sample_name])
 
-            if self._ebi_sample_accessions[sample_name] is None:
-                sample = ET.SubElement(sample_set, 'SAMPLE', {
-                    'alias': self._get_sample_alias(sample_name),
-                    'center_name': qiita_config.ebi_center_name}
-                )
+            sample_accession = self._ebi_sample_accessions[sample_name]
+            if self.action in ('ADD', 'VALIDATE'):
+                if sample_accession is not None:
+                    continue
+                else:
+                    sample = ET.SubElement(sample_set, 'SAMPLE', {
+                        'alias': self._get_sample_alias(sample_name),
+                        'center_name': qiita_config.ebi_center_name}
+                    )
             else:
                 sample = ET.SubElement(sample_set, 'SAMPLE', {
-                    'accession': self._ebi_sample_accessions[sample_name],
+                    'accession': sample_accession,
                     'center_name': qiita_config.ebi_center_name}
                 )
 

qiita_ware/test/test_commands.py

@@ -236,8 +236,6 @@ def test_max_ebiena_curl_error(self):
 
         rmtree(join(self.base_fp, '%d_ebi_submission' % aid), True)
 
-    submit_EBI
-
 
 FASTA_EXAMPLE = """>1.SKB2.640194_1 X orig_bc=X new_bc=X bc_diffs=0
 CCACCCAGTAAC

qiita_ware/test/test_ebi.py

@@ -177,7 +177,9 @@ def test_generate_sample_xml(self):
 
         samples = ['1.SKB2.640194', '1.SKB3.640195']
         obs = ET.tostring(submission.generate_sample_xml(samples=samples))
-        exp = ''.join([l.strip() for l in SAMPLEXML.splitlines()])
+        exp = ('<SAMPLE_SET xmlns:xsi="http://www.w3.org/2001/XMLSchema-'
+               'instance" xsi:noNamespaceSchemaLocation="ftp://ftp.sra.ebi.ac.'
+               'uk/meta/xsd/sra_1_3/SRA.sample.xsd" />')
         self.assertEqual(obs.decode('ascii'), exp)
 
         # removing samples so test text is easier to read
@@ -194,7 +196,6 @@ def test_generate_sample_xml(self):
             del(submission.samples[k])
             del(submission.samples_prep[k])
         obs = ET.tostring(submission.generate_sample_xml())
-        exp = ''.join([l.strip() for l in SAMPLEXML.splitlines()])
         self.assertEqual(obs.decode('ascii'), exp)
 
         obs = ET.tostring(submission.generate_sample_xml(samples=[]))