mirror of
https://github.com/galaxyproject/galaxy.git
synced 2026-10-24 12:13:33 +08:00
Allow sorting collection elements based on file
Given a collection with element_identifiers: ```yml element_1 element_2 ``` we can arbitrarily re-order collection elements using a dataset that specifies the new order, e.g ``` element_2 element_1 ``` I think this nicely complements numeric and alphanumeric sorts that are already possible with this tool.
This commit is contained in:
mvdbeek
@@ -9,6 +9,7 @@ import tarfile
|
||||
import tempfile
|
||||
import threading
|
||||
from cgi import FieldStorage
|
||||
from collections import OrderedDict
|
||||
from datetime import datetime
|
||||
from distutils.version import LooseVersion
|
||||
from xml.etree import ElementTree
|
||||
@@ -2497,14 +2498,28 @@ class SortTool(DatabaseOperationTool):
|
||||
|
||||
def produce_outputs(self, trans, out_data, output_collections, incoming, history):
|
||||
hdca = incoming["input"]
|
||||
sorttype = incoming["sort_type"]
|
||||
sorttype = incoming["sort_type"]["sort_type"]
|
||||
new_elements = odict()
|
||||
elements = hdca.collection.elements
|
||||
presort_elements = []
|
||||
if sorttype == 'alpha':
|
||||
presort_elements = [(dce.element_identifier, dce) for dce in elements]
|
||||
elif sorttype == 'numeric':
|
||||
presort_elements = [(int(re.sub('[^0-9]', '', dce.element_identifier)), dce) for dce in elements]
|
||||
sorted_elements = [x[1] for x in sorted(presort_elements, key=lambda x: x[0])]
|
||||
if presort_elements:
|
||||
sorted_elements = [x[1] for x in sorted(presort_elements, key=lambda x: x[0])]
|
||||
if sorttype == 'file':
|
||||
hda = incoming["sort_type"]["sort_file"]
|
||||
if hda.metadata and hda.metadata.get('data_lines', 0) == len(elements):
|
||||
old_elements_dict = OrderedDict()
|
||||
for element in elements:
|
||||
old_elements_dict[element.element_identifier] = element
|
||||
try:
|
||||
sorted_elements = [old_elements_dict[line.strip()] for line in open(hda.file_name)]
|
||||
except KeyError:
|
||||
hdca_history_name = "%s: %s" (hdca.hid, hdca.name)
|
||||
message = "List of element identifiers does not match element identifiers in collection '%s'" % hdca_history_name
|
||||
raise Exception(message)
|
||||
|
||||
for dce in sorted_elements:
|
||||
dce_object = dce.element_object
|
||||
|
||||
@@ -8,10 +8,16 @@
|
||||
class="ModelOperationToolAction"/>
|
||||
<inputs>
|
||||
<param type="data_collection" collection_type="list" name="input" label="Input Collection" />
|
||||
<param type="select" name="sort_type" label="Sort collection identifiers" help="">
|
||||
<option value="alpha">alphabetically</option>
|
||||
<option value="numeric">numerically (strips all characters except numbers)</option>
|
||||
</param>
|
||||
<conditional name="sort_type">
|
||||
<param type="select" name="sort_type" label="Sort collection identifiers" help="All element identifiers must be present once.">
|
||||
<option value="alpha">alphabetically</option>
|
||||
<option value="numeric">numerically (strips all characters except numbers)</option>
|
||||
<option value="file">Sort collection using order of identifiers in text file</option>
|
||||
</param>
|
||||
<when value="file">
|
||||
<param name="sort_file" type="data" format="txt" label="Select file that contains the new order of collection elements" help="Lines in text file must match collection"></param>
|
||||
</when>
|
||||
</conditional>
|
||||
</inputs>
|
||||
<outputs>
|
||||
<collection name="output" format_source="input" type="list" label="${on_string} (sorted)" >
|
||||
@@ -19,45 +25,68 @@
|
||||
</outputs>
|
||||
<tests>
|
||||
<test>
|
||||
<!-- test that we can reorder a collection using a file with element identifiers-->
|
||||
<param name="input">
|
||||
<collection type="list">
|
||||
<element name="def" value="simple_line_alternative.txt" />
|
||||
<element name="abc" value="simple_line.txt" />
|
||||
<element name="element_1" value="simple_line_alternative.txt" />
|
||||
<element name="element_2" value="simple_line.txt" />
|
||||
</collection>
|
||||
</param>
|
||||
<param name="sort_type" value="alpha" />
|
||||
<param name="sort_type" value="file"/>
|
||||
<param name="sort_file" value="new_element_sort_order.txt"/>
|
||||
<output_collection name="output" type="list">
|
||||
<element name="abc">
|
||||
<assert_contents>
|
||||
<has_text_matching expression="^This is a line of text.\n$" />
|
||||
</assert_contents>
|
||||
</element>
|
||||
<element name="def">
|
||||
<assert_contents>
|
||||
<has_text_matching expression="^This is a different line of text.\n$" />
|
||||
</assert_contents>
|
||||
</element>
|
||||
<element name="element_2">
|
||||
<assert_contents>
|
||||
<has_text_matching expression="^This is a line of text.\n$" />
|
||||
</assert_contents>
|
||||
</element>
|
||||
<element name="element_1">
|
||||
<assert_contents>
|
||||
<has_text_matching expression="^This is a different line of text.\n$" />
|
||||
</assert_contents>
|
||||
</element>
|
||||
</output_collection>
|
||||
</test>
|
||||
<test>
|
||||
<param name="input">
|
||||
<collection type="list">
|
||||
<element name="def0123400" value="simple_line_alternative.txt" />
|
||||
<element name="abc5678" value="simple_line.txt" />
|
||||
<element name="def" value="simple_line_alternative.txt" />
|
||||
<element name="abc" value="simple_line.txt" />
|
||||
</collection>
|
||||
</param>
|
||||
<param name="sort_type" value="numeric" />
|
||||
<param name="sort_type" value="alpha" />
|
||||
<output_collection name="output" type="list">
|
||||
<element name="abc5678">
|
||||
<assert_contents>
|
||||
<has_text_matching expression="^This is a line of text.\n$" />
|
||||
</assert_contents>
|
||||
</element>
|
||||
<element name="def0123400">
|
||||
<assert_contents>
|
||||
<has_text_matching expression="^This is a different line of text.\n$" />
|
||||
</assert_contents>
|
||||
</element>
|
||||
<element name="abc">
|
||||
<assert_contents>
|
||||
<has_text_matching expression="^This is a line of text.\n$" />
|
||||
</assert_contents>
|
||||
</element>
|
||||
<element name="def">
|
||||
<assert_contents>
|
||||
<has_text_matching expression="^This is a different line of text.\n$" />
|
||||
</assert_contents>
|
||||
</element>
|
||||
</output_collection>
|
||||
</test>
|
||||
<test>
|
||||
<param name="input">
|
||||
<collection type="list">
|
||||
<element name="def0123400" value="simple_line_alternative.txt" />
|
||||
<element name="abc5678" value="simple_line.txt" />
|
||||
</collection>
|
||||
</param>
|
||||
<param name="sort_type" value="numeric" />
|
||||
<output_collection name="output" type="list">
|
||||
<element name="abc5678">
|
||||
<assert_contents>
|
||||
<has_text_matching expression="^This is a line of text.\n$" />
|
||||
</assert_contents>
|
||||
</element>
|
||||
<element name="def0123400">
|
||||
<assert_contents>
|
||||
<has_text_matching expression="^This is a different line of text.\n$" />
|
||||
</assert_contents>
|
||||
</element>
|
||||
</output_collection>
|
||||
</test>
|
||||
</tests>
|
||||
|
||||
@@ -0,0 +1,2 @@
|
||||
element_2
|
||||
element_1
|
||||
Reference in New Issue
Block a user