Handing bedfile headers (#50)

* modified the docstring of the states_bedline_to_vector function

- Added more examples and descriptions to make ir more informative

* Added test for the states format option

* Fixed syntax error

* fixed syntax error 2

* Allows row_infos multiple columns

* Added code to identify and ignore bedfiles headers

* Removed --has-header option

* clean up the encode/decode conversions

* Fix line that handles bedfile headers

* Added test for ignoring bedfile headers

* Added bedfile with headers testfile

* Revert "Removed --has-header option"

This reverts commit 51a1af6. Brings back the --has-header option

clodius/cli/convert.py

@@ -74,7 +74,7 @@ def states_bedline_to_vector(bedlines,states_dic):
     chrom=parts[0]
     start=int(parts[1])
     end=int(parts[2])
-    state= states_dic[parts[3].encode('utf8')]
+    state= states_dic[parts[3]]
 
     states_vector = [ 1 if index == state else 0 for index in range(len(states_dic))]
 
@@ -135,13 +135,13 @@ def bedline_to_chrom_start_end_vector(bedlines, row_infos=None):
             start_set = set()
             end_set = set()
             all_vector = []
- 
+
             for bedline in bedlines:
                 parts = bedline.strip().split()
                 chrom = parts[chrom_col-1]
                 start = int(parts[from_pos_col-1])
                 end = int(parts[to_pos_col-1])
-                vector = [float(f) if not f == 'NA' else np.nan 
+                vector = [float(f) if not f == 'NA' else np.nan
                         for f in parts[value_col-1:value_col-1+num_rows]]
                 chrom_set.add(chrom)
                 start_set.add(start)
@@ -165,7 +165,8 @@ def bedline_to_chrom_start_end_vector(bedlines, row_infos=None):
                     starting_resolution, has_header, chunk_size)
         elif format == 'states':
             assert(row_infos != None), "A row_infos file must be provided for --format = 'states' "
-            states_dic = {row_infos[x]:x for x in range(len(row_infos))}
+            states_names = [lne.decode('utf8').split('\t')[0] for lne in row_infos]
+            states_dic = {states_names[x]:x for x in range(len(row_infos))}
 
             cmv.bedfile_to_multivec(filepaths, f_out, states_bedline_to_vector,
                     starting_resolution, has_header, chunk_size, states_dic)

clodius/multivec.py

@@ -14,6 +14,7 @@ def bedfile_to_multivec(input_filenames, f_out,
     '''
     Convert an epilogos bedfile to multivec format.
     '''
+
     files = []
     for input_filename in input_filenames:
         if op.splitext(input_filename)[1] == '.gz':
@@ -40,6 +41,10 @@ def bedfile_to_multivec(input_filenames, f_out,
     warned = False
 
     for lines in zip(*files):
+        # Identifies bedfile headers and ignore them
+        if "browser" == lines[0].decode('utf8')[0:7] or "track" in lines[0].decode('utf8')[0:6]:
+            continue
+
         chrom, start, end, vector = bedline_to_chrom_start_end_vector(lines, row_infos)
         # if vector[0] > 0 or vector[1] > 0:
         #     print("c,s,e,v", chrom, start, end, vector)

test/multivec_test.py

@@ -63,3 +63,22 @@ def test_states_format_befile_to_multivec():
     print("result.error", traceback.print_tb(tb))
     print("Exception:", a,b)
     '''
+
+
+def test_ignore_bedfile_headers():
+    runner = clt.CliRunner()
+    input_file = op.join(testdir, 'sample_data', '3_header_100_testfile.bed.gz')
+    rows_info_file = op.join(testdir, 'sample_data', '3_header_100_row_infos.txt')
+    f = tempfile.NamedTemporaryFile(delete=False)
+
+    result = runner.invoke(
+            ccc.bedfile_to_multivec,
+            [input_file,
+                '--format', 'states',
+                '--row-infos-filename', rows_info_file,
+                '--assembly', 'hg19',
+                '--starting-resolution', '200',
+                '--num-rows', '15'])
+
+    import traceback
+    a,b,tb = result.exc_info

test/sample_data/3_header_100_row_infos.txt

@@ -0,0 +1,15 @@
+15_Quies
+1_TssA
+14_ReprPCWk
+9_Het
+7_Enh
+5_TxWk
+2_TssAFlnk
+10_TssBiv
+4_Tx
+13_ReprPC
+12_EnhBiv
+11_BivFlnk
+8_ZNF/Rpts
+6_EnhG
+3_TxFlnk