Add Excel support and bump to 0.6.0

apcamargo · Jul 24, 2019 · 1227b3a · 1227b3a
1 parent dbdf92f
commit 1227b3a
Show file tree

Hide file tree

Showing 4 changed files with 13 additions and 8 deletions.
diff --git a/README.md b/README.md
@@ -54,7 +54,8 @@ usage: tspex [-h] [-l] [-d] [-t THRESHOLD] input_file output_file method
 Compute gene tissue-specificity from an expression matrix and save the output.
 
 positional arguments:
-  input_file            Expression matrix file in the TSV or CSV formats.
+  input_file            Expression matrix file in the TSV, CSV or Excel
+                        formats.
   output_file           Output TSV file containing tissue-specificity values.
   method                Tissue-specificity metric. Allowed values are:
                         "counts", "tau", "gini", "simpson",

diff --git a/docs/cli.md b/docs/cli.md
@@ -13,7 +13,8 @@ usage: tspex [-h] [-l] [-d] [-t THRESHOLD] input_file output_file method
 Compute gene tissue-specificity from an expression matrix and save the output.
 
 positional arguments:
-  input_file            Expression matrix file in the TSV or CSV formats.
+  input_file            Expression matrix file in the TSV, CSV or Excel
+                        formats.
   output_file           Output TSV file containing tissue-specificity values.
   method                Tissue-specificity metric. Allowed values are:
                         "counts", "tau", "gini", "simpson",

diff --git a/setup.py b/setup.py
@@ -25,15 +25,15 @@
 
 setup(
     name='tspex',
-    version='0.5.1',
+    version='0.6.0',
     packages=find_packages(),
     license='GNU General Public License v3.0',
     description='A Python package for calculating tissue-specificity metrics for gene expression.',
     long_description=open('README.md').read(),
     long_description_content_type='text/markdown',
-    install_requires=['matplotlib >= 2.2', 'numpy', 'pandas >= 0.23'],
-    python_requires= '>=3',
-    entry_points = {
+    install_requires=['matplotlib >= 2.2', 'numpy', 'pandas >= 0.23', 'xlrd >= 1.1.0'],
+    python_requires='>=3',
+    entry_points={
         'console_scripts': ['tspex=tspex.cli:main'],
     },
     url='https://apcamargo.github.io/tspex/',

diff --git a/tspex/cli.py b/tspex/cli.py
@@ -32,7 +32,10 @@
 def tspex_cli(input_file, output_file, method, log, disable_transformation, threshold):
     """Compute gene tissue-specificity from a expression matrix file and save an output file."""
     transform = not disable_transformation
-    expression_matrix = pd.read_csv(input_file, index_col=0, header=0, sep=None, engine='python')
+    if input_file.rsplit('.', 1)[1].lower() in ['xls', 'xlsx']:
+        expression_matrix = pd.read_excel(input_file, index_col=0, header=0)
+    else:
+        expression_matrix = pd.read_csv(input_file, index_col=0, header=0, sep=None, engine='python')
     tissue_specificity = tspex.TissueSpecificity(
         expression_matrix, method, log, transform=transform, threshold=threshold)
     tissue_specificity.tissue_specificity.to_csv(output_file, sep='\t')
@@ -45,7 +48,7 @@ def main():
         description='Compute gene tissue-specificity from an expression matrix and save the output.',
         formatter_class=argparse.ArgumentDefaultsHelpFormatter)
     parser.add_argument('input_file',
-                        help='Expression matrix file in the TSV or CSV formats.')
+                        help='Expression matrix file in the TSV, CSV or Excel formats.')
     parser.add_argument('output_file',
                         help='Output TSV file containing tissue-specificity values.')
     parser.add_argument(