New build instructions

MANIFEST.in

@@ -0,0 +1,15 @@
+# By default, all files are excluded
+# Include all source
+graft datasail
+
+# Include the files for the build process
+include setup.py
+include recipe/meta.yaml
+include recipe_lite/meta.yaml
+include base_recipe.yaml
+
+# Include housekeeping files
+include README.md
+include LICENSE
+include CONTRIBUTING.md
+

base_recipe.yaml

@@ -4,6 +4,9 @@ package:
 
 source:
   path: ..
+#     - ../datasail/*
+#     - ../datasail/args/*.yaml
+#     - setup.py
 
 build:
   entry_points:
@@ -44,4 +47,4 @@ about:
   home: https://github.com/kalininalab/DataSAIL
   license: MIT
   license_file: LICENSE
-  summary: DataSAIL is a tool to split datasets while reducing information leakage.
+  summary: DataSAIL is a tool to split datasets while reducing information leakage.

setup.py

@@ -24,9 +24,10 @@
         "Natural Language :: English",
         "Topic :: Scientific/Engineering :: Bio-Informatics",
     ],
-    packages=find_packages(),
-    include_package_data=False,
-    package_data={"datasail": ["datasail/args/*"]},
+    packages=["datasail"],
+    include_package_data=True,
+    # packages=find_packages(),
+    # include_package_data=False,
     install_requires=[],
     python_requires=">=3.8, <4.0.0",
     keywords="bioinformatics",

tests/data/pipeline/prot_strat.tsv

@@ -0,0 +1,37 @@
+# 1 "prot_weights.tsv"
+# 1 "<built-in>"
+# 1 "<command-line>"
+# 31 "<command-line>"
+# 1 "/usr/include/stdc-predef.h" 1 3 4
+# 32 "<command-line>" 2
+# 1 "prot_weights.tsv"
+Target_ID weight
+5CO1_A 1
+4OLI_A 0
+3ZON_A 1
+3WOV_A 0
+5TKD_A 1
+7AX4_A 0
+6NZR_A 1
+4L00_A 0
+4L01_A 1
+6B3J_R 0
+7RTB_R 1
+7FIM_R 0
+7X8R_R 1
+7DUQ_R 0
+6VCB_R 1
+5NX2_A 0
+7S15_R 1
+6LN2_A 0
+5VAI_R 1
+1XO7_A 0
+4S1E_A 1
+2HAQ_A 0
+4S1J_A 1
+6L2B_A 0
+3BT8_A 1
+8CU5_A 0
+4FRV_A 1
+4FRU_A 0
+1CYN_A 1