From bb6b7005a5f624dab5ac781e50e30e3bc8ed855b Mon Sep 17 00:00:00 2001
From: asistradition <asistradition@users.noreply.github.com>
Date: Wed, 16 Feb 2022 15:22:04 -0500
Subject: [PATCH] Fix parsing error v0.3.2

---
 CHANGELOG.md                     |  4 ++++
 inferelator_prior/motifs/meme.py |  4 ++++
 inferelator_prior/motifs/pwm.py  | 13 +++++++------
 setup.py                         |  4 ++--
 4 files changed, 17 insertions(+), 8 deletions(-)

diff --git a/CHANGELOG.md b/CHANGELOG.md
index 34343d9..12cd0b4 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -1,3 +1,7 @@
+### Version 0.3.2
+
+* Corrected a parsing error when reading CisBP PWM files 
+
 ### Version 0.3.1
 
 * Added `link_atac_bed_to_genes` module to link specific peaks from a BED file to nearby genes 
diff --git a/inferelator_prior/motifs/meme.py b/inferelator_prior/motifs/meme.py
index 7a95a59..49a3e46 100644
--- a/inferelator_prior/motifs/meme.py
+++ b/inferelator_prior/motifs/meme.py
@@ -39,6 +39,10 @@ def read(file_descript):
 def write(file_descript, motifs, alphabet=None, background=None, mode="w"):
 
     motifs = [motifs] if not isinstance(motifs, (list, tuple, pd.Series)) else motifs
+
+    if len(motifs) == 0:
+        raise RuntimeError("Unable to write motif file with zero motifs")
+
     alphabet = alphabet if alphabet is not None else motifs[0].alphabet
     background = np.array([[1 / len(alphabet)] * len(alphabet)]) if background is None else background
 
diff --git a/inferelator_prior/motifs/pwm.py b/inferelator_prior/motifs/pwm.py
index 48300fd..3bc3d51 100644
--- a/inferelator_prior/motifs/pwm.py
+++ b/inferelator_prior/motifs/pwm.py
@@ -23,19 +23,20 @@ def read(pwm_file_list, info_file, background=None, direct_only=False, pwm_has_i
 
     for pwm_file in pwm_file_list:
         pwm_id = os.path.splitext(os.path.basename(pwm_file))[0]
+        match_id = info_df[MOTIF_COL] == pwm_id
 
-        if pwm_id not in info_df[MOTIF_COL]:
+        if (match_id).sum() == 0:
             pwm_not_present.append(pwm_id)
             continue
 
         if direct_only:
-            direct = info_df.loc[info_df[MOTIF_COL] == pwm_id, TF_STATUS_COL].str.contains("D")
+            direct = info_df.loc[match_id, TF_STATUS_COL].str.contains("D")
             if not direct.any():
                 continue
             else:
-                pwm_names = info_df.loc[(info_df[MOTIF_COL] == pwm_id) & (info_df[TF_STATUS_COL] == "D"), TF_NAME_COL]
+                pwm_names = info_df.loc[match_id & (info_df[TF_STATUS_COL] == "D"), TF_NAME_COL]
         else:
-            pwm_names = info_df.loc[info_df[MOTIF_COL] == pwm_id, TF_NAME_COL]
+            pwm_names = info_df.loc[match_id, TF_NAME_COL]
 
         pwm_name = "/".join(pwm_names)
 
@@ -63,8 +64,8 @@ def read(pwm_file_list, info_file, background=None, direct_only=False, pwm_has_i
         motifs.append(motif)
 
     if len(pwm_not_present) > 0:
-        print("{pwm} PWM files not found in in {c} of {cf}".format(pwm=len(pwm_not_present), c=MOTIF_COL, cf=info_file))
+        print(f"{len(pwm_not_present)} PWM files not found in in {MOTIF_COL} of {info_file}")
     if len(pwm_malformed) > 0:
-        print("{pwm} PWM files malformed and improperly parsed".format(pwm=len(pwm_malformed)))
+        print(f"{len(pwm_malformed)} PWM files malformed and improperly parsed")
 
     return motifs
diff --git a/setup.py b/setup.py
index 43d09fd..fa5f01e 100644
--- a/setup.py
+++ b/setup.py
@@ -1,9 +1,9 @@
 import os
 from setuptools import setup, find_packages
 
-install_requires = ["numpy", "pandas>=1.0", "HTSeq", "pybedtools", "scipy", "pathos", "sklearn"]
+install_requires = ["numpy", "pandas>=1.0", "HTSeq", "pybedtools", "scipy", "pathos", "sklearn", "tqdm"]
 tests_require = ["coverage", "nose", "pysam"]
-version = "0.3.1"
+version = "0.3.2"
 
 # Description from README.md
 base_dir = os.path.dirname(os.path.abspath(__file__))