diff --git a/candidate_filter/candidate_filter.py b/candidate_filter/candidate_filter.py
index 3f912dd..77e0d6a 100755
--- a/candidate_filter/candidate_filter.py
+++ b/candidate_filter/candidate_filter.py
@@ -56,8 +56,13 @@ def main(args):
 
     # Read files into a single pandas DataFrame
     with open(args.input, "r") as f:
-        xml_list = [i.strip() for i in f.readlines() if i]
-    df_cands_ini, obs_meta_data = reading_cands.read_candidate_files(xml_list)
+        candidate_files_list = [i.strip() for i in f.readlines() if i]
+    
+    if candidate_files_list[0].endswith('.xml'):
+        df_cands_ini, obs_meta_data = reading_cands.read_xml_candidate_files(candidate_files_list)
+    
+    if candidate_files_list[0].endswith('.csv'):
+        df_cands_ini, obs_meta_data = reading_cands.read_csv_candidate_files(candidate_files_list)
 
 
    
diff --git a/candidate_filter/reading_cands.py b/candidate_filter/reading_cands.py
index 9af2ca1..7f69385 100644
--- a/candidate_filter/reading_cands.py
+++ b/candidate_filter/reading_cands.py
@@ -3,9 +3,10 @@
 import xml.etree.ElementTree as ET
 from astropy import units as u
 from astropy.coordinates import SkyCoord
+import numpy as np
 
 
-def read_candidate_files(files, verbose=True):
+def read_xml_candidate_files(files, verbose=True):
     # Reads candidates files and include the candidates in a single pandas DataFrame
 
     #files = glob.glob(path + '*/overview.xml')
@@ -57,6 +58,79 @@ def read_candidate_files(files, verbose=True):
     return df_candidates, obs_meta_data
 
 
+def read_csv_candidate_files(files, verbose=True):
+    # Reads candidates files and include the candidates in a single pandas DataFrame
+
+    #files = glob.glob(path + '*/candidates.csv')
+
+    if verbose:
+        print(f"{len(files)} candidates files found.")
+    
+    all_rows = []
+    file_index = 0
+    for file in files:
+        file = file.replace(',','') 
+    
+        candidates = np.genfromtxt(file,dtype='str',skip_header=1,  delimiter=',') #header is column names for candidate info
+        
+        for candidate_number, candidate  in reversed(list(enumerate(candidates))): #start with last row of candidate file, which is observation metadata, to get the beam coordinates for all candidates in that file
+        
+            if candidate_number == np.size(candidates,0) - 1: #read observation metadata from the last row of the candidate file
+                
+                tsamp = float(candidate[3])
+                fft_size = 0.0 #this is handled correctly in candidate filter
+                obs_length = float(candidate[2])
+                nsamples = int(obs_length/tsamp)
+                speed_of_light = 299792458.0
+                obs_length_over_c = obs_length / speed_of_light
+                obs_meta_data = {"tsamp": tsamp,
+                                 "nsamples": nsamples,
+                                 "obs_length": obs_length,
+                                 "fft_size": fft_size,           
+                                 'obs_length_over_c': obs_length_over_c}
+                src_raj = float(candidate[0])
+                src_dej = float(candidate[1])
+                src_rajd, src_dejd = convert_to_deg(src_raj, src_dej)
+                
+            else:
+                   
+                row = []
+                new_dict = {}
+                new_dict['cand_id_in_file'] = candidate_number
+                new_dict['src_raj'] = src_raj #should be defined from last row of file
+                new_dict['src_rajd'] = src_rajd
+                new_dict['src_dej'] = src_dej
+                new_dict['src_dejd'] = src_dejd
+                new_dict['file_index'] = file_index
+                new_dict['period'] = float(candidate[0])
+                new_dict['dm'] = float(candidate[2])
+                new_dict['snr'] = float(candidate[5])
+                new_dict['acc'] = 0.0
+                new_dict['file'] = file
+                new_dict['nassoc'] = 3 #set as the default minimum required nassoc to disable low nassoc filtering for FFA outputs
+                #don't have nh, is_adjacent, is_physical, ddm count/snr ratio, byte offset like peasoup
+                row.append(new_dict)
+                all_rows.extend(row)
+                
+        file_index += 1
+
+
+    df_candidates = pd.DataFrame(all_rows)
+
+    # Additional type casting may be necessary or not necessary at all
+    df_candidates = df_candidates.astype({"snr": float, "dm": float, "period": float,
+                                          "acc": float, "nassoc": int})
+
+    if verbose:
+        print(f"{len(df_candidates)} candidates read.")
+
+    # sort by snr
+    df_candidates.sort_values('snr', inplace=True, ascending=False)
+    df_candidates.reset_index(inplace=True, drop=True)
+
+    return df_candidates, obs_meta_data
+
+
 def create_row(root, candidates, file, file_index):
     # Read a candidate file and creates data rows