From f74b56ca209ec20736ca69cf70363ebc8dc0b8dc Mon Sep 17 00:00:00 2001
From: Andrew Page <andrew.page@quadram.ac.uk>
Date: Thu, 7 May 2020 14:50:32 +0100
Subject: [PATCH] extended species database details

---
 VERSION               |  2 +-
 scripts/socru_species |  7 ++++++-
 socru/Schemas.py      | 42 ++++++++++++++++++++++++++++++++++++++++++
 3 files changed, 49 insertions(+), 2 deletions(-)

diff --git a/VERSION b/VERSION
index 5859406..530cdd9 100644
--- a/VERSION
+++ b/VERSION
@@ -1 +1 @@
-2.2.3
+2.2.4
diff --git a/scripts/socru_species b/scripts/socru_species
index d08a90d..29bb555 100755
--- a/scripts/socru_species
+++ b/scripts/socru_species
@@ -17,10 +17,15 @@ parser = argparse.ArgumentParser(
 	description = 'List all available species',
 	usage = 'socru_species [options]',
     formatter_class=argparse.ArgumentDefaultsHelpFormatter)
+    
+parser.add_argument('--extended', '-e', action='store_true', help='Extended information about the species databases', default = False)
 parser.add_argument('--debug', action='store_true', help='Turn on debugging', default = False)
 parser.add_argument('--verbose', '-v', action='store_true', help='Turn on verbose output', default = False)
 parser.add_argument('--version', action='version', version=str(version))
 
 options = parser.parse_args()
 
-Schemas(options.verbose).print_all()
+if options.extended:
+    Schemas(options.verbose).print_extended()
+else:
+    Schemas(options.verbose).print_all()
diff --git a/socru/Schemas.py b/socru/Schemas.py
index bac42e6..286b476 100644
--- a/socru/Schemas.py
+++ b/socru/Schemas.py
@@ -2,6 +2,7 @@
 from os import listdir
 from os.path import isdir
 import pkg_resources
+import yaml
 
 class Schemas:
     def __init__(self, verbose):
@@ -14,6 +15,47 @@ def all_available(self):
     def print_all(self):
         for d in sorted(self.all_available()):
             print(d)
+            
+    def extended(self):
+        db_info = {}
+        for species in listdir(self.base_directory):
+            
+            db_dir = os.path.join(self.base_directory, species)
+            db_file = os.path.join(self.base_directory, species,'profile.txt.yml')
+
+            if not os.path.exists(db_file):
+                continue
+                
+            with open(db_file, 'r') as metadatafh:
+                try:
+                    metadata = yaml.load(metadatafh, yaml.SafeLoader)
+                    dnaA_fragment_number = 0
+                    dnaa_forward_orientation = False
+                    dif_fragment_number = 0
+                    reference_genome = ''
+                    
+                    if 'dnaa_fragment' in metadata:
+                        dnaA_fragment_number = int(metadata['dnaa_fragment'])
+                        dnaa_forward_orientation = metadata['dnaa_forward_orientation']
+					
+                    if 'dif_fragment' in metadata:
+                        dif_fragment_number = int(metadata['dif_fragment'])
+                        
+                    if 'reference_genome' in metadata:
+                        reference_genome = str(metadata['reference_genome'])
+                        
+                    db_info[species] = [species, dnaA_fragment_number, dnaa_forward_orientation, dif_fragment_number, reference_genome]
+                    
+					
+                except yaml.YAMLError as exc:
+                    print(exc)
+        return db_info
+        
+    def print_extended(self):
+        extended_details = self.extended()
+        print("\t".join(['Species','dnaA fragment No.', 'dnaA forward orientation', 'dif fragment No.', 'Reference']))
+        for species in sorted(extended_details):
+            print("\t".join( [ str(a) for a in extended_details[species] ]))
         
         
     def database_directory(self, db_dir, species):