add comments for athina

DessimozLab · Sep 25, 2024 · 349bdc9 · 349bdc9
1 parent 759ddcf
commit 349bdc9
Show file tree

Hide file tree

Showing 3 changed files with 11 additions and 4 deletions.
diff --git a/src/HogProf/lshbuilder.py b/src/HogProf/lshbuilder.py
@@ -161,10 +161,10 @@ def __init__(self,h5_oma=None,fileglob = None, taxa=None,masterTree=None, saving
 
         if self.h5OMA:
             self.HAM_PIPELINE = functools.partial( pyhamutils.get_ham_treemap_from_row, tree=self.tree_string ,  swap_ids=self.swap2taxcode , reformat_names = self.reformat_names , 
-                                                  orthoXML_as_string = True , use_phyloxml = self.use_phyloxml , orthomapper = self.idmapper ) 
+                                                  orthoXML_as_string = True , use_phyloxml = self.use_phyloxml , orthomapper = self.idmapper , levels = None ) 
         else:
             self.HAM_PIPELINE = functools.partial( pyhamutils.get_ham_treemap_from_row, tree=self.tree_string ,  swap_ids=self.swap2taxcode  , 
-                                                  orthoXML_as_string = False , reformat_names = self.reformat_names , use_phyloxml = self.use_phyloxml , orthomapper = self.idmapper )         
+                                                  orthoXML_as_string = False , reformat_names = self.reformat_names , use_phyloxml = self.use_phyloxml , orthomapper = self.idmapper , levels = None )         
 
         self.HASH_PIPELINE = functools.partial( hashutils.row2hash , taxaIndex=self.taxaIndex, treeweights=self.treeweights, wmg=wmg , lossonly = lossonly, duplonly = duplonly)
         if self.h5OMA:
@@ -254,6 +254,10 @@ def worker(self, i, q, retq, matq, l):
             df = q.get()
             if df is not None :
                 df['tree'] = df[['Fam', 'ortho']].apply(self.HAM_PIPELINE, axis=1)
+                #add a dictionary of results with subhogs { fam_sub1: { 'tree':tp , 'Fam':fam }  , fam_sub2: { 'tree':tp , 'Fam':fam } , ... }
+                #returned_df = pd.DataFrame.from_dict(df['tree'].to_dict(), orient='index')
+                #merge with pandas on right e.g. df.merge( returned_df , on = 'Fam' , how = 'right' )
+
                 df[['hash','rows']] = df[['Fam', 'tree']].apply(self.HASH_PIPELINE, axis=1)
                 if self.fileglob:
                     retq.put(df[['Fam', 'hash', 'ortho']])

diff --git a/src/HogProf/utils/hashutils.py b/src/HogProf/utils/hashutils.py
@@ -133,6 +133,7 @@ def row2hash(row , taxaIndex , treeweights , wmg , lossonly = False , duplonly =
     #convert a dataframe row to a weighted minhash
     fam, treemap = row.tolist()
     hog_matrix,weighted_hash = hash_tree(treemap , taxaIndex , treeweights , wmg , lossonly = lossonly , duplonly = duplonly)
+
     return  pd.Series([weighted_hash,hog_matrix], index=['hash','rows'])
 
 def fam2hash_hdf5(fam,  hdf5, dataset = None, nsamples = 128  ):

diff --git a/src/HogProf/utils/pyhamutils.py b/src/HogProf/utils/pyhamutils.py
@@ -84,7 +84,7 @@ def orthoxml2numerical(orthoxml , mapper):
     orthoxml = ET.tostring(root, encoding='unicode', method='xml')
     return orthoxml 
 
-def get_ham_treemap_from_row(row, tree , level = None , swap_ids = True , orthoXML_as_string = True , use_phyloxml = False , use_internal_name = True ,reformat_names= False, orthomapper = None ):  
+def get_ham_treemap_from_row(row, tree , levels = None , swap_ids = True , orthoXML_as_string = True , use_phyloxml = False , use_internal_name = True ,reformat_names= False, orthomapper = None ):  
     fam, orthoxml = row
     format = 'newick_string'
     if use_phyloxml:
@@ -99,8 +99,10 @@ def get_ham_treemap_from_row(row, tree , level = None , swap_ids = True , orthoX
         else:
             quoted = True
         try:
+            # return multiple treemaps corresponding to slices at different levels
             ham_obj = pyham.Ham(tree, orthoxml, type_hog_file="orthoxml" , tree_format = format  , use_internal_name=use_internal_name, orthoXML_as_string=orthoXML_as_string )            
-            tp = ham_obj.create_tree_profile(hog=ham_obj.get_list_top_level_hogs()[0])        
+            tp = ham_obj.create_tree_profile(hog=ham_obj.get_list_top_level_hogs()[0]) 
+            #check for losses / events and n leaves 
             return tp.treemap
         except Exception as e:
             # Capture the exception and format the traceback