From 106d42a4234b73125e8cc89b8d5c3b041ae079f4 Mon Sep 17 00:00:00 2001
From: blackrim <blackrim@gmail.com>
Date: Wed, 28 Feb 2024 14:02:03 -0500
Subject: [PATCH] more

---
 src/join_paftol_tax.py | 41 +++++++++++++++++++++++++++++++++++++----
 1 file changed, 37 insertions(+), 4 deletions(-)

diff --git a/src/join_paftol_tax.py b/src/join_paftol_tax.py
index 661c0ac..f4d9c15 100644
--- a/src/join_paftol_tax.py
+++ b/src/join_paftol_tax.py
@@ -3,17 +3,49 @@
 import tree_reader
 import tree_utils
 
+"""
+TODO
+- need to do the mrca using names that includes internal nodes
+- need to go back to the node that is the MRCA but includes the clade of unsampled 
+so if it is 
+(a,b,c),(d,e,f) and i have (((e,f),d),a) -> ((a,b,c),(d,(e,f)))
+
+"""
+
 def process_tax(t):
     return
 
+def intersect_taxa(n,t):
+    return len(set(n).intersection(t))
+
 def get_mrca_wnms(n,t):
     if len(n) == 1:
-        for i in t.leaves():
+        for i in t.iternodes():
             if i.label == n[0]:
                 return i
     else:
         return tree_utils.get_mrca_wnms(n,t)
 
+# n = names that we wanted to get mrca
+# nd = the mrca
+# paflvsnms = the paf tree lvs nms
+def walk_back_mrca(nd,paflvsnms):
+    rnd = nd
+    intn = intersect_taxa(rnd.lvsnms(),paflvsnms)
+    going = True
+    while going:
+        nintn = intersect_taxa(rnd.parent.lvsnms(),paflvsnms)
+        if nintn != intn:
+            break
+        else:
+            if rnd.parent == None:
+                break
+            rnd = rnd.parent
+    return rnd
+
+        
+
+
 if __name__ == "__main__":
     if len(sys.argv) != 3:
         print("python",sys.argv[0],"paf tax")
@@ -25,7 +57,7 @@ def get_mrca_wnms(n,t):
     for i in paf.leaves():
         i.data["original_name"] = i.label
         i.label = i.label.split("_")[-1]
-    for i in tax.leaves():
+    for i in tax.iternodes():
         i.data["original_name"] = i.label
         i.label = i.label.split("_")[0]
     count= 0
@@ -40,6 +72,7 @@ def get_mrca_wnms(n,t):
             k = get_mrca_wnms(j.lvsnms(),tax)
             if k == None:
                 continue
+            k = walk_back_mrca(k,paf.lvsnms())
             chds.append(k)
         if len(chds) == 1:
             continue
@@ -50,6 +83,6 @@ def get_mrca_wnms(n,t):
             n.add_child(j)
         p.add_child(n)
         count += 1
-        if count == 21:
-            break
+        #if count == 100:
+        #    break
     print(tax.get_newick_repr(False))