From f77a2a83d4fd85cd44faa048db8eba19d1a21522 Mon Sep 17 00:00:00 2001
From: johnandrea <46268748+johnandrea@users.noreply.github.com>
Date: Tue, 21 Dec 2021 14:48:44 -0400
Subject: [PATCH] Add tests for parameters

---
 diff.py | 949 +++++++++++++++++++++++++++++---------------------------
 1 file changed, 489 insertions(+), 460 deletions(-)

diff --git a/diff.py b/diff.py
index 6978991..5768bc9 100644
--- a/diff.py
+++ b/diff.py
@@ -1,460 +1,489 @@
-#!/usr/bin/python3
-
-"""
-Compare two gedcom trees.
-Arguments: tree1file  person1xref   tree2file  person2xref
-
-A person (child,partnermparent) which gets added in tree2 is not deteched,
-in order to do that run the program again reversing the order of the trees.
-
-This code is released under the MIT License: https://opensource.org/licenses/MIT
-Copyright (c) 2021 John A. Andrea
-v0.0.7
-"""
-
-import sys
-import re
-import difflib
-import readgedcom
-
-show_debug = False
-
-# how much change for a structure/branch difference
-branch_name_threshold = 0.88
-branch_date_threshold = 750 #days
-
-# how much change to report a person details difference
-report_name_threshold = 0.92
-report_date_threshold = 400  #days
-
-
-def check_config( start_ok ):
-    ok = start_ok
-
-    def check_val( start_ok, wanted_type, x_name, x ):
-        ok = start_ok
-        if isinstance( x, wanted_type ):
-           if x < 0:
-              print( x_name, 'cannot be less than zero', file=sys.stderr )
-              ok = False
-        else:
-           print( x_name, 'must be a', wanted_type, file=sys.stderr )
-           ok = False
-        return ok
-
-    ok = check_val( ok, float, 'branch_name_threshold', branch_name_threshold )
-    ok = check_val( ok, float, 'report_name_threshold', report_name_threshold )
-    ok = check_val( ok, int,   'branch_date_threshold', branch_date_threshold )
-    ok = check_val( ok, int,   'report_date_threshold', report_date_threshold )
-
-    return ok
-
-
-def days_between( d1, d2 ):
-    # expecting two dates as strings yyyymmdd
-    # return the approximate number of days between
-    # this is only for approximate comparisons, not date manipulations
-    # Assuming dates are C.E. Gregorian
-
-    def extract_parts( date ):
-        # yyyymmdd
-        # 01234567
-        return [ date[0:4], date[4:6], date[6:8] ]
-
-    def total_days( date ):
-        # [0]=year [1]=month [2]=day
-        return int(date[0]) * 365 + int(date[1]) * 30 + int(date[2])
-
-    parts = []
-    for date in [d1, d2]:
-        parts.append(extract_parts( date ))
-    return abs( total_days(parts[0]) - total_days(parts[1]) )
-
-
-def input_to_id( s ):
-    # might have been given as just a number
-    # convert to id style 'i58'
-    result = ''
-    if s:
-       result = s.replace('@','').replace('i','').replace('I','')
-    return 'i' + result
-
-
-def get_name( t, p ):
-    result = trees[t][ikey][p]['name'][0]['value']
-    if readgedcom.UNKNOWN_NAME in result:
-       result = 'unknown'
-    else:
-       # remove any suffix after the end slash
-       result = re.sub( r'/[^/]*$', '', result ).replace('/','').strip()
-    return result
-
-
-def get_best_date( t, p, date_name ):
-    best = None
-    if date_name in trees[t][ikey][p]['best-events']:
-       best = trees[t][ikey][p]['best-events'][date_name]
-    return best
-
-
-def get_full_date( t, p, date_name ):
-    # return the yyyymmdd value for the person's dated event, or None
-    result = None
-    best = get_best_date( t, p, date_name )
-    if best is not None:
-       if trees[t][ikey][p][date_name][best]['date']['is_known']:
-          # get the minimum date if its a range. if not a range min and max are equal
-          result = trees[t][ikey][p][date_name][best]['date']['min']['value']
-    return result
-
-
-def get_a_date( t, p, date_name ):
-    # return the date as given in the input file (or an empty string),
-    # which might contain before/after/etc or be a range
-    # not appropriate for comparison
-    result = ''
-    best = get_best_date( t, p, date_name )
-    if best is not None:
-       if trees[t][ikey][p][date_name][best]['date']['is_known']:
-          result = trees[t][ikey][p][date_name][best]['date']['in']
-    return result
-
-
-def get_dates( t, p ):
-    # for display purposes, not for comparison
-    return [ get_a_date(t, p, 'birt'), get_a_date(t, p, 'deat') ]
-
-
-def show_indi( t, p ):
-    # show an person's important details
-    dates = get_dates(t,p)
-    return get_name(t,p) + ' (' + dates[0] + '-' + dates[1] + ')'
-
-
-def get_other_partner( t, p, f ):
-    # in any given family, return the partner of the given person
-    result = None
-    for partner in ['wife','husb']:
-        if partner in trees[t][fkey][f]:
-           partner_id = trees[t][fkey][f][partner][0]
-           if partner_id != p:
-              result = partner_id
-    return result
-
-
-def list_all_partners( t, p ):
-    # for the given person in all the families in which they are a partner
-    # return   [family-id] = the-other-partner-id
-    # though partnerid might be unknown i.e. None
-    result = dict()
-    if 'fams' in trees[t][ikey][p]:
-       for fam in trees[t][ikey][p]['fams']:
-           result[fam] = get_other_partner( t, p, fam )
-    return result
-
-
-def show_person_header( t, p ):
-    print( '' )
-    print( show_indi( t, p ) )
-
-
-def get_name_match_value( n1, n2 ):
-    return difflib.SequenceMatcher(None, n1, n2).ratio()
-
-
-def compare_a_person( p1, p2 ):
-
-    def compare_person_dates( p1, title, date1, date2 ):
-        if date1:
-           if date2:
-              if days_between( date1, date2 ) >= report_date_threshold:
-                 show_person_header(1,p1)
-                 print( title, 'difference', date1, ' vs ', date2 )
-           else:
-              show_person_header(1,p1)
-              print( title, 'not in second tree' )
-        else:
-           if date2:
-              show_person_header(1,p1)
-              print( title, 'not in first tree' )
-
-    name1 = get_name( 1, p1 )
-    name2 = get_name( 2, p2 )
-
-    if get_name_match_value(name1,name2) < report_name_threshold:
-       show_person_header(1,p1)
-       print( 'Name difference:', name1, ' vs ', name2 )
-
-    for d in ['birt','deat']:
-        # first check whats given in the input file
-        d1 = get_a_date( 1, p1, d )
-        d2 = get_a_date( 2, p2, d )
-        if d1 != d2:
-           # then look closer at the parsed values
-           d1 = get_full_date( 1, p1, d )
-           d2 = get_full_date( 2, p2, d )
-           compare_person_dates(p1, d, d1, d2 )
-
-
-def person_match_value( t1, p1, t2, p2 ):
-    # should also check dates
-    name1 = get_name( t1, p1 )
-    name2 = get_name( t2, p2 )
-    value = get_name_match_value( name1, name2 )
-    if show_debug:
-       print( 'debug:sameness ', value, name1, ' vs ', name2 )
-    return value
-
-
-def is_same_person( t1, p1, t2, p2 ):
-    return person_match_value( t1, p1, t2, p2 ) >= branch_name_threshold
-
-
-def follow_parents( p1, p2 ):
-    if show_debug:
-       print( 'debug:follow parents of', get_name(1,p1) )
-
-    def get_famc( t, p ):
-        result = None
-        if 'famc' in trees[t][ikey][p]:
-           result = trees[t][ikey][p]['famc'][0]
-        return result
-
-    def get_partner( t, f, partner ):
-        result = None
-        if partner in trees[t][fkey][f]:
-           result = trees[t][fkey][f][partner][0]
-        return result
-
-    fam1 = get_famc( 1, p1 )
-    fam2 = get_famc( 2, p2 )
-
-    if fam1:
-       if fam2:
-          # this is going the be trouble for same sex couples
-          for partner in ['wife','husb']:
-              partner1 = get_partner( 1, fam1, partner )
-              partner2 = get_partner( 2, fam2, partner )
-
-              if partner1:
-                 if partner2:
-                    if is_same_person( 1, partner1, 2, partner2 ):
-                       # now what, check details
-                       if show_debug:
-                          print( 'debug:matched parent', partner, get_name(1,partner1) )
-                       follow_person( partner1, partner2 )
-                    else:
-                       show_person_header( 1, p1 )
-                       print( 'Parent(', partner, ') different from first to second' )
-                 else:
-                    show_person_header( 1, p1 )
-                    print( 'Parent(', partner, ') removed in second' )
-              else:
-                 if partner2:
-                    show_person_header( 1, p1 )
-                    print( 'Parent(', partner, ') added in second' )
-       else:
-          show_person_header( 1, p1 )
-          print( 'Parent(s) removed in second' )
-
-    else:
-      if fam2:
-         show_person_header( 1, p1 )
-         print( 'Parent(s) added in second' )
-
-
-def max_in_matrix( m ):
-    # should be a pythonic way to do this
-    x = -1
-    for i in m:
-        for j in m[i]:
-            x = max( x, m[i][j] )
-    return x
-
-
-def follow_children( p1, partner1, f1, f2 ):
-    global visited_fam
-    if f1 in visited_fam:
-       return
-    visited_fam.append(f1)
-
-    def match_children( p1, partner_name, children1, children2 ):
-        # make a matrix of matches to try gettimg the closest pairings
-        match_values = dict()
-        for c1 in children1:
-            match_values[c1] = dict()
-            for c2 in children2:
-                match_values[c1][c2] = person_match_value( 1, c1, 2, c2 )
-
-        matched1 = dict()
-        matched2 = dict()
-
-        best = max_in_matrix( match_values )
-
-        while best >= branch_name_threshold:
-            # where did it occur
-            # there might be a pythonic way to do this
-            for c1 in children1:
-                if c1 not in matched1:
-                   for c2 in children2:
-                       if c2 not in matched2:
-                          if match_values[c1][c2] >= best:
-                             match_values[c1][c2] = -1
-                             matched1[c1] = c2
-                             matched2[c2] = c1
-            best = max_in_matrix( match_values )
-
-        for c1 in children1:
-            if c1 in matched1:
-               follow_person( c1, matched1[c1] )
-
-            else:
-               show_person_header( 1, p1 )
-               print( 'with', partner_name, 'didnt match child', get_name(1,c1), 'first to second' )
-
-    partner_name = get_name(1,partner1)
-
-    if show_debug:
-       print( 'debug:follow children', get_name(1,p1),' and ', partner_name )
-
-    children1 = trees[1][fkey][f1]['chil']
-    children2 = trees[2][fkey][f2]['chil']
-    if children1:
-       if children2:
-          match_children( p1, partner_name, children1, children2 )
-
-       else:
-         show_person_header( 1, p1 )
-         print( 'All children with',partner_name,'removed in second' )
-    else:
-       if children2:
-          show_person_header( 1, p1 )
-          print( 'All children with',partner_name,'added in second' )
-
-
-def follow_partners( p1, p2 ):
-    if show_debug:
-       print( 'debug:in follow partners', get_name(1,p1) )
-
-    def match_partners( p1, partners1, partners2 ):
-        # make a matrix of tree1 people to tree2 people
-        # in order to find the best matches
-
-        match_values = dict()
-
-        for fam1 in partners1:
-            partner1 = partners1[fam1]
-            match_values[fam1] = dict()
-            for fam2 in partners2:
-                partner2 = partners2[fam2]
-                match_values[fam1][fam2] = person_match_value( 1, partner1, 2, partner2 )
-
-        # find the best match for each,
-        # so long as it isn't a better match for someone else
-
-        matched1 = dict()
-        matched2 = dict()
-
-        best = max_in_matrix( match_values )
-
-        while best >= branch_name_threshold:
-            # where did it occur
-            # there might be a pythonic way to do this
-            for fam1 in partners1:
-                if fam1 not in matched1:
-                   for fam2 in partners2:
-                       if fam2 not in matched2:
-                          if match_values[fam1][fam2] >= best:
-                             match_values[fam1][fam2] = -1
-                             matched1[fam1] = fam2
-                             matched2[fam2] = fam1
-            best = max_in_matrix( match_values )
-
-        for fam1 in partners1:
-            partner1 = partners1[fam1]
-            if fam1 in matched1:
-               fam2 = matched1[fam1]
-               follow_person( partner1, partners2[fam2] )
-
-               # now that families are known, do children within the family
-               follow_children( p1, partner1, fam1, fam2 )
-
-            else:
-               show_person_header( 1, p1 )
-               print( 'Didnt match partner', get_name(1,partner1), 'first to second' )
-
-
-    # check all the partners that person 1 might share with person 2
-    partners1 = list_all_partners( 1, p1 )
-    partners2 = list_all_partners( 2, p2 )
-
-    if partners1:
-       if partners2:
-          match_partners( p1, partners1, partners2 )
-
-       else:
-          show_person_header( 1, p1 )
-          print( 'Partner(s) removed in second' )
-    else:
-       if partners2:
-          show_person_header( 1, p1 )
-          print( 'Partner(s) added in second' )
-
-
-
-def follow_person( p1, p2 ):
-    global visited
-    if p1 in visited:
-       return
-    visited.append( p1 )
-
-    if show_debug:
-       print( 'debug:following person', show_indi( 1, p1 ) )
-
-    compare_a_person( p1, p2 )
-
-    follow_parents( p1, p2 )
-    follow_partners( p1, p2 )
-
-
-
-ikey = readgedcom.PARSED_INDI
-fkey = readgedcom.PARSED_FAM
-
-# the tree data will be globals
-trees = []
-starts = []
-file_names = []
-
-# add an initial zero'th element so that the rest of the program uses 1 and 2
-trees.append(0)
-starts.append(0)
-file_names.append(0)
-
-# params 1,2 then 3,4
-for i in [1,3]:
-    file_names.append( sys.argv[i] )
-    trees.append( readgedcom.read_file( sys.argv[i] ) )
-    starts.append( input_to_id( sys.argv[i+1] ) )
-
-ok = True
-print( 'Starting points' )
-for i in [1,2]:
-    if starts[i] in trees[i][ikey]:
-       print( i, '=', show_indi( i, starts[i] ) )
-    else:
-       print( 'Given key', starts[i], 'not in tree', i, file_names[i], file=sys.stderr )
-       ok = False
-
-ok = check_config( ok )
-
-if not ok:
-   sys.exit(1)
-
-# match the trees
-
-# prevent double visitations of the same person
-visited = []
-visited_fam = []
-
-follow_person( starts[1], starts[2] )
+#!/usr/bin/python3
+
+"""
+Compare two gedcom trees.
+Arguments: tree1file  person1xref   tree2file  person2xref
+
+A person (child,partnermparent) which gets added in tree2 is not deteched,
+in order to do that run the program again reversing the order of the trees.
+
+This code is released under the MIT License: https://opensource.org/licenses/MIT
+Copyright (c) 2021 John A. Andrea
+v0.0.9
+"""
+
+import sys
+import os
+import re
+import difflib
+import readgedcom
+
+show_debug = False
+
+# how much change for a structure/branch difference
+branch_name_threshold = 0.88
+branch_date_threshold = 750 #days
+
+# how much change to report a person details difference
+report_name_threshold = 0.92
+report_date_threshold = 400  #days
+
+
+def check_config( start_ok ):
+    ok = start_ok
+
+    def check_val( start_ok, wanted_type, x_name, x ):
+        ok = start_ok
+        if isinstance( x, wanted_type ):
+           if x < 0:
+              print( x_name, 'cannot be less than zero', file=sys.stderr )
+              ok = False
+        else:
+           print( x_name, 'must be a', wanted_type, file=sys.stderr )
+           ok = False
+        return ok
+
+    ok = check_val( ok, float, 'branch_name_threshold', branch_name_threshold )
+    ok = check_val( ok, float, 'report_name_threshold', report_name_threshold )
+    ok = check_val( ok, int,   'branch_date_threshold', branch_date_threshold )
+    ok = check_val( ok, int,   'report_date_threshold', report_date_threshold )
+
+    return ok
+
+
+def days_between( d1, d2 ):
+    # expecting two dates as strings yyyymmdd
+    # return the approximate number of days between
+    # this is only for approximate comparisons, not date manipulations
+    # Assuming dates are C.E. Gregorian
+
+    def extract_parts( date ):
+        # yyyymmdd
+        # 01234567
+        return [ date[0:4], date[4:6], date[6:8] ]
+
+    def total_days( date ):
+        # [0]=year [1]=month [2]=day
+        return int(date[0]) * 365 + int(date[1]) * 30 + int(date[2])
+
+    parts = []
+    for date in [d1, d2]:
+        parts.append(extract_parts( date ))
+    return abs( total_days(parts[0]) - total_days(parts[1]) )
+
+
+def input_to_id( s ):
+    # might have been given as just a number
+    # convert to id style 'i58'
+    result = ''
+    if s:
+       result = s.replace('@','').replace('i','').replace('I','')
+    return 'i' + result
+
+
+def get_name( t, p ):
+    result = trees[t][ikey][p]['name'][0]['value']
+    if readgedcom.UNKNOWN_NAME in result:
+       result = 'unknown'
+    else:
+       # remove any suffix after the end slash
+       result = re.sub( r'/[^/]*$', '', result ).replace('/','').strip()
+    return result
+
+
+def get_best_date( t, p, date_name ):
+    best = None
+    if date_name in trees[t][ikey][p]['best-events']:
+       best = trees[t][ikey][p]['best-events'][date_name]
+    return best
+
+
+def get_full_date( t, p, date_name ):
+    # return the yyyymmdd value for the person's dated event, or None
+    result = None
+    best = get_best_date( t, p, date_name )
+    if best is not None:
+       if trees[t][ikey][p][date_name][best]['date']['is_known']:
+          # get the minimum date if its a range. if not a range min and max are equal
+          result = trees[t][ikey][p][date_name][best]['date']['min']['value']
+    return result
+
+
+def get_a_date( t, p, date_name ):
+    # return the date as given in the input file (or an empty string),
+    # which might contain before/after/etc or be a range
+    # not appropriate for comparison
+    result = ''
+    best = get_best_date( t, p, date_name )
+    if best is not None:
+       if trees[t][ikey][p][date_name][best]['date']['is_known']:
+          result = trees[t][ikey][p][date_name][best]['date']['in']
+    return result
+
+
+def get_dates( t, p ):
+    # for display purposes, not for comparison
+    return [ get_a_date(t, p, 'birt'), get_a_date(t, p, 'deat') ]
+
+
+def show_indi( t, p ):
+    # show an person's important details
+    dates = get_dates(t,p)
+    return get_name(t,p) + ' (' + dates[0] + '-' + dates[1] + ')'
+
+
+def get_other_partner( t, p, f ):
+    # in any given family, return the partner of the given person
+    result = None
+    for partner in ['wife','husb']:
+        if partner in trees[t][fkey][f]:
+           partner_id = trees[t][fkey][f][partner][0]
+           if partner_id != p:
+              result = partner_id
+    return result
+
+
+def list_all_partners( t, p ):
+    # for the given person in all the families in which they are a partner
+    # return   [family-id] = the-other-partner-id
+    # though partnerid might be unknown i.e. None
+    result = dict()
+    if 'fams' in trees[t][ikey][p]:
+       for fam in trees[t][ikey][p]['fams']:
+           result[fam] = get_other_partner( t, p, fam )
+    return result
+
+
+def show_person_header( t, p ):
+    print( '' )
+    print( show_indi( t, p ) )
+
+
+def get_name_match_value( n1, n2 ):
+    return difflib.SequenceMatcher(None, n1, n2).ratio()
+
+
+def compare_a_person( p1, p2 ):
+
+    def compare_person_dates( p1, title, date1, date2 ):
+        if date1:
+           if date2:
+              if days_between( date1, date2 ) >= report_date_threshold:
+                 show_person_header(1,p1)
+                 print( title, 'difference', date1, ' vs ', date2 )
+           else:
+              show_person_header(1,p1)
+              print( title, 'not in second tree' )
+        else:
+           if date2:
+              show_person_header(1,p1)
+              print( title, 'not in first tree' )
+
+    name1 = get_name( 1, p1 )
+    name2 = get_name( 2, p2 )
+
+    if get_name_match_value(name1,name2) < report_name_threshold:
+       show_person_header(1,p1)
+       print( 'Name difference:', name1, ' vs ', name2 )
+
+    for d in ['birt','deat']:
+        # first check whats given in the input file
+        d1 = get_a_date( 1, p1, d )
+        d2 = get_a_date( 2, p2, d )
+        if d1 != d2:
+           # then look closer at the parsed values
+           d1 = get_full_date( 1, p1, d )
+           d2 = get_full_date( 2, p2, d )
+           compare_person_dates(p1, d, d1, d2 )
+
+
+def person_match_value( t1, p1, t2, p2 ):
+    # should also check dates
+    name1 = get_name( t1, p1 )
+    name2 = get_name( t2, p2 )
+    value = get_name_match_value( name1, name2 )
+    if show_debug:
+       print( 'debug:sameness ', value, name1, ' vs ', name2 )
+    return value
+
+
+def is_same_person( t1, p1, t2, p2 ):
+    return person_match_value( t1, p1, t2, p2 ) >= branch_name_threshold
+
+
+def follow_parents( p1, p2 ):
+    if show_debug:
+       print( 'debug:follow parents of', get_name(1,p1) )
+
+    def get_famc( t, p ):
+        result = None
+        if 'famc' in trees[t][ikey][p]:
+           result = trees[t][ikey][p]['famc'][0]
+        return result
+
+    def get_partner( t, f, partner ):
+        result = None
+        if partner in trees[t][fkey][f]:
+           result = trees[t][fkey][f][partner][0]
+        return result
+
+    fam1 = get_famc( 1, p1 )
+    fam2 = get_famc( 2, p2 )
+
+    if fam1:
+       if fam2:
+          # this is going the be trouble for same sex couples
+          for partner in ['wife','husb']:
+              partner1 = get_partner( 1, fam1, partner )
+              partner2 = get_partner( 2, fam2, partner )
+
+              if partner1:
+                 if partner2:
+                    if is_same_person( 1, partner1, 2, partner2 ):
+                       # now what, check details
+                       if show_debug:
+                          print( 'debug:matched parent', partner, get_name(1,partner1) )
+                       follow_person( partner1, partner2 )
+                    else:
+                       show_person_header( 1, p1 )
+                       print( 'Parent(', partner, ') different from first to second' )
+                 else:
+                    show_person_header( 1, p1 )
+                    print( 'Parent(', partner, ') removed in second' )
+              else:
+                 if partner2:
+                    show_person_header( 1, p1 )
+                    print( 'Parent(', partner, ') added in second' )
+       else:
+          show_person_header( 1, p1 )
+          print( 'Parent(s) removed in second' )
+
+    else:
+      if fam2:
+         show_person_header( 1, p1 )
+         print( 'Parent(s) added in second' )
+
+
+def max_in_matrix( m ):
+    # should be a pythonic way to do this
+    x = -1
+    for i in m:
+        for j in m[i]:
+            x = max( x, m[i][j] )
+    return x
+
+
+def follow_children( p1, partner1, f1, f2 ):
+    global visited_fam
+    if f1 in visited_fam:
+       return
+    visited_fam.append(f1)
+
+    def match_children( p1, partner_name, children1, children2 ):
+        # make a matrix of matches to try gettimg the closest pairings
+        match_values = dict()
+        for c1 in children1:
+            match_values[c1] = dict()
+            for c2 in children2:
+                match_values[c1][c2] = person_match_value( 1, c1, 2, c2 )
+
+        matched1 = dict()
+        matched2 = dict()
+
+        best = max_in_matrix( match_values )
+
+        while best >= branch_name_threshold:
+            # where did it occur
+            # there might be a pythonic way to do this
+            for c1 in children1:
+                if c1 not in matched1:
+                   for c2 in children2:
+                       if c2 not in matched2:
+                          if match_values[c1][c2] >= best:
+                             match_values[c1][c2] = -1
+                             matched1[c1] = c2
+                             matched2[c2] = c1
+            best = max_in_matrix( match_values )
+
+        for c1 in children1:
+            if c1 in matched1:
+               follow_person( c1, matched1[c1] )
+
+            else:
+               show_person_header( 1, p1 )
+               print( 'with', partner_name, 'didnt match child', get_name(1,c1), 'first to second' )
+
+    partner_name = get_name(1,partner1)
+
+    if show_debug:
+       print( 'debug:follow children', get_name(1,p1),' and ', partner_name )
+
+    children1 = trees[1][fkey][f1]['chil']
+    children2 = trees[2][fkey][f2]['chil']
+    if children1:
+       if children2:
+          match_children( p1, partner_name, children1, children2 )
+
+       else:
+         show_person_header( 1, p1 )
+         print( 'All children with',partner_name,'removed in second' )
+    else:
+       if children2:
+          show_person_header( 1, p1 )
+          print( 'All children with',partner_name,'added in second' )
+
+
+def follow_partners( p1, p2 ):
+    if show_debug:
+       print( 'debug:in follow partners', get_name(1,p1) )
+
+    def match_partners( p1, partners1, partners2 ):
+        # make a matrix of tree1 people to tree2 people
+        # in order to find the best matches
+
+        match_values = dict()
+
+        for fam1 in partners1:
+            partner1 = partners1[fam1]
+            match_values[fam1] = dict()
+            for fam2 in partners2:
+                partner2 = partners2[fam2]
+                match_values[fam1][fam2] = person_match_value( 1, partner1, 2, partner2 )
+
+        # find the best match for each,
+        # so long as it isn't a better match for someone else
+
+        matched1 = dict()
+        matched2 = dict()
+
+        best = max_in_matrix( match_values )
+
+        while best >= branch_name_threshold:
+            # where did it occur
+            # there might be a pythonic way to do this
+            for fam1 in partners1:
+                if fam1 not in matched1:
+                   for fam2 in partners2:
+                       if fam2 not in matched2:
+                          if match_values[fam1][fam2] >= best:
+                             match_values[fam1][fam2] = -1
+                             matched1[fam1] = fam2
+                             matched2[fam2] = fam1
+            best = max_in_matrix( match_values )
+
+        for fam1 in partners1:
+            partner1 = partners1[fam1]
+            if fam1 in matched1:
+               fam2 = matched1[fam1]
+               follow_person( partner1, partners2[fam2] )
+
+               # now that families are known, do children within the family
+               follow_children( p1, partner1, fam1, fam2 )
+
+            else:
+               show_person_header( 1, p1 )
+               print( 'Didnt match partner', get_name(1,partner1), 'first to second' )
+
+
+    # check all the partners that person 1 might share with person 2
+    partners1 = list_all_partners( 1, p1 )
+    partners2 = list_all_partners( 2, p2 )
+
+    if partners1:
+       if partners2:
+          match_partners( p1, partners1, partners2 )
+
+       else:
+          show_person_header( 1, p1 )
+          print( 'Partner(s) removed in second' )
+    else:
+       if partners2:
+          show_person_header( 1, p1 )
+          print( 'Partner(s) added in second' )
+
+
+
+def follow_person( p1, p2 ):
+    global visited
+    if p1 in visited:
+       return
+    visited.append( p1 )
+
+    if show_debug:
+       print( 'debug:following person', show_indi( 1, p1 ) )
+
+    compare_a_person( p1, p2 )
+
+    follow_parents( p1, p2 )
+    follow_partners( p1, p2 )
+
+
+def show_usage():
+    print( '' )
+    print( 'Arguments: tree1file  person1xref   tree2file  person2xref' )
+    print( 'Output report to stdout' )
+
+
+
+if len(sys.argv) != 5:
+   show_usage()
+   sys.exit(1)
+
+ikey = readgedcom.PARSED_INDI
+fkey = readgedcom.PARSED_FAM
+
+# the tree data will be globals
+trees = []
+starts = []
+file_names = []
+
+# add an initial zero'th element so that the rest of the program uses 1 and 2
+trees.append(0)
+starts.append(0)
+file_names.append(0)
+
+# params 1,2 then 3,4
+for i in [1,3]:
+    file_names.append( sys.argv[i] )
+    starts.append( input_to_id( sys.argv[i+1] ) )
+
+ok = True
+for i in [1,2]:
+    if os.path.isfile( file_names[i] ):
+       trees.append( readgedcom.read_file( file_names[i] ) )
+
+    else:
+       print( 'Data file does not exist:', file_names[i], file=sys.stderr )
+       ok = False
+
+if file_names[1] == file_names[2]:
+   print( 'Identical files', file=sys.stderr )
+   ok = False
+
+if not ok:
+   sys.exit(1)
+
+print( 'Starting points' )
+for i in [1,2]:
+    if isinstance( trees[i], dict ) and ikey in trees[i]:
+       if starts[i] in trees[i][ikey]:
+          print( i, '=', show_indi( i, starts[i] ) )
+       else:
+          print( 'Given key', starts[i], 'not in tree', i, file_names[i], file=sys.stderr )
+          ok = False
+    else:
+       print( 'Tree', i, 'not fully parsed data', file=sys.stderr )
+       ok = False
+
+ok = check_config( ok )
+
+if not ok:
+   sys.exit(1)
+
+# match the trees
+
+# prevent double visitations of the same person
+visited = []
+visited_fam = []
+
+follow_person( starts[1], starts[2] )