[july] step6 ?

s5.py

@@ -7,7 +7,6 @@ from __future__ import print_function
 from base import create_csv, TARGET_DIR, MODEL_LIST
 import os
 import csv
-import sys
 
 
 def get_sorted_list(model_name):
@@ -53,7 +52,7 @@ for m in MODEL_LIST:
 rows.append(row)
 
 for i in xrange(0, max_count):
-    row = [i, ]
+    row = [(i+1), ]
     for m in MODEL_LIST:
         try:
             row += sorted_by_weight[m][i]

s6.py

@@ -0,0 +1,154 @@
+'''
+Step 6
+'''
+# -*- coding:utf-8 -*-
+#!/usr/bin/env python
+from __future__ import print_function
+from base import create_csv, TARGET_DIR, MODEL_LIST
+import os
+import csv
+
+m2case = {
+    'I': 'LS',
+    'P': 'LS',
+    'C': 'LS',
+    'M': 'Gender',
+    'F': 'Gender',
+    'H': 'Level',
+    'S': 'Level',
+    'D': 'SciF',
+    'f': 'SciF',
+}
+
+x2tiny_conv = [0, 0, 0, 4, 3, 4, 3, 5, 3, 3, 3, 0]
+merging_col = [
+    [19, 6],
+    [20, 7],
+    [21, 10],
+    [22, 3],
+    [23, 6],
+    [24, 11],
+    [25, 4],
+    [26, 12],
+    [27, 14],
+]
+
+
+def break_into_tiny_list(incoming, length):
+    if incoming:
+        return [incoming[i] for i in xrange(0, length)]
+    else:
+        return ['' for i in xrange(0, length)]
+
+
+def get_yynn(four_char_model):
+    g = {}
+    for m in four_char_model:
+        g[m] = {}
+        src_name = 'Case1_%s.csv-output.csv' % m2case[m]
+        src_path = os.path.join(TARGET_DIR, src_name)
+        with open(src_path, 'rb') as f:
+            rows = csv.reader(f)
+            for r in rows:
+                if m != r[0]:
+                    continue
+                for l in r[1].split(','):
+                    # g[m][l] = r[3:11]
+                    # break into pieces
+                    g[m][l] = []
+                    for c in xrange(3, 11):
+                        g[m][l] += break_into_tiny_list(r[c], x2tiny_conv[c])
+
+    return g
+
+
+def get_lo_yynn(lo_id, char_model, **kwargs):
+    '''
+    This doesn't do anything special, just try-catch and leave other def clean
+    '''
+    if not kwargs.get('src') or not lo_id or not char_model:
+        return ['' for i in xrange(0, 28)]
+    src = kwargs.get('src')
+    try:
+        return src[char_model][lo_id]
+    except KeyError:
+        return ['' for i in xrange(0, 28)]
+
+
+def looop_each_4char_model():
+    '''
+    Each model has 6 columns of data + # (1st col)
+    '''
+    loop_order = 0
+    offset = 6
+    for m in MODEL_LIST:
+        yynn = get_yynn(m)
+        data = []
+        fpath = os.path.join(TARGET_DIR, 'part5', 'sorted_all.csv')
+        with open(fpath, 'rb') as f:
+            rows = csv.reader(f)
+            rows.next()
+            for r in rows:
+                b = [r[0], ]
+                st = (offset*loop_order)+1
+                b += r[st: st+offset]
+                if not b[2]:
+                    continue
+                l = {'lo_id': b[2], 'children': []}
+                for sm in m:
+                    c = b[:]
+                    c.append(sm)
+                    c += get_lo_yynn(b[2], sm, src=yynn)
+                    l['children'].append(c)
+
+                l['merged_row'] = merge_row(l['children'])
+                l['merged_col'] = merge_col(l['merged_row'])
+                data.append(l)
+        loop_order += 1
+        write_output(m, data)
+
+
+def merge_row(children):
+    if len(children) < 1:
+        return []
+    llen = len(children[0])
+    merged = children[0][0:8] + ['' for i in xrange(8, llen)]
+    for i in children:
+        for j in xrange(8, llen):
+            if merged[j] == '':
+                merged[j] = i[j]
+            elif merged[j] != i[j]:
+                if i[j] != '':
+                    merged[j] = 'Y/N'
+    # get rid of 1-char-model name
+    merged = merged[0:7]+merged[8:]
+    return merged
+
+
+def merge_col(child):
+    if len(child) < 27:
+        return child
+    offset = 7
+    merged = child[0:19+offset]
+    for x, y in merging_col:
+        x += offset
+        y += offset
+        if merged[y] == '':
+            merged[y] = child[x]
+        elif merged[y] != child[x]:
+            if child[x] != '':
+                merged[y] = 'Y/N'
+    return merged
+
+
+def write_output(model_name, raw_data):
+    r = []
+    for i in raw_data:
+        row = [
+            'R_%s' % i['lo_id'][3:],
+        ] + i['merged_col']
+        r.append(row)
+    create_csv('%s-6.csv' % model_name, r, directory='part6')
+
+
+looop_each_4char_model()