[aug] s9

s8.py

@@ -80,11 +80,9 @@ def main():
             m_header = header[:]
             m_header[1] = m
             ros = get_ro_data(m, i)
-            for r in ros:
+            m_header += [r[0] for r in ros]
-                m_header.append(r[0])
             result = middleman(lo_data, ros)
             output_name = '%s-%s.csv' % (m, i)
-
             # process header -- get ind of first R_xxx
             number_ind = 0
             cnt_ind = 0
@@ -96,7 +94,7 @@ def main():
                     break
                 except ValueError:
                     cnt_ind += 1
-            __h = m_header[:2] + ['_' for i in xrange(0, number_ind-2)] + m_header[3:]
+            __h = m_header[:2] + ['_' for i in xrange(0, number_ind-2)] + m_header[2:]
             result = [__h] + result
             create_csv(output_name, result, directory='part8')
 

s9.py

@@ -0,0 +1,92 @@
+#!/usr/bin/env python
+# -*- coding:utf-8 -*-
+'''
+Step 9
+'''
+from __future__ import print_function
+from base import create_csv, TARGET_DIR, MODEL_LIST, RAW_DIR
+import os
+import csv
+import re
+
+FN_CONST = 19
+
+def get_retrieved_yynn_count(model):
+    result = {}
+    for i in '1234':
+        P = os.path.join(TARGET_DIR, 'part7', '%s-%s.csv' % (model, i))
+        with open(P, 'rb') as f:
+            rows = csv.reader(f)
+            for row in rows:
+                if row[0] not in result:
+                    result[row[0]] = row[len(row)-1]
+    return result
+
+
+def process_file(seven_data, data, header):
+    start_ind = 0
+    for i in xrange(0, len(header)):
+        if re.search(r'^R', header[i]):
+            start_ind = i
+            break
+
+    prf_output = []
+    for i in xrange(start_ind, len(data)):
+        R_id = header[i]
+        if R_id not in seven_data:
+            print("%s missing in step7" % R_id)
+            continue
+        TP = float(data[i])
+        FP = float(seven_data[R_id]) - TP
+        FN = float(FN_CONST) - TP
+        # print('TP: %s, FP: %s, FN: %s' % (TP, FP, FN))
+
+        P = (TP / (TP + FP)) * 100
+        R = (TP / (TP + FN)) * 100
+        try:
+            F = 2 * (P * R) / (P + R)
+        except ZeroDivisionError:
+            F = 'inf'
+            print('ZeroDivisionError\n    %s -- R_id: %s | P: %s | R: %s' % (data[0], R_id, P, R))
+
+        prf_output.append({
+            'R_id': R_id,
+            "P": P,
+            "R": R,
+            "F": F,
+        })
+    # print(prf_output)
+    return prf_output
+
+
+def loop_thru_step8():
+    BASE_PATH = os.path.join(TARGET_DIR, 'part8')
+    for i in os.walk(BASE_PATH):
+        for j in i[2]:
+            output = []
+            seven = get_retrieved_yynn_count(j[:4])
+            fpath = os.path.join(BASE_PATH, j)
+            # print(fpath)
+            with open(fpath, 'rb') as f:
+                output_header = ['lo_id', j[:4]]
+                output_sub_header = ['', '', ]
+                rows = csv.reader(f)
+                header = rows.next()
+                is_header_created = False
+                for r in rows:
+                    output_row = [r[0], j[:4]]
+                    prf = process_file(seven, r, header)
+                    for i in prf:
+                        if not is_header_created:
+                            output_header += [i['R_id'], '', '']
+                            output_sub_header += ['P', 'R', 'F']
+                        output_row += [i['P'], i['R'], i['F']]
+                    output.append(output_row)
+                    is_header_created = True
+
+            output_name = j
+            result = [output_header] + [output_sub_header] + output
+            create_csv(output_name, result, directory='part9')
+            is_header_created = False
+
+loop_thru_step8()