From f2f14cd090595d67c6abbc20e45ad3d907fd10d4 Mon Sep 17 00:00:00 2001
From: Javi Ribera <jprat@purdue.edu>
Date: Mon, 20 Nov 2017 23:33:58 -0500
Subject: [PATCH] script to generate CSV from Yuhao

---
 generate_csv.py | 91 +++++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 91 insertions(+)
 create mode 100644 generate_csv.py

diff --git a/generate_csv.py b/generate_csv.py
new file mode 100644
index 0000000..bb4cce0
--- /dev/null
+++ b/generate_csv.py
@@ -0,0 +1,91 @@
+import pandas as pd
+import cv2
+import numpy as np
+import sys
+import os
+import ast
+import random
+import shutil
+from tqdm import tqdm
+
+np.random.seed(0)
+
+train_df = pd.DataFrame(columns=['plant_count'])
+test_df = pd.DataFrame(columns=['plant_count'])
+validate_df = pd.DataFrame(columns=['plant_count'])
+
+if not os.path.exists('train'):
+    os.makedirs('train')
+if not os.path.exists('test'):
+    os.makedirs('test')
+if not os.path.exists('validate'):
+    os.makedirs('validate')
+
+dirs = [i for i in range(1, 18)]
+
+for dirnum in tqdm(dirs):
+    dirname = 'dataset' + str(dirnum).zfill(2)
+    
+    fd = open(os.path.join(dirname,'gt.txt'))
+
+    data = []
+    for line in fd:
+        line = line.strip()
+        imgnum = line.split(' ')[1]
+        x = line.split(' ')[2]
+        if (x == 'X'):
+            continue
+        y = line.split(' ')[3]
+
+        imagename = imgnum.zfill(10)+'.png'
+        if not os.path.exists(os.path.join(dirname,imagename)):
+            continue
+        image = cv2.imread(os.path.join(dirname,imagename))
+
+        h = image.shape[0]
+        x = int(x)/2
+        y = h - int(y)/2
+        data.append([imagename, y, x])
+
+        #print(imagename)
+        #print(x, y)
+
+    random.shuffle(data)
+    for i in range(len(data)):
+        item = data[i]
+        imagename = item[0]
+        y = item[1]
+        x = item[2]
+
+        # newname = str(filecounter).zfill(10) + '.png'
+        newname = dirname + '_' + imagename
+        df = pd.DataFrame(data=[[1, [[y, x]]]],
+        index=[newname],
+                          columns=['plant_count', 'plant_locations'])
+        if (i < len(data)*0.8):
+            if os.path.isfile('train/'+newname):
+                print('%s exists' % 'train/'+newname)
+                exit(-1)
+            shutil.move(os.path.join(dirname,imagename), 'train/'+newname)
+            train_df = train_df.append(df)
+        elif (i < len(data)*0.9):
+            if os.path.isfile('train/'+newname):
+                print('%s exists' % 'test/'+newname)
+                exit(-1)
+            shutil.move(os.path.join(dirname,imagename), 'test/'+newname)
+            test_df = test_df.append(df)
+        else:
+            if os.path.isfile('train/'+newname):
+                print('%s exists' % 'test/'+newname)
+                exit(-1)
+            shutil.move(os.path.join(dirname,imagename), 'validate/'+newname)
+            validate_df = validate_df.append(df)
+
+train_df.to_csv('train.csv')
+shutil.move('train.csv', 'train')
+test_df.to_csv('test.csv')
+shutil.move('test.csv', 'test')
+validate_df.to_csv('validate.csv')
+shutil.move('validate.csv', 'validate')
+
+