Skip to content

Commit 03a1247

Browse files
committed
Final commit
0 parents  commit 03a1247

File tree

3 files changed

+108
-0
lines changed

3 files changed

+108
-0
lines changed

.gitignore

+1
Original file line numberDiff line numberDiff line change
@@ -0,0 +1 @@
1+
myenv/

mydbscheme_dump.sql

+55
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,55 @@
1+
-- MySQL dump 10.13 Distrib 5.7.20, for Win32 (AMD64)
2+
--
3+
-- Host: localhost Database: zauba
4+
-- ------------------------------------------------------
5+
-- Server version 5.7.20-log
6+
7+
/*!40101 SET @OLD_CHARACTER_SET_CLIENT=@@CHARACTER_SET_CLIENT */;
8+
/*!40101 SET @OLD_CHARACTER_SET_RESULTS=@@CHARACTER_SET_RESULTS */;
9+
/*!40101 SET @OLD_COLLATION_CONNECTION=@@COLLATION_CONNECTION */;
10+
/*!40101 SET NAMES utf8 */;
11+
/*!40103 SET @OLD_TIME_ZONE=@@TIME_ZONE */;
12+
/*!40103 SET TIME_ZONE='+00:00' */;
13+
/*!40014 SET @OLD_UNIQUE_CHECKS=@@UNIQUE_CHECKS, UNIQUE_CHECKS=0 */;
14+
/*!40014 SET @OLD_FOREIGN_KEY_CHECKS=@@FOREIGN_KEY_CHECKS, FOREIGN_KEY_CHECKS=0 */;
15+
/*!40101 SET @OLD_SQL_MODE=@@SQL_MODE, SQL_MODE='NO_AUTO_VALUE_ON_ZERO' */;
16+
/*!40111 SET @OLD_SQL_NOTES=@@SQL_NOTES, SQL_NOTES=0 */;
17+
18+
--
19+
-- Table structure for table `zauba`
20+
--
21+
22+
DROP TABLE IF EXISTS `zauba`;
23+
/*!40101 SET @saved_cs_client = @@character_set_client */;
24+
/*!40101 SET character_set_client = utf8 */;
25+
CREATE TABLE `zauba` (
26+
`srn` varchar(50) NOT NULL,
27+
`service_request_date` date DEFAULT NULL,
28+
`service_description` varchar(100) DEFAULT NULL,
29+
`t_o_f` varchar(50) DEFAULT NULL,
30+
`amount` float DEFAULT NULL,
31+
`total` float DEFAULT NULL,
32+
PRIMARY KEY (`srn`)
33+
) ENGINE=InnoDB DEFAULT CHARSET=utf8;
34+
/*!40101 SET character_set_client = @saved_cs_client */;
35+
36+
--
37+
-- Dumping data for table `zauba`
38+
--
39+
40+
LOCK TABLES `zauba` WRITE;
41+
/*!40000 ALTER TABLE `zauba` DISABLE KEYS */;
42+
INSERT INTO `zauba` VALUES ('U16571275','0000-00-00','U45200MH1995PTC094208','Normal',100,100),('U16572745','0000-00-00','U74999DL2007PTC163299','Normal',100,100),('U16573131','0000-00-00','U72200KA2006PTC039676','Normal',100,100);
43+
44+
UNLOCK TABLES;
45+
/*!40103 SET TIME_ZONE=@OLD_TIME_ZONE */;
46+
47+
/*!40101 SET SQL_MODE=@OLD_SQL_MODE */;
48+
/*!40014 SET FOREIGN_KEY_CHECKS=@OLD_FOREIGN_KEY_CHECKS */;
49+
/*!40014 SET UNIQUE_CHECKS=@OLD_UNIQUE_CHECKS */;
50+
/*!40101 SET CHARACTER_SET_CLIENT=@OLD_CHARACTER_SET_CLIENT */;
51+
/*!40101 SET CHARACTER_SET_RESULTS=@OLD_CHARACTER_SET_RESULTS */;
52+
/*!40101 SET COLLATION_CONNECTION=@OLD_COLLATION_CONNECTION */;
53+
/*!40111 SET SQL_NOTES=@OLD_SQL_NOTES */;
54+
55+
-- Dump completed on 2018-01-03 2:27:00

parse.py

+52
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,52 @@
1+
from io import StringIO
2+
from pdfminer.pdfinterp import PDFResourceManager, PDFPageInterpreter
3+
from pdfminer.converter import TextConverter
4+
from pdfminer.layout import LAParams
5+
from pdfminer.pdfpage import PDFPage
6+
import re
7+
import MySQLdb
8+
9+
def convert(fname, pages=None):
10+
if not pages:
11+
pagenums = set()
12+
else:
13+
pagenums = set(pages)
14+
15+
output = StringIO()
16+
manager = PDFResourceManager()
17+
converter = TextConverter(manager, output, laparams=LAParams())
18+
interpreter = PDFPageInterpreter(manager, converter)
19+
20+
infile = open(fname, 'rb')
21+
for page in PDFPage.get_pages(infile, pagenums):
22+
interpreter.process_page(page)
23+
infile.close()
24+
converter.close()
25+
text = output.getvalue()
26+
output.close
27+
return text
28+
29+
s=convert('U16571275.pdf', pages=[0,1])
30+
#print(s)
31+
32+
33+
srn=re.findall(r'[A-Z]\d+\d+\d+\d+\d+\d+\d', str(s))
34+
sr_date=re.findall(r'(.\d\W.\d\W.\d..)', str(s))
35+
#k=re.findall(r't=(.*)\s+a:.*\s+(\d+)\s+.*=(.*)',str(s))
36+
serv_desc=re.findall(r'([A-Z]\d.*\d)+\s+\W', str(s))
37+
tof=re.findall(r'\W.*\W+\n+(.*?)+\n\w', str(s))
38+
amt=re.findiall(r'(\d+\.\d*)', str(s))
39+
40+
41+
42+
db = MySQLdb.connect("localhost","usermm","root","zauba" )
43+
cursor = db.cursor()
44+
# Create table as per requirement
45+
sql = " insert into zauba values("+ .join(srn) +","+.join(sr_date)+","+serv_desc+","+ .join(tof)+","+.join(amt)+");"
46+
47+
cursor.execute(sql)
48+
49+
50+
db.close()
51+
52+

0 commit comments

Comments
 (0)