我正在寻找 Scrapy 中 SQLite 管道的一些示例代码。我知道没有内置的支持，但我确信它已经完成了。只有实际的代码可以帮助我，因为我只知道足够的 Python 和 Scrapy 来完成我非常有限的任务，并且需要代码作为起点。
我做了这样的事情：
#
# Author: Jay Vaughan
#
# Pipelines for processing items returned from a scrape.
# Dont forget to add pipeline to the ITEM_PIPELINES setting
# See: http://doc.scrapy.org/topics/item-pipeline.html
#
from scrapy import log
from pysqlite2 import dbapi2 as sqlite
# This pipeline takes the Item and stuffs it into scrapedata.db
class scrapeDatasqLitePipeline(object):
def __init__(self):
# Possible we should be doing this in spider_open instead, but okay
self.connection = sqlite.connect('./scrapedata.db')
self.cursor = self.connection.cursor()
self.cursor.execute('CREATE TABLE IF NOT EXISTS myscrapedata ' \
'(id INTEGER PRIMARY KEY, url VARCHAR(80), desc VARCHAR(80))')
# Take the item and put it in database - do not allow duplicates
def process_item(self, item, spider):
self.cursor.execute("select * from myscrapedata where url=?", item['url'])
result = self.cursor.fetchone()
if result:
log.msg("Item already in database: %s" % item, level=log.DEBUG)
else:
self.cursor.execute(
"insert into myscrapedata (url, desc) values (?, ?)",
(item['url'][0], item['desc'][0]))
self.connection.commit()
log.msg("Item stored : " % item, level=log.DEBUG)
return item
def handle_error(self, e):
log.err(e)
这是一个带有 sqlalchemy 的 sqlite 管道。 使用 sqlalchemy，您可以根据需要轻松更改数据库。
在
settings.py# settings.py
# ...
DATABASE = {
'drivername': 'sqlite',
# 'host': 'localhost',
# 'port': '5432',
# 'username': 'YOUR_USERNAME',
# 'password': 'YOUR_PASSWORD',
'database': 'books.sqlite'
}
然后在
pipelines.py# pipelines.py
import logging
from scrapy import signals
from sqlalchemy import Column, Integer, String, DateTime
from sqlalchemy import create_engine
from sqlalchemy.engine.url import URL
from sqlalchemy.ext.declarative import declarative_base
from sqlalchemy.orm import sessionmaker
from sqlalchemy.pool import NullPool
logger = logging.getLogger(__name__)
DeclarativeBase = declarative_base()
class Book(DeclarativeBase):
__tablename__ = "books"
id = Column(Integer, primary_key=True)
title = Column('title', String)
author = Column('author', String)
publisher = Column('publisher', String)
url = Column('url', String)
scrape_date = Column('scrape_date', DateTime)
def __repr__(self):
return "<Book({})>".format(self.url)
class SqlitePipeline(object):
def __init__(self, settings):
self.database = settings.get('DATABASE')
self.sessions = {}
@classmethod
def from_crawler(cls, crawler):
pipeline = cls(crawler.settings)
crawler.signals.connect(pipeline.spider_opened, signals.spider_opened)
crawler.signals.connect(pipeline.spider_closed, signals.spider_closed)
return pipeline
def create_engine(self):
engine = create_engine(URL(**self.database), poolclass=NullPool, connect_args = {'charset':'utf8'})
return engine
def create_tables(self, engine):
DeclarativeBase.metadata.create_all(engine, checkfirst=True)
def create_session(self, engine):
session = sessionmaker(bind=engine)()
return session
def spider_opened(self, spider):
engine = self.create_engine()
self.create_tables(engine)
session = self.create_session(engine)
self.sessions[spider] = session
def spider_closed(self, spider):
session = self.sessions.pop(spider)
session.close()
def process_item(self, item, spider):
session = self.sessions[spider]
book = Book(**item)
link_exists = session.query(Book).filter_by(url=item['url']).first() is not None
if link_exists:
logger.info('Item {} is in db'.format(book))
return item
try:
session.add(book)
session.commit()
logger.info('Item {} stored in db'.format(book))
except:
logger.info('Failed to add {} to db'.format(book))
session.rollback()
raise
return item
和
items.py#items.py
import scrapy
class BookItem(scrapy.Item):
title = scrapy.Field()
author = scrapy.Field()
publisher = scrapy.Field()
scrape_date = scrapy.Field()
您也可以考虑将
class Bookitems.py如果您对twisted的adbapi感到满意，您可以将此mysql管道作为起点：http://github.com/darkrho/scrapy-googledir-mysql/blob/master/googledir/pipelines.py
并在
__init__self.dbpool = adbapi.ConnectionPool("sqlite3", database="/path/sqlite.db")
以下内容应兼容：
# Python 3.6.4
# Scrapy 1.5.1
# SQLite 3.21.0+
# APSW 3.9.2.post1
如果你想使用APSW，只需按照评论中的说明替换sqlite即可。
如果您的
items.py...
class NewAdsItem(Item):
AdId = Field()
DateR = Field()
DateA = Field()
DateN = Field()
DateE = Field()
URL = Field() # AdURL
在
settings.pyITEM_PIPELINES = {
'adbot.pipelines.DbPipeline': 100,
}
SQLITE_FILE = 'mad.db'
SQLITE_TABLE = 'ads'
在
pipelines.pyimport os
import sqlite3 # pip install pysqlite3
#import apsw # pip install apsw
from scrapy import signals
from scrapy.conf import settings
from adbot.items import NewAdsItem # Get items from "items.py"
con = None
ikeys = None
class DbPipeline(object):
dbfile = settings.get('SQLITE_FILE') # './test.db'
dbtable = settings.get('SQLITE_TABLE')
def __init__(self):
self.setupDBCon()
self.createTables()
def setupDBCon(self):
#self.con = apsw.Connection(self.dbfile) # apsw
self.con = sqlite3.connect(self.dbfile) # sqlite3
self.cur = self.con.cursor()
def createTables(self):
self.dropDbTable()
self.createDbTable()
def dropDbTable(self):
print("Dropping old table: %s" % self.dbtable )
self.cur.execute("DROP TABLE IF EXISTS %s" % self.dbtable )
def closeDB(self):
self.con.close()
def __del__(self):
self.closeDB()
def createDbTable(self):
print("Creating new table: %s" % self.dbtable )
#------------------------------
# Construct the item strings:
#------------------------------
# NOTE: This does not work, because items.py class re-orders the items!
#self.ikeys = NewAdsItem.fields.keys()
#print("Keys in creatDbTable: \t%s" % ",".join(self.ikeys) )
#cols = " TEXT, ".join(self.ikeys) + " TEXT"
#print("cols: \t%s:" % cols, flush=True)
#------------------------------
cols = "AdId TEXT, DateR TEXT, DateA TEXT, DateN TEXT, DateE TEXT, URL TEXT"
# NOTE: Use "INSERT OR IGNORE", if you also use: "AdId TEXT NOT NULL UNIQUE"
sql = "CREATE TABLE IF NOT EXISTS %s (id INTEGER PRIMARY KEY NOT NULL, %s)" % (self.dbtable, cols )
#print (sql)
self.cur.execute(sql)
def process_item(self, item, spider):
self.storeInDb(item)
return item
def storeInDb(self, item):
# NOTE: Use "INSERT OR IGNORE", if you also use: "AdId TEXT NOT NULL UNIQUE"
# "INSERT INTO ads ( AdId, DateR, AdURL ) VALUES (?, ?, ?)"
sql = "INSERT INTO {0} ({1}) VALUES ({2})".format(self.dbtable, ','.join(item.keys()), ','.join(['?'] * len(item.keys())) )
# (item.get('AdId',''),item.get('DateR',''),item.get('AdURL',''), ...)
itkeys = ','.join(item.keys()) # item keys as a list
itvals = ','.join(item.values()) # item values as a list
ivals = tuple(item.values()) # item values as a tuple
#print (sql)
#print(" itkeys: \t%s" % itkeys, flush=True)
#print(" itvals: \t%s" % itvals, flush=True)
self.cur.execute(sql, ivals) # WARNING: Does NOT handle '[]'s ==> use: '' in spider
self.con.commit() # used in sqlite3 ONLY! (Remove for APSW)
然后您可以使用以下命令从命令行检查数据库：
echo "select * from ads;" | sqlite3 -csv -header mad.db
警告：
因为
items.py项目钥匙的排序方式存在差异 通过或导入您的物品类别获得时 直接通过item.keys()，你会发现 第一个案例根据出现顺序（在文件中）排序， 而在第二种情况下，它是按“字母顺序”排序的。这是非常 悲伤，因为当您尝试创建数据库时它会带来麻烦 在执行self.ikeys = NewAdsItem.fields.keys()process_item()之前动态表。