scrapy源码解读命令行工具

Posted 2021-02-28 tianyiliang

一、前言

scrapy的命令分为全局命令和项目命令。顾名思义，全局命令是不管什么时候都能使用，项目命令只能用于具体已存在的项目上。

二、全局命令

• startproject（创建爬虫项目，一个项目下有一个或多个爬虫 -- scrapy startproject project_name [project_dir]）
• genspider（进入爬虫项目目录，创建爬虫 -- cd project_name/project_dir && scrapy genspider spider_name domain_name)
• settings（获取项目配置 -- scrapy settings --get BOT_NAME）
• runspider（运行爬虫 -- scrapy runspider myspider.py）
• shell（scrapy交互终端，在不启动爬虫的情况下调试代码 -- scrapy shell http://www.baidu.com）
• fetch（对指定url发起get请求，并将返回的内容输出到控制台 -- scrapy fetch http://www.baidu.com）
• view（用不到，忽略）
• version（查看scrapy版本 -- scrapy version）

三、项目命令（局部命令）

• crawl（启动爬虫 -- scrapy crawl spider_name）
• check（用不到，忽略）
• list（列出当前的所有项目的所有爬虫 -- scrapy list）
• edit（修改spider模块的代码，很少用 -- scrapy edit spider_name）
• parse（用不到。。）
• bench（用不到。。）

四、命令源码

以上为scrapy命令行工具的基本介绍，不明白的请对具体命令自行百度。本文的重点是介绍命令的源码。如有错误之处还请指正，scrapy源码解读系列文章会一直更新。

• startproject

  1 import re
  2 import os
  3 import string
  4 from importlib import import_module
  5 from os.path import join, exists, abspath
  6 from shutil import ignore_patterns, move, copy2, copystat
  7 
  8 import scrapy
  9 from scrapy.commands import ScrapyCommand
 10 from scrapy.utils.template import render_templatefile, string_camelcase
 11 from scrapy.exceptions import UsageError
 12 
 13 
 14 TEMPLATES_TO_RENDER = (
 15     (‘scrapy.cfg‘,),
 16     (‘${project_name}‘, ‘settings.py.tmpl‘),
 17     (‘${project_name}‘, ‘items.py.tmpl‘),
 18     (‘${project_name}‘, ‘pipelines.py.tmpl‘),
 19     (‘${project_name}‘, ‘middlewares.py.tmpl‘),
 20 )
 21 
 22 IGNORE = ignore_patterns(‘*.pyc‘, ‘.svn‘)
 23 
 24 
 25 class Command(ScrapyCommand):
 26 
 27     requires_project = False
 28     default_settings = {‘LOG_ENABLED‘: False,
 29                         ‘SPIDER_LOADER_WARN_ONLY‘: True}
 30 
 31     def syntax(self):
 32         return "<project_name> [project_dir]"
 33 
 34     def short_desc(self):
 35         return "Create new project"
 36 
 37     def _is_valid_name(self, project_name):
 38         """
 39         校验项目名称，避免出现特殊符号
 40         :param project_name: 项目名称
 41         :return:
 42         """
 43         def _module_exists(module_name):
 44             try:
 45                 import_module(module_name)
 46                 return True
 47             except ImportError:
 48                 return False
 49 
 50         if not re.search(r‘^[_a-zA-Z]w*$‘, project_name):
 51             print(‘Error: Project names must begin with a letter and contain‘
 52                   ‘ only
letters, numbers and underscores‘)
 53         elif _module_exists(project_name):  # 判断项目名是否已经存在
 54             print(‘Error: Module %r already exists‘ % project_name)
 55         else:
 56             return True
 57         return False
 58 
 59     def _copytree(self, src, dst):
 60         """
 61         复制src目录的所有文件至dst目录下
 62         Since the original function always creates the directory, to resolve
 63         the issue a new function had to be created. It‘s a simple copy and
 64         was reduced for this case.
 65         More info at:
 66         https://github.com/scrapy/scrapy/pull/2005
 67         """
 68         ignore = IGNORE
 69         names = os.listdir(src)
 70         ignored_names = ignore(src, names)
 71 
 72         if not os.path.exists(dst):
 73             os.makedirs(dst)
 74 
 75         for name in names:
 76             if name in ignored_names:
 77                 continue
 78 
 79             srcname = os.path.join(src, name)
 80             dstname = os.path.join(dst, name)
 81             if os.path.isdir(srcname):
 82                 self._copytree(srcname, dstname)  # 循环调用
 83             else:
 84                 copy2(srcname, dstname)  # 复制srcname文件内容到dstname文件
 85         copystat(src, dst)  # 复制文件的元数据
 86 
 87     def run(self, args, opts):
 88         """
 89         命令的运行方法，1-验证传参是否合法 2-验证项目名是否存在 3-复制模板文件到项目目录下 
 90         4-渲染项目目录下的模板文件，实现项目初始文件的创建 5-打印创建成功的输出
 91         :param args:
 92         :param opts:
 93         :return:
 94         """
 95         if len(args) not in (1, 2):  # args为命令的传参，project_name, [project_dir(可选)]
 96             raise UsageError()
 97 
 98         project_name = args[0]
 99         project_dir = args[0]
100 
101         if len(args) == 2:
102             project_dir = args[1]
103 
104         if exists(join(project_dir, ‘scrapy.cfg‘)):
105             self.exitcode = 1
106             print(‘Error: scrapy.cfg already exists in %s‘ % abspath(project_dir))
107             return
108 
109         if not self._is_valid_name(project_name):
110             self.exitcode = 1  # 项目名不正确时，正常退出
111             return
112 
113         self._copytree(self.templates_dir, abspath(project_dir))  # 复制文件
114         move(join(project_dir, ‘module‘), join(project_dir, project_name))  # 移动文件
115         for paths in TEMPLATES_TO_RENDER:
116             path = join(*paths)
117             tplfile = join(project_dir,
118                 string.Template(path).substitute(project_name=project_name))  # string.Template方法，实现了字符串的替换，类似于字符串的%操作符和format方法
119             render_templatefile(tplfile, project_name=project_name,
120                 ProjectName=string_camelcase(project_name))  # 也是利用string.Template方法，实现模板文件内容的动态替换（替换project_name和ProjectName）
121         print("New Scrapy project ‘%s‘, using template directory ‘%s‘, "
122               "created in:" % (project_name, self.templates_dir))
123         print("    %s
" % abspath(project_dir))
124         print("You can start your first spider with:")
125         print("    cd %s" % project_dir)
126         print("    scrapy genspider example example.com")
127 
128     @property
129     def templates_dir(self):
130         _templates_base_dir = self.settings[‘TEMPLATES_DIR‘] or 131             join(scrapy.__path__[0], ‘templates‘)
132         return join(_templates_base_dir, ‘project‘)

View Code

• genspider

  1 import os
  2 import shutil
  3 import string
  4 
  5 from importlib import import_module
  6 from os.path import join, dirname, abspath, exists, splitext
  7 
  8 import scrapy
  9 from scrapy.commands import ScrapyCommand
 10 from scrapy.utils.template import render_templatefile, string_camelcase
 11 from scrapy.exceptions import UsageError
 12 
 13 
 14 def sanitize_module_name(module_name):
 15     """Sanitize the given module name, by replacing dashes and points
 16     with underscores and prefixing it with a letter if it doesn‘t start
 17     with one
 18     """
 19     module_name = module_name.replace(‘-‘, ‘_‘).replace(‘.‘, ‘_‘)
 20     if module_name[0] not in string.ascii_letters:
 21         module_name = "a" + module_name
 22     return module_name
 23 
 24 
 25 class Command(ScrapyCommand):
 26 
 27     requires_project = False
 28     default_settings = {‘LOG_ENABLED‘: False}
 29 
 30     def syntax(self):
 31         return "[options] <name> <domain>"
 32 
 33     def short_desc(self):
 34         return "Generate new spider using pre-defined templates"
 35 
 36     def add_options(self, parser):
 37         """
 38         在genspider命令后添加可选参数
 39         :param parser: 
 40         :return: 
 41         """
 42         ScrapyCommand.add_options(self, parser)
 43         parser.add_option("-l", "--list", dest="list", action="store_true",  # 列出可选的用来生成初始文件的模板
 44             help="List available templates")
 45         parser.add_option("-e", "--edit", dest="edit", action="store_true",  # 创建spider.py文件后编辑它
 46             help="Edit spider after creating it")
 47         parser.add_option("-d", "--dump", dest="dump", metavar="TEMPLATE",  # 将模板文件输出到控制台
 48             help="Dump template to standard output")
 49         parser.add_option("-t", "--template", dest="template", default="basic",  # 使用指定模板，默认使用basic.tmpl模板文件
 50             help="Uses a custom template.")
 51         parser.add_option("--force", dest="force", action="store_true",  # 强制创建spider.py文件
 52             help="If the spider already exists, overwrite it with the template")
 53 
 54     def run(self, args, opts):
 55         """
 56         命令的主方法
 57         :param args: 
 58         :param opts: 
 59         :return: 
 60         """
 61         if opts.list:  # 如果有可选参数-l，则列出模板文件并退出运行
 62             self._list_templates()
 63             return
 64         if opts.dump:  # 有可选参数-d，则读取模板文件并打印
 65             template_file = self._find_template(opts.dump)
 66             if template_file:
 67                 with open(template_file, "r") as f:
 68                     print(f.read())
 69             return
 70         if len(args) != 2:  # 如果参数不是2个，则报错（scrapy genspider spider_name domain_name）
 71             raise UsageError()
 72 
 73         name, domain = args[0:2]
 74         module = sanitize_module_name(name)
 75 
 76         if self.settings.get(‘BOT_NAME‘) == module:  # 爬虫名不能和项目名相同
 77             print("Cannot create a spider with the same name as your project")
 78             return
 79 
 80         try:
 81             spidercls = self.crawler_process.spider_loader.load(name)  # 加载爬虫名模块，如果加载出来，标明该爬虫已经存在，需要换个名字
 82         except KeyError:
 83             pass
 84         else:
 85             # if spider already exists and not --force then halt
 86             if not opts.force:
 87                 print("Spider %r already exists in module:" % name)
 88                 print("  %s" % spidercls.__module__)
 89                 return
 90         template_file = self._find_template(opts.template)  # 获取模板文件
 91         if template_file:
 92             self._genspider(module, name, domain, opts.template, template_file)  # 创建爬虫的初始文件
 93             if opts.edit:
 94                 self.exitcode = os.system(‘scrapy edit "%s"‘ % name)
 95 
 96     def _genspider(self, module, name, domain, template_name, template_file):
 97         """利用模板文件和string.Template方法创建spider初始文件"""
 98         """Generate the spider module, based on the given template"""
 99         tvars = {
100             ‘project_name‘: self.settings.get(‘BOT_NAME‘),
101             ‘ProjectName‘: string_camelcase(self.settings.get(‘BOT_NAME‘)),
102             ‘module‘: module,
103             ‘name‘: name,
104             ‘domain‘: domain,
105             ‘classname‘: ‘%sSpider‘ % ‘‘.join(s.capitalize() 106                 for s in module.split(‘_‘))
107         }
108         if self.settings.get(‘NEWSPIDER_MODULE‘):
109             spiders_module = import_module(self.settings[‘NEWSPIDER_MODULE‘])
110             spiders_dir = abspath(dirname(spiders_module.__file__))
111         else:
112             spiders_module = None
113             spiders_dir = "."
114         spider_file = "%s.py" % join(spiders_dir, module)
115         shutil.copyfile(template_file, spider_file)  # 复制模板文件内容到spider_file文件中
116         render_templatefile(spider_file, **tvars)  # 使用string.Template方法（类似字符串的format方法实现指定字符串的替换）
117         print("Created spider %r using template %r " % (name, 118             template_name), end=(‘‘ if spiders_module else ‘
‘))
119         if spiders_module:
120             print("in module:
  %s.%s" % (spiders_module.__name__, module))
121 
122     def _find_template(self, template):
123         """获取spider模板的文件路径"""
124         template_file = join(self.templates_dir, ‘%s.tmpl‘ % template)
125         if exists(template_file):
126             return template_file
127         print("Unable to find template: %s
" % template)
128         print(‘Use "scrapy genspider --list" to see all available templates.‘)
129 
130     def _list_templates(self):
131         print("Available templates:")
132         for filename in sorted(os.listdir(self.templates_dir)):
133             if filename.endswith(‘.tmpl‘):
134                 print("  %s" % splitext(filename)[0])
135 
136     @property
137     def templates_dir(self):
138         """获取模板文件的目录路径"""
139         _templates_base_dir = self.settings[‘TEMPLATES_DIR‘] or 140             join(scrapy.__path__[0], ‘templates‘)
141         return join(_templates_base_dir, ‘spiders‘)

View Code

• crawl

 1 from scrapy.commands import ScrapyCommand
 2 from scrapy.utils.conf import arglist_to_dict, feed_process_params_from_cli
 3 from scrapy.exceptions import UsageError
 4 
 5 
 6 class Command(ScrapyCommand):
 7 
 8     requires_project = True
 9 
10     def syntax(self):
11         return "[options] <spider>"
12 
13     def short_desc(self):
14         return "Run a spider"
15 
16     def add_options(self, parser):
17         ScrapyCommand.add_options(self, parser)
18         parser.add_option("-a", dest="spargs", action="append", default=[], metavar="NAME=VALUE",  # -a 爬虫运行时的参数
19                           help="set spider argument (may be repeated)")
20         parser.add_option("-o", "--output", metavar="FILE", action="append",  # -o 爬虫结果输出到指定文件中（如json文件）
21                           help="dump scraped items into FILE (use - for stdout)")
22         parser.add_option("-t", "--output-format", metavar="FORMAT",  # -t 爬虫抓取结果输出的格式
23                           help="format to use for dumping items with -o")
24 
25     def process_options(self, args, opts):
26         ScrapyCommand.process_options(self, args, opts)
27         try:
28             opts.spargs = arglist_to_dict(opts.spargs)
29         except ValueError:
30             raise UsageError("Invalid -a value, use -a NAME=VALUE", print_help=False)
31         if opts.output:
32             feeds = feed_process_params_from_cli(self.settings, opts.output, opts.output_format)
33             self.settings.set(‘FEEDS‘, feeds, priority=‘cmdline‘)
34 
35     def run(self, args, opts):
36         if len(args) < 1: # crawl命令的参数只有一个--即为要运行的爬虫名（scrapy crawl spider_name）
37             raise UsageError()
38         elif len(args) > 1:
39             raise UsageError("running ‘scrapy crawl‘ with more than one spider is no longer supported")
40         spname = args[0]
41 
42         # 这里调用的是scrapy/crawler.py文件中Crawler类的crawl方法
43         # 1-先通过spname参数创建Spider类的实例化对象，创建ExecutionEngine类的实例化对象，
44         # 2-通过调用ExecutionEngine类的open_spider方法，开启爬虫，开始调用请求进行数据的采集（仔细看看open_spider()方法！！！）
45         # 最终返回的defer对象，包含正常结果和异常结果的两种回调函数，如果爬虫的异步网络请求是正常的，则调用callback；否则调用errback（twisted的Deferred对象）
46         crawl_defer = self.crawler_process.crawl(spname, **opts.spargs)
47 
48 
49         if getattr(crawl_defer, ‘result‘, None) is not None and issubclass(crawl_defer.result.type, Exception):  # 如果result值为None而且类型为Exception的子类
50             self.exitcode = 1  # 错误返回码1
51         else:
52             self.crawler_process.start()  # 启动reactor事件监听循环（有IO事件就调用逻辑处理代码，没有就一直监听等待IO事件的到来，且几乎不消耗CPU）
53 
54             if self.crawler_process.bootstrap_failed or 55                     (hasattr(self.crawler_process, ‘has_exception‘) and self.crawler_process.has_exception):
56                 self.exitcode = 1  # 有异常退出

View Code

 

• 未完待续