README.rst

scrapy-fake-useragent

Random User-Agent middleware for Scrapy scraping framework based on fake-useragent, which picks up User-Agent strings based on usage statistics from a real world database, but also has the option to configure a generator of fake UA strings, as a backup, powered by Faker.

It also has the possibility of extending the capabilities of the middleware, by adding your own providers.

Changes

Please see CHANGELOG.

Installation

The simplest way is to install it via pip:

pip install scrapy-fake-useragent

Configuration

Turn off the built-in UserAgentMiddleware and RetryMiddleware and add RandomUserAgentMiddleware and RetryUserAgentMiddleware.

In Scrapy >=1.0:

DOWNLOADER_MIDDLEWARES = {
    'scrapy.downloadermiddlewares.useragent.UserAgentMiddleware': None,
    'scrapy.downloadermiddlewares.retry.RetryMiddleware': None,
    'scrapy_fake_useragent.middleware.RandomUserAgentMiddleware': 400,
    'scrapy_fake_useragent.middleware.RetryUserAgentMiddleware': 401,
}

In Scrapy <1.0:

DOWNLOADER_MIDDLEWARES = {
    'scrapy.contrib.downloadermiddleware.useragent.UserAgentMiddleware': None,
    'scrapy.contrib.downloadermiddleware.retry.RetryMiddleware': None,
    'scrapy_fake_useragent.middleware.RandomUserAgentMiddleware': 400,
    'scrapy_fake_useragent.middleware.RetryUserAgentMiddleware': 401,
}

Recommended setting (1.3.0+):

FAKEUSERAGENT_PROVIDERS = [
    'scrapy_fake_useragent.providers.FakeUserAgentProvider',  # this is the first provider we'll try
    'scrapy_fake_useragent.providers.FakerProvider',  # if FakeUserAgentProvider fails, we'll use faker to generate a user-agent string for us
    'scrapy_fake_useragent.providers.FixedUserAgentProvider',  # fall back to USER_AGENT value
]
USER_AGENT = '<your user agent string which you will fall back to if all other providers fail>'

---

Additional configuration information

Enabling providers

The package comes with a thin abstraction layer of User-Agent providers, which for purposes of backwards compatibility defaults to:

FAKEUSERAGENT_PROVIDERS = [
    'scrapy_fake_useragent.providers.FakeUserAgentProvider'
]

The package has also FakerProvider (powered by Faker library) and FixedUserAgentProvider implemented and available for use if needed.

Each provider is enabled individually, and used in the order they are defined. In case a provider fails execute (for instance, it can happen to fake-useragent because of it's dependency with an online service), the next one will be used.

Example of what FAKEUSERAGENT_PROVIDERS setting may look like in your case:

FAKEUSERAGENT_PROVIDERS = [
    'scrapy_fake_useragent.providers.FakeUserAgentProvider',
    'scrapy_fake_useragent.providers.FakerProvider',
    'scrapy_fake_useragent.providers.FixedUserAgentProvider',
    'mypackage.providers.CustomProvider'
]

Configuring fake-useragent

Parameter: FAKE_USERAGENT_RANDOM_UA_TYPE defaulting to random.

Other options, as example:

• firefox to mimic only Firefox browsers
• msie to mimic Internet Explorer only
• etc.

You can also set the FAKEUSERAGENT_FALLBACK option, which is a fake-useragent specific fallback. For example:

FAKEUSERAGENT_FALLBACK = 'Mozilla/5.0 (Android; Mobile; rv:40.0)'

What it does is, if the selected FAKE_USERAGENT_RANDOM_UA_TYPE fails to retrieve a UA, it will use the type set in FAKEUSERAGENT_FALLBACK.

Configuring faker

Parameter: FAKER_RANDOM_UA_TYPE defaulting to user_agent which is the way of selecting totally random User-Agents values. Other options, as example:

• chrome
• firefox
• safari
• etc. (please refer to Faker UserAgent provider documentation for the available options)

Configuring FixedUserAgent

It also comes with a fixed provider (only provides one user agent), reusing the Scrapy's default USER_AGENT setting value.

Usage with scrapy-proxies

To use with middlewares of random proxy such as scrapy-proxies, you need:

1. set RANDOM_UA_PER_PROXY to True to allow switch per proxy
2. set priority of RandomUserAgentMiddleware to be greater than scrapy-proxies, so that proxy is set before handle UA

License

The package is under MIT license. Please see LICENSE.