Grab Settings

Network options

url

Type

string

Default

None

The URL of the requested web page. You can use relative URLS, in which case Grab will build the absolute url by joining the relative URL with the URL or previously requested document. Be aware that Grab does not automatically escape unsafe characters in the URL. This is a design feature. You can use urllib.quote and urllib.quote_plus functions to make your URLs safe.

More info about valid URLs is in RFC 2396.

timeout

Type

int

Default

15

Maximum time for a network operation. If it is exceeded, GrabNetworkTimeout is raised.

connect_timeout

Type

int

Default

3

Maximum time for connection to the remote server and receipt of an initial response. If it is exceeded, GrabNetworkTimeout is raised.

process_redirect

Type

bool

Default

True

Automatically process HTTP 30* redirects.

redirect_limit

Type

int

Default

10

Set the maximum number of redirects that Grab will do for one request. Redirects follow the “Location” header in 301/302 network responses, and also follow the URL specified in meta refresh tags.

method

Type

string

Default

“GET”

Possible values

“GET”, “POST”, “PUT”, “DELETE”

The HTTP request method to use. By default, GET is used. If you specify post or multipart_post options, then Grab automatically changes the method to POST.

fields

Type

sequence of pairs or dict

Default

None

Data to be sent in serialized form. Serialization depends on the type of request. For GET on mulitpart post requests the “urlencode” method is used. By default POST/PUT requests are multipart.

multipart

Type

boolean

Default

True

Control if multipart encoding must be used for PUT/POST requests.

body

Type

bytes

Default

None

Raw bytes content to send

headers

Type

dict

Default

None

Additional HTTP-headers. The value of this option will be added to headers that Grab generates by default. See details in Work with HTTP Headers.

common_headers

Type

dict

Default

None

By default, Grab generates some common HTTP headers to mimic the behaviour of a real web browser. If you have trouble with these default headers, you can specify your own headers with this option. Please note that the usual way to specify a header is to use the headers option. See details in Work with HTTP Headers.

reuse_cookies

Type

bool

Default

True

If this option is enabled, then all cookies in each network response are stored locally and sent back with further requests to the same server.

cookies

Type

dict

Default

None

Cookies to send to the server. If the option reuse_cookies is also enabled, then cookies from the cookies option will be joined with stored cookies.

cookiefile

Type

string

Default

None

Before each request, Grab will read cookies from this file and join them with stored cookies. After each response, Grab will save all cookies to that file. The data stored in the file is a dict serialized as JSON.

Proxy Options

proxy

Type

string

Default

None

The address of the proxy server, in either “domain:port” or “ip:port” format.

proxy_userpwd

Type

string

Default

None

Security data to submit to the proxy if it requires authentication. Form of data is “username:password”

proxy_type

Type

string

Default

None

Type of proxy server. Available values are “http”, “socks4” and “socks5”.

proxy_auto_change

Type

bool

Default

True

If Grab should change the proxy before every network request.

Response Processing Options

encoding

Type

string

Default

None

The encoding (character set) is used to store document’s content as bytes. By default Grab detects encoding of document automatically. If it detects the encoding incorrectly you can specify exact encoding with this option. The encoding option is used to convert document’s bytes content into Unicode text also for biilding DOM tree of the document.

content_type

Type

string

Default

“html”

Available values

“html” and “xml”

This option controls which lxml parser is used to process the body of the response. By default, the html parser is used. If you want to parse XML, then you may need to change this option to “xml” to force the use of an XML parser which does not strip the content of CDATA nodes.

Debugging